Research ArticleMusical Rhythms Affect Heart Rate VariabilityAlgorithm and Models

Hui-Min Wang and Sheng-Chieh Huang

The Department of Electrical Engineering National Chiao Tung University Room 720 Engineering Building 51001 University Road Hsinchu 30010 Taiwan

Correspondence should be addressed to Hui-Min Wang marcelwanggmailcom

Received 15 April 2014 Accepted 21 August 2014 Published 17 September 2014

Academic Editor George E Tsekouras

Copyright copy 2014 H-M Wang and S-C Huang This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

There were a lot of psychological music experiments andmodels but there were few psychological rhythm experiments andmodelsThere were a lot of physiological music experiments but there were few physiological music models There were few physiologicalrhythm experiments but there was no physiological rhythm model We proposed a physiological rhythm model to fill this gapTwenty-two participants 4 drum loops as stimuli and electrocardiogram (ECG) were employed in this work We designed analgorithm tomap tempo complexity and energy into two heart rate variability (HRV)measures the standard deviation of normal-to-normal heartbeats (SDNN) and the ratio of low- and high-frequency powers (LFHF) these twomeasures form the physiologicalvalencearousal plane There were four major findings Initially simple and loud rhythms enhanced arousal Secondly the removalof fast and loud rhythms decreased arousal Thirdly fast rhythms increased valence Finally the removal of fast and quiet rhythmsincreased valence Our work extended the psychological model to the physiological model and deepened the musical model intothe rhythmic model Moreover this model could be the rules of automatic music generating systems

1 Introduction

The relation of music to emotion has been studied fordecades and the literature is fruitful [1] There exist a lot ofpsychologicalmodels betweenmusic and emotion [2] but thephysiological models betweenmusic and emotion are limited[3] One of the physiological actions heart rate variability(HRV) which is controlled by the autonomic nervous system(ANS) is tightly connected with emotions [4] Previously wehad analyzed the relationship between musical rhythms andHRV [5] and built two heuristic models [6 7] In this papera systematic algorithm is proposed to construct new models

Musical emotions change with psychophysiological mea-sures and musical features [8] whilst three basic questionsare highlighted [9] how do musical features evoke emotionshow do actions involved in musical emotions progress andwhich actions and brain processes are involved in musicalemotions Basically people feel what music expresses butneed not be always in a simple case only 61 of 45participants felt what they perceived [10] More particularlya stronger correlation is suggested by a recently developed

theory that the aesthetic awe accompanies by being moved(cognitive) emotions (psychological) and thrills (physio-logical) in the same time [11] The three associated levelsof musical response should be analyzed individually andcan be discussed together In the research about musicpsychology sometimes the perceived and felt emotions areexamined separately sometimes the music clips are justlabeled artificially whether they are perceived or felt had notbeen mentioned yet [12]Themappings between music spaceand emotion space [13 14] employ the following synonymsmusic mood detection [15] music emotion measurement[12] characterization [12] recognition [16 17] classification[12 18] predicting [19] or modeling [20] the review articlesdemonstrate the fruits of expertsrsquo interests [16ndash20]

In a valuable model four elementary properties aregoodness-of-fit [21] (generalization [22]) simplicity [21 22]predictability [21 22] and the relation to existing theories[22] Beyond cognitive reactions the prediction of emotionalreactions is also important in music psychology [22 23]In addition to an existing review [24] a survey of themodels [25ndash45] for emotional responses is given in this

Hindawi Publishing CorporationAdvances in Electrical EngineeringVolume 2014 Article ID 851796 14 pageshttpdxdoiorg1011552014851796

2 Advances in Electrical Engineering

paperThe acoustic features emotion spaces and methods togenerate these models are listed in Table 1 with four majorobservations Initially regression analysis [30 31 36 38ndash40 42 43 45] is widely employed while some soft computingmethods (fuzzy [41] support vector machine [34 37] neuralnetwork [35] K-nearest neighborhood [27] Gaussian mix-ture model [25]) are also used Secondly beyond the enviablevalence and arousal tension [29 32 36 40 44] is anothercommon dimension too Thirdly the felt emotion [28 35] isinfrequently examined compared to the perceived emotionFinally most works employ a wide range of acoustic featureswhile the interest in a single featuremoves frompitch [40 44]to timbre [28] and rhythm [26] For the emotion spaceRussell [13] proposed the valencearousal space and Thayer[14] reduced it to four labels the following synonyms arecommonly used valence as pleasantness [38] arousal asactivity [36 45] tension [29 32 36 40 44] as interest [43]expectancy [40] strength [38] potency [38] and resonance[27]

Whether perceived or felt emotion the three commonlyused methods including valencearousal dimensions listsof basic emotions and diverse emotion inventories are notcomplete in the musical emotion study [46] To have a morecomprehensive overview [4] the underlying mechanisms ofthe central nervous system (CNS) [46 47] and the ANS [48ndash50] should be also deliberated In the most essential sensethe acoustic properties can be perceived by the nervous sys-tem and evoke psychophysiological responses For examplefast and loud voices usually cause emotional arousal andincreased respiratory and heart rates through the auditoryand limbic systems [46 47] Some related physiological reac-tions coincidedwith the emotion expressed inmusic [48] andrated valencearousal level [49] Therefore a valencearousalmodel [50] of musical emotion had been built based onthe associated physiological measures including electromyo-gram electrocardiogram skin conductivity and respirationchanges Extended linear discriminant analysis (pLDA) wasemployed in the classification of musical emotions

Since emotion correlates with physiological responses[48ndash50] too the models of musical emotions were devel-oped by integrating the acoustic features and physiologicalmeasures [51ndash53] Despite combining with physiologicalmeasures these models [51ndash53] tend to be similar to thepsychological models mentioned above [25ndash45] in spite ofthe same perceived or felt emotion space In addition usingacoustic features to model [3] the physiological responsesseems more radical if the underlying mechanisms [46 47]are considered The model [3] employs 11 musical charac-teristics and the conclusion is that rhythmic features are themajor factors of the physiological responses (respiration skinconductance and heart rate) to music It also points out alimitation that some acoustic features would correlate witheach otherThismakes it difficult to discriminate their relativecontributions to the detected relationships

The rest of this paper is organized as follows In Section 2the rhythmic features for modeling and the HRV features forthe physiological emotion space are introduced In Section 3the experiment and developed algorithm are presentedIn Section 4 the models of the relationships between the

rhythmic and the HRV features are illustrated in figuresand the equations and tables of statistics are provided InSection 5 how tempo complexity and energy work in thepsychological experiments psychological models physio-logical experiments and physiological models are reviewedand compared Finally in Section 6 we summarize thecontributions made in this study and suggest the directionsfor the further research

2 Preliminary

21 Why Are Physiological Models Necessary To study therelation of musical features to emotion most psychologicalmodels focus on the perceived emotion as Table 1 demon-strated However the perceived emotion need not be equalto the felt emotion [10 54] and the felt emotion may nothave the related physiological responses neither [55] If themusic clips are selected by the participants themselves theperceived emotion has no statistical significance with thefelt emotion in the ranks Otherwise the differences arestatistically significant [54] Moreover some patterns of thephysiological responses appear while there is no relatedself-report [55] Since the perceived emotion felt emotionand physiological actions play different roles a physiologicalmodel is necessary beyond the psychological models

22 ValenceArousal Model Although four dimensions arenecessary for emotion spaces [56] it is difficult to real-ize a four-dimensional model Three dimensions are alsoemployed in the psychological models [27 36 38 43]Some work [57] announced that three dimensions could bereduced into two without significant loss of goodness-of-fit it also shows that the dimensional model is better thanthe discrete (categorical) model in resolution Thus Russellrsquostwo-dimensional valencearousal model [13] states its meritwithin the ebb and flow of relevant woks in the discipline ofpsychology

23 A Physiological ValenceArousal Model SDNN andLFHF two measures of HRV [58] are employed as twodimensions in our physiological valencearousal modelSDNN is the standard deviation of normal-to-normal heart-beat intervals in time domain and LFHF is the ratio oflow- and high-frequency powers after the fast Fourier trans-form (FFT) SDNN presents the variation of the circulatorysystem and LFHF presents the balance of the sympatheticand parasympathetic nervous systems [59] HRV is highlycorrelated with emotion [60ndash64] and some evidence [65ndash70] reveals that SDNN is a good indicator of valence inthe physiological perspective In general cases the normalsubjectsrsquo SDNNs are higher than the depression subjectsrsquo [65ndash67] For the normal subjectsrsquo case the SDNN levels of subjectswith positive mood are higher than the negative mood [68]For the depression subjectsrsquo case the SDNN levels of low-depression patientsrsquo are higher than the high-depressionpatientsrsquo [69 70] All of these findings reach a statisticalsignificance level Hence SDNN could be a proper indicatoras the physiological valence In addition increased LFHFvalues denote that the sympathetic nerve activity tended to be

Advances in Electrical Engineering 3

Table 1 Psychological models of musical features

Reference Year Acoustic features Emotion space MethodNumber Catalog Type A rhythm B pitch C timbre D

other

[25] 2012 70 A(1)B(17)C(52) A dynamic (1) B tonal (17) C spectral(11) timbre (41) Russell Perceived GMM

[26] 2012 11 A(11) A rhythm (11) Thayer Perceived C5

[27] 2012 20 A(10)B(6)C(4) A rhythm (7) intensity (3) B pitch andtonality (4) harmony (2) C timbre (4) VA-Resonance Perceived KNN

[28] 2012 60 C(60) C timbre (60) Thayer Felt C4

[29] 2012 9 A(6)B(3)A dynamics (1) onset frequency (1)tempo (1) meter (1) rhythmic regularity(1) syncopation (1) B harmony (1)pitch height (1) melodic expectation (1)

T Perceived MPW

[30] 2012 16 A(6)B(10)

A dynamic (1) crescendo (1) density(1) speed (1) tempo (1) articulation (1)B direction (1) surprise (1) tendency(1) mode (1) dissonance (1) harmonicsurprise (1) harmonic tempo (1) highpitch (1) low pitch (1) closure (1)

D11 Perceived Regression

[31] 2011 46 A(16)B(10)C(10)D(10)A temporal (6) rhythmic (10) Bmelodyharmony (10) C spectral (10)D lyrics (10)

Russell Perceived Regression

[32] 2011 4 A(2)B(2) Atempo (1) dynamics (1) B register (1)contour (1) T Perceived ANOVA

[33] 2011 32 C(32) C spectral contrast and MFCCs (20)echo nest timbre (12) Russell Perceived CRF

[34] 2010 292 A(87)B(22)C(18)D(165) A rhythm (87) B chords (22) C spectral(18) D lyrics (12) metadata (153) Thayer Perceived SVM

[35] 2009 13 A(3)B(6)C(4)A dynamics (2) tempo (1) B meanpitch (2) pitch variation (3) texture (1)C timbre (4)

Russell Felt NN

[36] 2009 18 A(7)B(8)C(3)A dynamics (1) rhythm (3) articulation(3) B harmony (5) register (3) Ctimbre (3)

VAT Perceived Regression

[37] 2008 29 B(M)C(M)D(F) B pitch (many) C timbre (many) Dother (few) Russell Perceived SVM

[38] 2008 9 A(4)B(3)D(2) A temporal surface (2) dynamics (2) Bregister (2) tonality (1) D other (2) VA-Strength Perceived Regression

[39] 2008 15 A(2)B(7)C(6)

A loudness (1) volume (1) B tonaldissonance (2) pure tonal (1) complextonal (1) multiplicity (1) tonality (1)chord (1) C spectral centroid (1)sharpness (2) timbral width (1) spectraldissonance (2)

Thayer Perceived Regression

[40] 2007 1 B(1) B pitch (1) T Perceived Regression[41] 2006 15 A(7)B(8) A loudness (1) duration (6) B pitch (8) Thayer Perceived Fuzzy

[42] 2006 18 A(3)B(11)C(4)A dynamics (2) tempo (1) B meanpitch (2) pitch variation (3) harmony(5) texture (1) C timbre (4)

Russell Perceived Regression

[43] 2007 8 A(3)B(4)C(1)A tempo (1) loudness (1) articulation(1) B roughness (1) melody (1) ambitus(1) register (1) C brightness (1)

VA-Interest Perceived Regression

4 Advances in Electrical Engineering

Table 1 Continued

Reference Year Acoustic features Emotion space MethodNumber Catalog Type A rhythm B pitch C timbre D

other

[44] 2005 4 B(4) B pitch (stability proximity directionmobility) T Perceived Formula

[45] 2004 6 A(3)B(3)A loudness (1) spectral centroid (1)tempo (1) B melodic (1) contour (1)texture (1)

Russell Perceived Regression

Emotion space V valence A arousal T tension D11 11 dimensionsMethod GMM Gaussian mixture model C5 classifiers 5 KNN K-nearest neighborhood C4 classifiers 4 MPW moving perceptual window ANOVAanalysis of variance CRF conditional random field SVM support vector machine NN neural network

strong thus LFHF could be used as a physiological indicatorof arousal levels [59]

24 Musical Features Employed in the Model The three mostimportant features of music are rhythm pitch and timbre[71ndash76] Although pitch [77] and timbre [78] have beenemployed to recognize emotion in music the role of rhythmseems more rudimentary In fact the rhythmic features [26]had been used in modeling the emotional responses andacquired a reasonable performance

An all-encompassing representation of the rhythm maybe the temporal organization of sound tightly connectedwith meter which refers to the periodic composition inmusic [23] Musical rhythm may have its roots in the motorrhythms controlling heart rate breathing and locomotion[79] dominated by brain stem the ancient structure [47]fast loud sounds produce an increased activation of the brain[47] and some evidence suggests that the musical rhythm canregulate emotion and motivational states [80]

Rhythm is less tractable than pitch and timbre in local-izing its specific neural substances [81] However manycultures emphasize the rhythm with two others playing a lesscrucial role [82] To study the neural basis of rhythm brainimaging psychophysical and computational approacheshave been employed [83] Beat and meter induction are thefundamental elements of cognitive mechanisms [84] whilethe representations of metric structure and neural markers ofexpectation of the beat have been found in both electroen-cephalography (EEG) and magnetoencephalography (MEG)[85] Beat perception is innate newborn infants expectdownbeats (onsets of rhythmic cycles) even unmarked bystress or the spectral features [86] Infants also engage in therhythmic movement to rhythmically regular sounds and thefaster movement tempo is associated with the faster auditorytempo [87] Although the beat perception is innate the abilityto detect rhythmic changes is more developed in adults thaninfants [88] and the trained musicians than the untrainedindividuals [89]

Of all rhythmic features experts suggest that tempo com-plexity (regularity) and energy (intensity strength dynamicloudness and volume) [15 27 90 91] are themost significant

3 Methods

31 Experiment

311 Participants There were 22 healthy subjects 15 malesand 7 females engaged in the experiment The average agewas 23 None of them had been professionally trained inmusic

312 Musical Stimuli There were four drum loops in thisstudy (L1 to L4) The parameters (tempo complexity andenergy) are listed in Table 2 and the rankings are illustratedin Figure 1

313 Apparatus TheECGsignalwas captured by a 3-channelportable device (MSI E3-80 FDA 510(k) K071085) at 500Hzsampling rate from the chest surface of the body Only thechannel-1 data was taken to be analyzed

314 Procedures All experiments were carried out in mod-erate temperature humidity and light with subjects sittingand wearing headphones (eyes closed) in a quiet room Eachsubject accepted 4 rounds of experiment in different daysEach round took 15 minutes separated as 5 stages as Figure 2illustrated Stage 1 had 5 minutes to let the subjects calmdown Stage 2 had 3 minutes of rest as the baseline for theresponses of the drum loops Stage 3 had 2 minutes as thebaseline for the responses after the drum loops Stage 4 had 3minutes of stimulus of some drum loop Stage 5 had 2minutesof rest The ECG signal from stage 2 to stage 5 was recordedand separated as epoch 1 (E1) to epoch 4 (E4) Comparison 1(C1) was the difference of E1 and E3 and comparison 2 (C2)was the difference of E2 and E4

32 Signal Processing

321 Musical Features Musical rhythm is expressed by thesuccessive notes that record the relating temporal informa-tion Tempo (119879119901) could be decided as the unit beats perminute (bpm) [92]

Advances in Electrical Engineering 5

Table 2 Rhythmic features model factors and HRV data collection of four drum loops (L1simL4)

Definition L1 L2 L3 L4Tp Tempo 8960 10530 13950 17650Cp Perceptual complexity 273 227 114 386Eg Energy sum(1198962)(1013) 199 102 383 156Valence (C1) minus1Tp minus418 minus223 050 226Valence (C2) TpEg minus233 448 minus334 564Arousal (C1) EgCp minus002 minus005 027 minus005Arousal (C2) minusTp times Eg minus015 minus010 minus042 minus023SDNN (C1) Standard deviation of all RR intervals of C1 minus501 plusmn695 minus081 plusmn1057 minus010 plusmn1163 227 plusmn1206SDNN (C2) Standard deviation of all RR intervals of C2 minus266 plusmn731 517 plusmn966 minus314 plusmn1552 508 plusmn1029LFHF (C1) Ratio of low frequency and high frequency of C1 minus004 plusmn167 minus002 plusmn081 028 plusmn089 minus006 plusmn066LFHF (C2) Ratio of low frequency and high frequency of C2 minus010 plusmn076 minus010 plusmn068 minus041 plusmn059 minus029 plusmn105

Ener

gy

4

4

35

3

3

25

2

2

15

1

1

05

0

Complexity

Tem

po

4

L4

3

L3

2

L2

1

L1

Figure 1 Four drum loop patterns (L1simL4) employed in this studywith ranked tempo (L4 gt L3 gt L2 gt L1) complexity (L4 gt L1 gt L2 gtL3) and energy (L3 gt L1 gt L4 gt L2)

The other characteristic perceptual complexity (119862119901)was obtained by asking the human subjects to judge thecomplexity of the rhythms they had listened to It was assessedby the subjects using a subjective rating of 1 to 4 on a LikertScale (4 being the most complex) [93]

The energy parameter (119864119892) was defined as Σ1198962 the

summation of square of 119896 where 119896 is the amplitude of thesignal [94 95]

322 HRV Features To acquire the measures of HRV fea-tures QRS detection was the first step [96ndash98] where 119877denotes the peak in a heartbeat signal After the abnor-mal beats were rejected [99] the mean of R-R intervals

5 8 10 13 150 (min)

Resting

C1

C2

Figure 2 The procedure of each experiment the epoch of 5minto 8min is the baseline of the physiological responses during musiclistening and the epoch of 8min to 10min is the baseline of thephysiological responses after music listening

(MRR) standard deviation of normal-to-normal R-R inter-vals (SDNN) and root ofmean of sumof square of differencesof adjacent R-R intervals (RMSSD) were measured in timedomain [59] After interpolation [100] (prepared for FFT)and detrending [101] (to filter the respiratory signal) theFFT [102] was applied to calculate the low- (LF) and high-frequency (HF) powers and their ratio (LFHF) The resultsof SDNN and LFHF are listed in Table 2 Four groups ofdata SDNN C1 SDNN C2 LFHF C1 and LFHF C2 wereobserved in our analysis

33 Algorithm For modeling the responses of some HRVmeasure and the related musical rhythms our algorithmincluded 3 steps Initially all possible combinations of therhythmic features were explored Secondly the values ofthe combinations were linearly transformed Finally thecoefficients of the linear transformationswere calculated suchthat the Euclidianmetric between the results of step 2 and therelated HRV responses is the minimum

The stimuli are 4 drum loops (rhythm) named as 1198771 to1198774 here

119877119903119894 119903119894 isin 1 2 3 4 (1)

Each rhythm relates to some HRV measure119867119903119894

119867119903119894 119903119894 isin 1 2 3 4 (2)

6 Advances in Electrical Engineering

For each rhythm there are three rhythmic features 119879119901119862119901 and 119864119892 named as 1198651 to 1198653

119865119891119894 119891119894 isin 1 2 3 (3)

Since a musical stimulus contains multiple combinationsand interactions of various features it is difficult to realizewhich feature is contributing to the perceived emotion [45]or physiological responses Our solution is considering allpossible combinations the influence of each feature 119865119891119894 islinear (order 1) of no effect (order 0) or inverse (orderminus1) Although the higher orders are plausible order one isstill the most suitable to construct a simplified model forunderstanding the relation of the musical rhythms to theirrelated HRV measures

119864119903119894 =

3

prod

119891119894=1

(119877119903119894119865119891119894)119891119894119901

119891119894119901 isin minus1 0 1 (4)

Linear transformation is necessary because the units ofrhythmic and HRV features are not uniform All we need arethe related correspondences Consider

119879119903119894 = 119909 (119864119903119894) + 119910 119909 isin 119877 119910 isin 119877 (5)

Thus for some subjectrsquos HRV responses 1198671 to 1198674 (egSDNN) to rhythms 1198771 to 1198774 some combination of rhythmicfeatures can model the relation if the metric between 119879119903119894 and119867119903119894 is the minimum Euclidean metric [27 103] is employedin this work

119863 = radic

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (6)

After squaring each side of (6) we can acquire thecoefficients of the linear transformation to get the minimummetric if the partial derivatives of 1198632 (with respect to 119909 and119910) are both zero

1198632=

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (7)

34 Statistics

341 Outliers The judgment and removal of outliers arefundamental in the processing of experimental data [104]In our study the experimental data was separated into threegroups by theminimummetric mentioned in Section 33 andthe group with larger metric was eliminated

The basic idea is to rank a group of data by the algorithmwe proposed and make the data with larger metric obsoleteThere were four groups of data SDNN (C1) SDNN (C2)LFHF (C1) and LFHF (C2) Each group had 22 records ofthe 22 participants Each record had four subrecords of drumloops L1 to L4 First we used the algorithm in each group andranked the 22 records as 1 to 22 by their minimum metricAfter summing the ranks of the four groups of these 22 par-ticipants these participants were partitioned into three levels

small medium and large metric (7 7 and 8 participants)We defined the predictable class with the small and mediummetrics and the nonpredictable class with the large metricThe nonpredictable class was excluded from our experi-mental data After removing the nonpredictable class therhythmic features and the four averages of these four groupsof predictable class were modeled by our algorithm again

342 ANOVA and 119905-Test Repeated measurements havefour major advantages obtaining the individual patterns ofchange less subjects serving as subjectsrsquo own controls andreliability the disadvantages are the complication by thedependence among repeated observations and less controlof the circumstances [105] In our experiment the repeatedmeasurements were employed

Two basic methods of ANOVA are one-way between-groups and one-way within-groups A more complicatedmethod is two-way factorial ANOVA with one between-groups and one within-groups factor [106] The repeated-measures ANOVA can be considered as a special caseof two-way ANOVA [107] The formula of repeated mea-sures ANOVA with one within-groups factor (119882) and onebetween-groups factor (119861) is listed as [106]

119910 sim 119861 lowast119882 + Error(Subject119882

) (8)

For each HRV measure there are two ANOVA tables119882is the participant If the epoch is fixed 119861 is the rhythm Ifthe rhythm is fixed 119861 is the epoch Then the Tukey honestsignificant difference test was used to acquire pair-by-paircomparisons [108] for each between-groups pair

Finally the pairwise 119905-tests [109 110] were applied for C1and C2 to realize whether the HRV responses are influencedby some particular rhythm

4 Results

Table 2 collected the musical features factors of models andHRV data in this study Valence C1 and arousal C1 revealhow rhythmic features influenced the HRV responses whilelistening to music and valence C2 and arousal C2 revealhow rhythmic features influenced the HRV responses afterlistening tomusicThevalues ofmodified combinations of therhythmic and HRV features were illustrated in Figures 3(a)and 3(b) Furthermore (9) to (12) demonstrated the relation-ships

Fast tempo enhanced SDNN that is people prefer fastertempi

SDNN (C1) prop minus1

119879119901

(9)

High intensity and low complexity enhanced LFHF thephysiological arousal

LFHF

(C1) prop119864119892

119862119901

(10)

Advances in Electrical Engineering 7

Valence

0

5

10

15

20

1 2 3 4 5 6 7 8 9(ms)

RhythmSDNN

minus5

minus10

minus15

minus20

minus25

(a)

Arousal

0

05

1

15

2

1 2 3 4 5 6 7 8 9

RhythmLFHF

minus2

minus15

minus1

minus05

(b)

Figure 3 (a) The gray bars represent the averages of SDNN and the white bars represent the approached values of the model Items 1 to 4represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to 4 in C2 (b) gray bars represent the averages of LFHF and whitebars represent the approached values of the model Items 1 to 4 represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to4 in C2 (note related numeric data is presented in Table 2)

People prefer faster tempiHowever the sound should notbe too loud if we hope the effect can remain after the musicstops

SDNN (C2) prop119879119901

119864119892

(11)

If fast and loud music stops people feel relaxed

LFHF

(C2) prop minus119879119901 times 119864119892 (12)

Table 3 is the statistical resultsThe SDNN (C2) of the fourrhythms had a significance value 001 The LFHF of L3 had asignificance value 003 while and after listening to music

5 Discussion

51 Advantages of the Proposed Model There are four majoradvantages in this work Initially the psychological musicmodels are many [2 25ndash45] but the physiological musicmodels are rare [3] Secondly the psychologicalmusicmodels[2 25ndash45] and physiological musical experiments [1 111] aremany but studies of the psychological rhythmmodel [26] andthe physiological rhythmic experiment [112] are rareThirdlyalmost all studies concern the responses during themusic butdiscussions about the responses after the music are limited[113] Finally there aremanymusic studies aboutHRV [1 111]but there is no model Our work is the first one

52 Comparison of the Models There are six levels of musicalmodel illustrated in Figure 4 (a) themodel from the results ofthe psychological experiments (b) themodel of the perceivedemotion (c) the model of the felt emotion (d) the modelfrom the physiological results (e) the proposed physiolog-ical valencearousal model while listening to the musical

rhythms and (f) the proposed physiological valencearousalmodel after listening to the musical rhythms

The six levels of model need not reach a consensusThereare four reasons Initially the perceived emotion is not alwaysthe same as the felt emotion [10 54] These two emotionsare closer if the music clips are chosen by the participantsthemselves instead of others Secondly somemeasured phys-iological responses might not have been reported by self-report [55] Thirdly if the music clips are simplified to theform of rhythm or tempo the responses are changed [112]Finally the responses while and after listening to music neednot be the same obviously

Figure 4(a) shows that 119879119901 and 119864119892 dominate arousaland 119862119901 dominates valence in a survey of psychologicalexperiments (pp 383ndash392) [1] The effects of 119879119901 and 119864119892

on arousal are clear As for valence the regular and variedrhythms derive positive emotions and the irregular rhythmsderive negative emotions (pp 383ndash392) [1] Hence valence isnegatively correlated to 119862119901

Figure 4(b) shows that 119879119901 [26 31 45] and 119864119892 [26 2731 45] dominate arousal and 119879119901 [27] and 119862119901 [27] dominatevalence in the perceived models The results of arousal forthe perceived models are the same as the psychologicalexperiments But 119879119901 and 119862119901 are viewed as the valence-basedfeatures Whether they are positively correlated to valence ornot has not been highlighted in the paper [27]

Figure 4(c) shows that 119879119901 [35] and 119864119892 [35] dominatearousal and 119879119901 [35] dominates valence in the felt modelThe results of felt and perceived models are similar But thevalence-based feature is limited at 119879119901 in the felt model [35]

Figure 4(d) shows that 119879119901 [55 113 114] and 119864119892 [114]dominate arousal and 119879119901 [55] dominates valence in the phys-iological experiments Both psychological studies (Figures4(a) 4(b) and 4(c)) and physiological studies (Figure 4(d))have the same results about arousal But the faster tempi arefound to decrease SDNN the physiological valence [55]

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

2 Advances in Electrical Engineering

paperThe acoustic features emotion spaces and methods togenerate these models are listed in Table 1 with four majorobservations Initially regression analysis [30 31 36 38ndash40 42 43 45] is widely employed while some soft computingmethods (fuzzy [41] support vector machine [34 37] neuralnetwork [35] K-nearest neighborhood [27] Gaussian mix-ture model [25]) are also used Secondly beyond the enviablevalence and arousal tension [29 32 36 40 44] is anothercommon dimension too Thirdly the felt emotion [28 35] isinfrequently examined compared to the perceived emotionFinally most works employ a wide range of acoustic featureswhile the interest in a single featuremoves frompitch [40 44]to timbre [28] and rhythm [26] For the emotion spaceRussell [13] proposed the valencearousal space and Thayer[14] reduced it to four labels the following synonyms arecommonly used valence as pleasantness [38] arousal asactivity [36 45] tension [29 32 36 40 44] as interest [43]expectancy [40] strength [38] potency [38] and resonance[27]

Whether perceived or felt emotion the three commonlyused methods including valencearousal dimensions listsof basic emotions and diverse emotion inventories are notcomplete in the musical emotion study [46] To have a morecomprehensive overview [4] the underlying mechanisms ofthe central nervous system (CNS) [46 47] and the ANS [48ndash50] should be also deliberated In the most essential sensethe acoustic properties can be perceived by the nervous sys-tem and evoke psychophysiological responses For examplefast and loud voices usually cause emotional arousal andincreased respiratory and heart rates through the auditoryand limbic systems [46 47] Some related physiological reac-tions coincidedwith the emotion expressed inmusic [48] andrated valencearousal level [49] Therefore a valencearousalmodel [50] of musical emotion had been built based onthe associated physiological measures including electromyo-gram electrocardiogram skin conductivity and respirationchanges Extended linear discriminant analysis (pLDA) wasemployed in the classification of musical emotions

Since emotion correlates with physiological responses[48ndash50] too the models of musical emotions were devel-oped by integrating the acoustic features and physiologicalmeasures [51ndash53] Despite combining with physiologicalmeasures these models [51ndash53] tend to be similar to thepsychological models mentioned above [25ndash45] in spite ofthe same perceived or felt emotion space In addition usingacoustic features to model [3] the physiological responsesseems more radical if the underlying mechanisms [46 47]are considered The model [3] employs 11 musical charac-teristics and the conclusion is that rhythmic features are themajor factors of the physiological responses (respiration skinconductance and heart rate) to music It also points out alimitation that some acoustic features would correlate witheach otherThismakes it difficult to discriminate their relativecontributions to the detected relationships

The rest of this paper is organized as follows In Section 2the rhythmic features for modeling and the HRV features forthe physiological emotion space are introduced In Section 3the experiment and developed algorithm are presentedIn Section 4 the models of the relationships between the

rhythmic and the HRV features are illustrated in figuresand the equations and tables of statistics are provided InSection 5 how tempo complexity and energy work in thepsychological experiments psychological models physio-logical experiments and physiological models are reviewedand compared Finally in Section 6 we summarize thecontributions made in this study and suggest the directionsfor the further research

2 Preliminary

21 Why Are Physiological Models Necessary To study therelation of musical features to emotion most psychologicalmodels focus on the perceived emotion as Table 1 demon-strated However the perceived emotion need not be equalto the felt emotion [10 54] and the felt emotion may nothave the related physiological responses neither [55] If themusic clips are selected by the participants themselves theperceived emotion has no statistical significance with thefelt emotion in the ranks Otherwise the differences arestatistically significant [54] Moreover some patterns of thephysiological responses appear while there is no relatedself-report [55] Since the perceived emotion felt emotionand physiological actions play different roles a physiologicalmodel is necessary beyond the psychological models

22 ValenceArousal Model Although four dimensions arenecessary for emotion spaces [56] it is difficult to real-ize a four-dimensional model Three dimensions are alsoemployed in the psychological models [27 36 38 43]Some work [57] announced that three dimensions could bereduced into two without significant loss of goodness-of-fit it also shows that the dimensional model is better thanthe discrete (categorical) model in resolution Thus Russellrsquostwo-dimensional valencearousal model [13] states its meritwithin the ebb and flow of relevant woks in the discipline ofpsychology

23 A Physiological ValenceArousal Model SDNN andLFHF two measures of HRV [58] are employed as twodimensions in our physiological valencearousal modelSDNN is the standard deviation of normal-to-normal heart-beat intervals in time domain and LFHF is the ratio oflow- and high-frequency powers after the fast Fourier trans-form (FFT) SDNN presents the variation of the circulatorysystem and LFHF presents the balance of the sympatheticand parasympathetic nervous systems [59] HRV is highlycorrelated with emotion [60ndash64] and some evidence [65ndash70] reveals that SDNN is a good indicator of valence inthe physiological perspective In general cases the normalsubjectsrsquo SDNNs are higher than the depression subjectsrsquo [65ndash67] For the normal subjectsrsquo case the SDNN levels of subjectswith positive mood are higher than the negative mood [68]For the depression subjectsrsquo case the SDNN levels of low-depression patientsrsquo are higher than the high-depressionpatientsrsquo [69 70] All of these findings reach a statisticalsignificance level Hence SDNN could be a proper indicatoras the physiological valence In addition increased LFHFvalues denote that the sympathetic nerve activity tended to be

Advances in Electrical Engineering 3

Table 1 Psychological models of musical features

Reference Year Acoustic features Emotion space MethodNumber Catalog Type A rhythm B pitch C timbre D

other

[25] 2012 70 A(1)B(17)C(52) A dynamic (1) B tonal (17) C spectral(11) timbre (41) Russell Perceived GMM

[26] 2012 11 A(11) A rhythm (11) Thayer Perceived C5

[27] 2012 20 A(10)B(6)C(4) A rhythm (7) intensity (3) B pitch andtonality (4) harmony (2) C timbre (4) VA-Resonance Perceived KNN

[28] 2012 60 C(60) C timbre (60) Thayer Felt C4

[29] 2012 9 A(6)B(3)A dynamics (1) onset frequency (1)tempo (1) meter (1) rhythmic regularity(1) syncopation (1) B harmony (1)pitch height (1) melodic expectation (1)

T Perceived MPW

[30] 2012 16 A(6)B(10)

A dynamic (1) crescendo (1) density(1) speed (1) tempo (1) articulation (1)B direction (1) surprise (1) tendency(1) mode (1) dissonance (1) harmonicsurprise (1) harmonic tempo (1) highpitch (1) low pitch (1) closure (1)

D11 Perceived Regression

[31] 2011 46 A(16)B(10)C(10)D(10)A temporal (6) rhythmic (10) Bmelodyharmony (10) C spectral (10)D lyrics (10)

Russell Perceived Regression

[32] 2011 4 A(2)B(2) Atempo (1) dynamics (1) B register (1)contour (1) T Perceived ANOVA

[33] 2011 32 C(32) C spectral contrast and MFCCs (20)echo nest timbre (12) Russell Perceived CRF

[34] 2010 292 A(87)B(22)C(18)D(165) A rhythm (87) B chords (22) C spectral(18) D lyrics (12) metadata (153) Thayer Perceived SVM

[35] 2009 13 A(3)B(6)C(4)A dynamics (2) tempo (1) B meanpitch (2) pitch variation (3) texture (1)C timbre (4)

Russell Felt NN

[36] 2009 18 A(7)B(8)C(3)A dynamics (1) rhythm (3) articulation(3) B harmony (5) register (3) Ctimbre (3)

VAT Perceived Regression

[37] 2008 29 B(M)C(M)D(F) B pitch (many) C timbre (many) Dother (few) Russell Perceived SVM

[38] 2008 9 A(4)B(3)D(2) A temporal surface (2) dynamics (2) Bregister (2) tonality (1) D other (2) VA-Strength Perceived Regression

[39] 2008 15 A(2)B(7)C(6)

A loudness (1) volume (1) B tonaldissonance (2) pure tonal (1) complextonal (1) multiplicity (1) tonality (1)chord (1) C spectral centroid (1)sharpness (2) timbral width (1) spectraldissonance (2)

Thayer Perceived Regression

[40] 2007 1 B(1) B pitch (1) T Perceived Regression[41] 2006 15 A(7)B(8) A loudness (1) duration (6) B pitch (8) Thayer Perceived Fuzzy

[42] 2006 18 A(3)B(11)C(4)A dynamics (2) tempo (1) B meanpitch (2) pitch variation (3) harmony(5) texture (1) C timbre (4)

Russell Perceived Regression

[43] 2007 8 A(3)B(4)C(1)A tempo (1) loudness (1) articulation(1) B roughness (1) melody (1) ambitus(1) register (1) C brightness (1)

VA-Interest Perceived Regression

4 Advances in Electrical Engineering

Table 1 Continued

Reference Year Acoustic features Emotion space MethodNumber Catalog Type A rhythm B pitch C timbre D

other

[44] 2005 4 B(4) B pitch (stability proximity directionmobility) T Perceived Formula

[45] 2004 6 A(3)B(3)A loudness (1) spectral centroid (1)tempo (1) B melodic (1) contour (1)texture (1)

Russell Perceived Regression

Emotion space V valence A arousal T tension D11 11 dimensionsMethod GMM Gaussian mixture model C5 classifiers 5 KNN K-nearest neighborhood C4 classifiers 4 MPW moving perceptual window ANOVAanalysis of variance CRF conditional random field SVM support vector machine NN neural network

strong thus LFHF could be used as a physiological indicatorof arousal levels [59]

24 Musical Features Employed in the Model The three mostimportant features of music are rhythm pitch and timbre[71ndash76] Although pitch [77] and timbre [78] have beenemployed to recognize emotion in music the role of rhythmseems more rudimentary In fact the rhythmic features [26]had been used in modeling the emotional responses andacquired a reasonable performance

An all-encompassing representation of the rhythm maybe the temporal organization of sound tightly connectedwith meter which refers to the periodic composition inmusic [23] Musical rhythm may have its roots in the motorrhythms controlling heart rate breathing and locomotion[79] dominated by brain stem the ancient structure [47]fast loud sounds produce an increased activation of the brain[47] and some evidence suggests that the musical rhythm canregulate emotion and motivational states [80]

Rhythm is less tractable than pitch and timbre in local-izing its specific neural substances [81] However manycultures emphasize the rhythm with two others playing a lesscrucial role [82] To study the neural basis of rhythm brainimaging psychophysical and computational approacheshave been employed [83] Beat and meter induction are thefundamental elements of cognitive mechanisms [84] whilethe representations of metric structure and neural markers ofexpectation of the beat have been found in both electroen-cephalography (EEG) and magnetoencephalography (MEG)[85] Beat perception is innate newborn infants expectdownbeats (onsets of rhythmic cycles) even unmarked bystress or the spectral features [86] Infants also engage in therhythmic movement to rhythmically regular sounds and thefaster movement tempo is associated with the faster auditorytempo [87] Although the beat perception is innate the abilityto detect rhythmic changes is more developed in adults thaninfants [88] and the trained musicians than the untrainedindividuals [89]

Of all rhythmic features experts suggest that tempo com-plexity (regularity) and energy (intensity strength dynamicloudness and volume) [15 27 90 91] are themost significant

3 Methods

31 Experiment

311 Participants There were 22 healthy subjects 15 malesand 7 females engaged in the experiment The average agewas 23 None of them had been professionally trained inmusic

312 Musical Stimuli There were four drum loops in thisstudy (L1 to L4) The parameters (tempo complexity andenergy) are listed in Table 2 and the rankings are illustratedin Figure 1

313 Apparatus TheECGsignalwas captured by a 3-channelportable device (MSI E3-80 FDA 510(k) K071085) at 500Hzsampling rate from the chest surface of the body Only thechannel-1 data was taken to be analyzed

314 Procedures All experiments were carried out in mod-erate temperature humidity and light with subjects sittingand wearing headphones (eyes closed) in a quiet room Eachsubject accepted 4 rounds of experiment in different daysEach round took 15 minutes separated as 5 stages as Figure 2illustrated Stage 1 had 5 minutes to let the subjects calmdown Stage 2 had 3 minutes of rest as the baseline for theresponses of the drum loops Stage 3 had 2 minutes as thebaseline for the responses after the drum loops Stage 4 had 3minutes of stimulus of some drum loop Stage 5 had 2minutesof rest The ECG signal from stage 2 to stage 5 was recordedand separated as epoch 1 (E1) to epoch 4 (E4) Comparison 1(C1) was the difference of E1 and E3 and comparison 2 (C2)was the difference of E2 and E4

32 Signal Processing

321 Musical Features Musical rhythm is expressed by thesuccessive notes that record the relating temporal informa-tion Tempo (119879119901) could be decided as the unit beats perminute (bpm) [92]

Advances in Electrical Engineering 5

Table 2 Rhythmic features model factors and HRV data collection of four drum loops (L1simL4)

Definition L1 L2 L3 L4Tp Tempo 8960 10530 13950 17650Cp Perceptual complexity 273 227 114 386Eg Energy sum(1198962)(1013) 199 102 383 156Valence (C1) minus1Tp minus418 minus223 050 226Valence (C2) TpEg minus233 448 minus334 564Arousal (C1) EgCp minus002 minus005 027 minus005Arousal (C2) minusTp times Eg minus015 minus010 minus042 minus023SDNN (C1) Standard deviation of all RR intervals of C1 minus501 plusmn695 minus081 plusmn1057 minus010 plusmn1163 227 plusmn1206SDNN (C2) Standard deviation of all RR intervals of C2 minus266 plusmn731 517 plusmn966 minus314 plusmn1552 508 plusmn1029LFHF (C1) Ratio of low frequency and high frequency of C1 minus004 plusmn167 minus002 plusmn081 028 plusmn089 minus006 plusmn066LFHF (C2) Ratio of low frequency and high frequency of C2 minus010 plusmn076 minus010 plusmn068 minus041 plusmn059 minus029 plusmn105

Ener

gy

4

4

35

3

3

25

2

2

15

1

1

05

0

Complexity

Tem

po

4

L4

3

L3

2

L2

1

L1

Figure 1 Four drum loop patterns (L1simL4) employed in this studywith ranked tempo (L4 gt L3 gt L2 gt L1) complexity (L4 gt L1 gt L2 gtL3) and energy (L3 gt L1 gt L4 gt L2)

The other characteristic perceptual complexity (119862119901)was obtained by asking the human subjects to judge thecomplexity of the rhythms they had listened to It was assessedby the subjects using a subjective rating of 1 to 4 on a LikertScale (4 being the most complex) [93]

The energy parameter (119864119892) was defined as Σ1198962 the

summation of square of 119896 where 119896 is the amplitude of thesignal [94 95]

322 HRV Features To acquire the measures of HRV fea-tures QRS detection was the first step [96ndash98] where 119877denotes the peak in a heartbeat signal After the abnor-mal beats were rejected [99] the mean of R-R intervals

5 8 10 13 150 (min)

Resting

C1

C2

Figure 2 The procedure of each experiment the epoch of 5minto 8min is the baseline of the physiological responses during musiclistening and the epoch of 8min to 10min is the baseline of thephysiological responses after music listening

(MRR) standard deviation of normal-to-normal R-R inter-vals (SDNN) and root ofmean of sumof square of differencesof adjacent R-R intervals (RMSSD) were measured in timedomain [59] After interpolation [100] (prepared for FFT)and detrending [101] (to filter the respiratory signal) theFFT [102] was applied to calculate the low- (LF) and high-frequency (HF) powers and their ratio (LFHF) The resultsof SDNN and LFHF are listed in Table 2 Four groups ofdata SDNN C1 SDNN C2 LFHF C1 and LFHF C2 wereobserved in our analysis

33 Algorithm For modeling the responses of some HRVmeasure and the related musical rhythms our algorithmincluded 3 steps Initially all possible combinations of therhythmic features were explored Secondly the values ofthe combinations were linearly transformed Finally thecoefficients of the linear transformationswere calculated suchthat the Euclidianmetric between the results of step 2 and therelated HRV responses is the minimum

The stimuli are 4 drum loops (rhythm) named as 1198771 to1198774 here

119877119903119894 119903119894 isin 1 2 3 4 (1)

Each rhythm relates to some HRV measure119867119903119894

119867119903119894 119903119894 isin 1 2 3 4 (2)

6 Advances in Electrical Engineering

For each rhythm there are three rhythmic features 119879119901119862119901 and 119864119892 named as 1198651 to 1198653

119865119891119894 119891119894 isin 1 2 3 (3)

Since a musical stimulus contains multiple combinationsand interactions of various features it is difficult to realizewhich feature is contributing to the perceived emotion [45]or physiological responses Our solution is considering allpossible combinations the influence of each feature 119865119891119894 islinear (order 1) of no effect (order 0) or inverse (orderminus1) Although the higher orders are plausible order one isstill the most suitable to construct a simplified model forunderstanding the relation of the musical rhythms to theirrelated HRV measures

119864119903119894 =

3

prod

119891119894=1

(119877119903119894119865119891119894)119891119894119901

119891119894119901 isin minus1 0 1 (4)

Linear transformation is necessary because the units ofrhythmic and HRV features are not uniform All we need arethe related correspondences Consider

119879119903119894 = 119909 (119864119903119894) + 119910 119909 isin 119877 119910 isin 119877 (5)

Thus for some subjectrsquos HRV responses 1198671 to 1198674 (egSDNN) to rhythms 1198771 to 1198774 some combination of rhythmicfeatures can model the relation if the metric between 119879119903119894 and119867119903119894 is the minimum Euclidean metric [27 103] is employedin this work

119863 = radic

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (6)

After squaring each side of (6) we can acquire thecoefficients of the linear transformation to get the minimummetric if the partial derivatives of 1198632 (with respect to 119909 and119910) are both zero

1198632=

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (7)

34 Statistics

341 Outliers The judgment and removal of outliers arefundamental in the processing of experimental data [104]In our study the experimental data was separated into threegroups by theminimummetric mentioned in Section 33 andthe group with larger metric was eliminated

The basic idea is to rank a group of data by the algorithmwe proposed and make the data with larger metric obsoleteThere were four groups of data SDNN (C1) SDNN (C2)LFHF (C1) and LFHF (C2) Each group had 22 records ofthe 22 participants Each record had four subrecords of drumloops L1 to L4 First we used the algorithm in each group andranked the 22 records as 1 to 22 by their minimum metricAfter summing the ranks of the four groups of these 22 par-ticipants these participants were partitioned into three levels

small medium and large metric (7 7 and 8 participants)We defined the predictable class with the small and mediummetrics and the nonpredictable class with the large metricThe nonpredictable class was excluded from our experi-mental data After removing the nonpredictable class therhythmic features and the four averages of these four groupsof predictable class were modeled by our algorithm again

342 ANOVA and 119905-Test Repeated measurements havefour major advantages obtaining the individual patterns ofchange less subjects serving as subjectsrsquo own controls andreliability the disadvantages are the complication by thedependence among repeated observations and less controlof the circumstances [105] In our experiment the repeatedmeasurements were employed

Two basic methods of ANOVA are one-way between-groups and one-way within-groups A more complicatedmethod is two-way factorial ANOVA with one between-groups and one within-groups factor [106] The repeated-measures ANOVA can be considered as a special caseof two-way ANOVA [107] The formula of repeated mea-sures ANOVA with one within-groups factor (119882) and onebetween-groups factor (119861) is listed as [106]

119910 sim 119861 lowast119882 + Error(Subject119882

) (8)

For each HRV measure there are two ANOVA tables119882is the participant If the epoch is fixed 119861 is the rhythm Ifthe rhythm is fixed 119861 is the epoch Then the Tukey honestsignificant difference test was used to acquire pair-by-paircomparisons [108] for each between-groups pair

Finally the pairwise 119905-tests [109 110] were applied for C1and C2 to realize whether the HRV responses are influencedby some particular rhythm

4 Results

Table 2 collected the musical features factors of models andHRV data in this study Valence C1 and arousal C1 revealhow rhythmic features influenced the HRV responses whilelistening to music and valence C2 and arousal C2 revealhow rhythmic features influenced the HRV responses afterlistening tomusicThevalues ofmodified combinations of therhythmic and HRV features were illustrated in Figures 3(a)and 3(b) Furthermore (9) to (12) demonstrated the relation-ships

Fast tempo enhanced SDNN that is people prefer fastertempi

SDNN (C1) prop minus1

119879119901

(9)

High intensity and low complexity enhanced LFHF thephysiological arousal

LFHF

(C1) prop119864119892

119862119901

(10)

Advances in Electrical Engineering 7

Valence

0

5

10

15

20

1 2 3 4 5 6 7 8 9(ms)

RhythmSDNN

minus5

minus10

minus15

minus20

minus25

(a)

Arousal

0

05

1

15

2

1 2 3 4 5 6 7 8 9

RhythmLFHF

minus2

minus15

minus1

minus05

(b)

Figure 3 (a) The gray bars represent the averages of SDNN and the white bars represent the approached values of the model Items 1 to 4represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to 4 in C2 (b) gray bars represent the averages of LFHF and whitebars represent the approached values of the model Items 1 to 4 represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to4 in C2 (note related numeric data is presented in Table 2)

People prefer faster tempiHowever the sound should notbe too loud if we hope the effect can remain after the musicstops

SDNN (C2) prop119879119901

119864119892

(11)

If fast and loud music stops people feel relaxed

LFHF

(C2) prop minus119879119901 times 119864119892 (12)

Table 3 is the statistical resultsThe SDNN (C2) of the fourrhythms had a significance value 001 The LFHF of L3 had asignificance value 003 while and after listening to music

5 Discussion

51 Advantages of the Proposed Model There are four majoradvantages in this work Initially the psychological musicmodels are many [2 25ndash45] but the physiological musicmodels are rare [3] Secondly the psychologicalmusicmodels[2 25ndash45] and physiological musical experiments [1 111] aremany but studies of the psychological rhythmmodel [26] andthe physiological rhythmic experiment [112] are rareThirdlyalmost all studies concern the responses during themusic butdiscussions about the responses after the music are limited[113] Finally there aremanymusic studies aboutHRV [1 111]but there is no model Our work is the first one

52 Comparison of the Models There are six levels of musicalmodel illustrated in Figure 4 (a) themodel from the results ofthe psychological experiments (b) themodel of the perceivedemotion (c) the model of the felt emotion (d) the modelfrom the physiological results (e) the proposed physiolog-ical valencearousal model while listening to the musical

rhythms and (f) the proposed physiological valencearousalmodel after listening to the musical rhythms

The six levels of model need not reach a consensusThereare four reasons Initially the perceived emotion is not alwaysthe same as the felt emotion [10 54] These two emotionsare closer if the music clips are chosen by the participantsthemselves instead of others Secondly somemeasured phys-iological responses might not have been reported by self-report [55] Thirdly if the music clips are simplified to theform of rhythm or tempo the responses are changed [112]Finally the responses while and after listening to music neednot be the same obviously

Figure 4(a) shows that 119879119901 and 119864119892 dominate arousaland 119862119901 dominates valence in a survey of psychologicalexperiments (pp 383ndash392) [1] The effects of 119879119901 and 119864119892

on arousal are clear As for valence the regular and variedrhythms derive positive emotions and the irregular rhythmsderive negative emotions (pp 383ndash392) [1] Hence valence isnegatively correlated to 119862119901

Figure 4(b) shows that 119879119901 [26 31 45] and 119864119892 [26 2731 45] dominate arousal and 119879119901 [27] and 119862119901 [27] dominatevalence in the perceived models The results of arousal forthe perceived models are the same as the psychologicalexperiments But 119879119901 and 119862119901 are viewed as the valence-basedfeatures Whether they are positively correlated to valence ornot has not been highlighted in the paper [27]

Figure 4(c) shows that 119879119901 [35] and 119864119892 [35] dominatearousal and 119879119901 [35] dominates valence in the felt modelThe results of felt and perceived models are similar But thevalence-based feature is limited at 119879119901 in the felt model [35]

Figure 4(d) shows that 119879119901 [55 113 114] and 119864119892 [114]dominate arousal and 119879119901 [55] dominates valence in the phys-iological experiments Both psychological studies (Figures4(a) 4(b) and 4(c)) and physiological studies (Figure 4(d))have the same results about arousal But the faster tempi arefound to decrease SDNN the physiological valence [55]

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

Advances in Electrical Engineering 3

Table 1 Psychological models of musical features

Reference Year Acoustic features Emotion space MethodNumber Catalog Type A rhythm B pitch C timbre D

other

[25] 2012 70 A(1)B(17)C(52) A dynamic (1) B tonal (17) C spectral(11) timbre (41) Russell Perceived GMM

[26] 2012 11 A(11) A rhythm (11) Thayer Perceived C5

[27] 2012 20 A(10)B(6)C(4) A rhythm (7) intensity (3) B pitch andtonality (4) harmony (2) C timbre (4) VA-Resonance Perceived KNN

[28] 2012 60 C(60) C timbre (60) Thayer Felt C4

[29] 2012 9 A(6)B(3)A dynamics (1) onset frequency (1)tempo (1) meter (1) rhythmic regularity(1) syncopation (1) B harmony (1)pitch height (1) melodic expectation (1)

T Perceived MPW

[30] 2012 16 A(6)B(10)

A dynamic (1) crescendo (1) density(1) speed (1) tempo (1) articulation (1)B direction (1) surprise (1) tendency(1) mode (1) dissonance (1) harmonicsurprise (1) harmonic tempo (1) highpitch (1) low pitch (1) closure (1)

D11 Perceived Regression

[31] 2011 46 A(16)B(10)C(10)D(10)A temporal (6) rhythmic (10) Bmelodyharmony (10) C spectral (10)D lyrics (10)

Russell Perceived Regression

[32] 2011 4 A(2)B(2) Atempo (1) dynamics (1) B register (1)contour (1) T Perceived ANOVA

[33] 2011 32 C(32) C spectral contrast and MFCCs (20)echo nest timbre (12) Russell Perceived CRF

[34] 2010 292 A(87)B(22)C(18)D(165) A rhythm (87) B chords (22) C spectral(18) D lyrics (12) metadata (153) Thayer Perceived SVM

[35] 2009 13 A(3)B(6)C(4)A dynamics (2) tempo (1) B meanpitch (2) pitch variation (3) texture (1)C timbre (4)

Russell Felt NN

[36] 2009 18 A(7)B(8)C(3)A dynamics (1) rhythm (3) articulation(3) B harmony (5) register (3) Ctimbre (3)

VAT Perceived Regression

[37] 2008 29 B(M)C(M)D(F) B pitch (many) C timbre (many) Dother (few) Russell Perceived SVM

[38] 2008 9 A(4)B(3)D(2) A temporal surface (2) dynamics (2) Bregister (2) tonality (1) D other (2) VA-Strength Perceived Regression

[39] 2008 15 A(2)B(7)C(6)

A loudness (1) volume (1) B tonaldissonance (2) pure tonal (1) complextonal (1) multiplicity (1) tonality (1)chord (1) C spectral centroid (1)sharpness (2) timbral width (1) spectraldissonance (2)

Thayer Perceived Regression

[40] 2007 1 B(1) B pitch (1) T Perceived Regression[41] 2006 15 A(7)B(8) A loudness (1) duration (6) B pitch (8) Thayer Perceived Fuzzy

[42] 2006 18 A(3)B(11)C(4)A dynamics (2) tempo (1) B meanpitch (2) pitch variation (3) harmony(5) texture (1) C timbre (4)

Russell Perceived Regression

[43] 2007 8 A(3)B(4)C(1)A tempo (1) loudness (1) articulation(1) B roughness (1) melody (1) ambitus(1) register (1) C brightness (1)

VA-Interest Perceived Regression

4 Advances in Electrical Engineering

Table 1 Continued

Reference Year Acoustic features Emotion space MethodNumber Catalog Type A rhythm B pitch C timbre D

other

[44] 2005 4 B(4) B pitch (stability proximity directionmobility) T Perceived Formula

[45] 2004 6 A(3)B(3)A loudness (1) spectral centroid (1)tempo (1) B melodic (1) contour (1)texture (1)

Russell Perceived Regression

Emotion space V valence A arousal T tension D11 11 dimensionsMethod GMM Gaussian mixture model C5 classifiers 5 KNN K-nearest neighborhood C4 classifiers 4 MPW moving perceptual window ANOVAanalysis of variance CRF conditional random field SVM support vector machine NN neural network

strong thus LFHF could be used as a physiological indicatorof arousal levels [59]

24 Musical Features Employed in the Model The three mostimportant features of music are rhythm pitch and timbre[71ndash76] Although pitch [77] and timbre [78] have beenemployed to recognize emotion in music the role of rhythmseems more rudimentary In fact the rhythmic features [26]had been used in modeling the emotional responses andacquired a reasonable performance

An all-encompassing representation of the rhythm maybe the temporal organization of sound tightly connectedwith meter which refers to the periodic composition inmusic [23] Musical rhythm may have its roots in the motorrhythms controlling heart rate breathing and locomotion[79] dominated by brain stem the ancient structure [47]fast loud sounds produce an increased activation of the brain[47] and some evidence suggests that the musical rhythm canregulate emotion and motivational states [80]

Rhythm is less tractable than pitch and timbre in local-izing its specific neural substances [81] However manycultures emphasize the rhythm with two others playing a lesscrucial role [82] To study the neural basis of rhythm brainimaging psychophysical and computational approacheshave been employed [83] Beat and meter induction are thefundamental elements of cognitive mechanisms [84] whilethe representations of metric structure and neural markers ofexpectation of the beat have been found in both electroen-cephalography (EEG) and magnetoencephalography (MEG)[85] Beat perception is innate newborn infants expectdownbeats (onsets of rhythmic cycles) even unmarked bystress or the spectral features [86] Infants also engage in therhythmic movement to rhythmically regular sounds and thefaster movement tempo is associated with the faster auditorytempo [87] Although the beat perception is innate the abilityto detect rhythmic changes is more developed in adults thaninfants [88] and the trained musicians than the untrainedindividuals [89]

Of all rhythmic features experts suggest that tempo com-plexity (regularity) and energy (intensity strength dynamicloudness and volume) [15 27 90 91] are themost significant

3 Methods

31 Experiment

311 Participants There were 22 healthy subjects 15 malesand 7 females engaged in the experiment The average agewas 23 None of them had been professionally trained inmusic

312 Musical Stimuli There were four drum loops in thisstudy (L1 to L4) The parameters (tempo complexity andenergy) are listed in Table 2 and the rankings are illustratedin Figure 1

313 Apparatus TheECGsignalwas captured by a 3-channelportable device (MSI E3-80 FDA 510(k) K071085) at 500Hzsampling rate from the chest surface of the body Only thechannel-1 data was taken to be analyzed

314 Procedures All experiments were carried out in mod-erate temperature humidity and light with subjects sittingand wearing headphones (eyes closed) in a quiet room Eachsubject accepted 4 rounds of experiment in different daysEach round took 15 minutes separated as 5 stages as Figure 2illustrated Stage 1 had 5 minutes to let the subjects calmdown Stage 2 had 3 minutes of rest as the baseline for theresponses of the drum loops Stage 3 had 2 minutes as thebaseline for the responses after the drum loops Stage 4 had 3minutes of stimulus of some drum loop Stage 5 had 2minutesof rest The ECG signal from stage 2 to stage 5 was recordedand separated as epoch 1 (E1) to epoch 4 (E4) Comparison 1(C1) was the difference of E1 and E3 and comparison 2 (C2)was the difference of E2 and E4

32 Signal Processing

321 Musical Features Musical rhythm is expressed by thesuccessive notes that record the relating temporal informa-tion Tempo (119879119901) could be decided as the unit beats perminute (bpm) [92]

Advances in Electrical Engineering 5

Table 2 Rhythmic features model factors and HRV data collection of four drum loops (L1simL4)

Definition L1 L2 L3 L4Tp Tempo 8960 10530 13950 17650Cp Perceptual complexity 273 227 114 386Eg Energy sum(1198962)(1013) 199 102 383 156Valence (C1) minus1Tp minus418 minus223 050 226Valence (C2) TpEg minus233 448 minus334 564Arousal (C1) EgCp minus002 minus005 027 minus005Arousal (C2) minusTp times Eg minus015 minus010 minus042 minus023SDNN (C1) Standard deviation of all RR intervals of C1 minus501 plusmn695 minus081 plusmn1057 minus010 plusmn1163 227 plusmn1206SDNN (C2) Standard deviation of all RR intervals of C2 minus266 plusmn731 517 plusmn966 minus314 plusmn1552 508 plusmn1029LFHF (C1) Ratio of low frequency and high frequency of C1 minus004 plusmn167 minus002 plusmn081 028 plusmn089 minus006 plusmn066LFHF (C2) Ratio of low frequency and high frequency of C2 minus010 plusmn076 minus010 plusmn068 minus041 plusmn059 minus029 plusmn105

Ener

gy

4

4

35

3

3

25

2

2

15

1

1

05

0

Complexity

Tem

po

4

L4

3

L3

2

L2

1

L1

Figure 1 Four drum loop patterns (L1simL4) employed in this studywith ranked tempo (L4 gt L3 gt L2 gt L1) complexity (L4 gt L1 gt L2 gtL3) and energy (L3 gt L1 gt L4 gt L2)

The other characteristic perceptual complexity (119862119901)was obtained by asking the human subjects to judge thecomplexity of the rhythms they had listened to It was assessedby the subjects using a subjective rating of 1 to 4 on a LikertScale (4 being the most complex) [93]

The energy parameter (119864119892) was defined as Σ1198962 the

summation of square of 119896 where 119896 is the amplitude of thesignal [94 95]

322 HRV Features To acquire the measures of HRV fea-tures QRS detection was the first step [96ndash98] where 119877denotes the peak in a heartbeat signal After the abnor-mal beats were rejected [99] the mean of R-R intervals

5 8 10 13 150 (min)

Resting

C1

C2

Figure 2 The procedure of each experiment the epoch of 5minto 8min is the baseline of the physiological responses during musiclistening and the epoch of 8min to 10min is the baseline of thephysiological responses after music listening

(MRR) standard deviation of normal-to-normal R-R inter-vals (SDNN) and root ofmean of sumof square of differencesof adjacent R-R intervals (RMSSD) were measured in timedomain [59] After interpolation [100] (prepared for FFT)and detrending [101] (to filter the respiratory signal) theFFT [102] was applied to calculate the low- (LF) and high-frequency (HF) powers and their ratio (LFHF) The resultsof SDNN and LFHF are listed in Table 2 Four groups ofdata SDNN C1 SDNN C2 LFHF C1 and LFHF C2 wereobserved in our analysis

33 Algorithm For modeling the responses of some HRVmeasure and the related musical rhythms our algorithmincluded 3 steps Initially all possible combinations of therhythmic features were explored Secondly the values ofthe combinations were linearly transformed Finally thecoefficients of the linear transformationswere calculated suchthat the Euclidianmetric between the results of step 2 and therelated HRV responses is the minimum

The stimuli are 4 drum loops (rhythm) named as 1198771 to1198774 here

119877119903119894 119903119894 isin 1 2 3 4 (1)

Each rhythm relates to some HRV measure119867119903119894

119867119903119894 119903119894 isin 1 2 3 4 (2)

6 Advances in Electrical Engineering

For each rhythm there are three rhythmic features 119879119901119862119901 and 119864119892 named as 1198651 to 1198653

119865119891119894 119891119894 isin 1 2 3 (3)

Since a musical stimulus contains multiple combinationsand interactions of various features it is difficult to realizewhich feature is contributing to the perceived emotion [45]or physiological responses Our solution is considering allpossible combinations the influence of each feature 119865119891119894 islinear (order 1) of no effect (order 0) or inverse (orderminus1) Although the higher orders are plausible order one isstill the most suitable to construct a simplified model forunderstanding the relation of the musical rhythms to theirrelated HRV measures

119864119903119894 =

3

prod

119891119894=1

(119877119903119894119865119891119894)119891119894119901

119891119894119901 isin minus1 0 1 (4)

Linear transformation is necessary because the units ofrhythmic and HRV features are not uniform All we need arethe related correspondences Consider

119879119903119894 = 119909 (119864119903119894) + 119910 119909 isin 119877 119910 isin 119877 (5)

Thus for some subjectrsquos HRV responses 1198671 to 1198674 (egSDNN) to rhythms 1198771 to 1198774 some combination of rhythmicfeatures can model the relation if the metric between 119879119903119894 and119867119903119894 is the minimum Euclidean metric [27 103] is employedin this work

119863 = radic

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (6)

After squaring each side of (6) we can acquire thecoefficients of the linear transformation to get the minimummetric if the partial derivatives of 1198632 (with respect to 119909 and119910) are both zero

1198632=

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (7)

34 Statistics

341 Outliers The judgment and removal of outliers arefundamental in the processing of experimental data [104]In our study the experimental data was separated into threegroups by theminimummetric mentioned in Section 33 andthe group with larger metric was eliminated

The basic idea is to rank a group of data by the algorithmwe proposed and make the data with larger metric obsoleteThere were four groups of data SDNN (C1) SDNN (C2)LFHF (C1) and LFHF (C2) Each group had 22 records ofthe 22 participants Each record had four subrecords of drumloops L1 to L4 First we used the algorithm in each group andranked the 22 records as 1 to 22 by their minimum metricAfter summing the ranks of the four groups of these 22 par-ticipants these participants were partitioned into three levels

small medium and large metric (7 7 and 8 participants)We defined the predictable class with the small and mediummetrics and the nonpredictable class with the large metricThe nonpredictable class was excluded from our experi-mental data After removing the nonpredictable class therhythmic features and the four averages of these four groupsof predictable class were modeled by our algorithm again

342 ANOVA and 119905-Test Repeated measurements havefour major advantages obtaining the individual patterns ofchange less subjects serving as subjectsrsquo own controls andreliability the disadvantages are the complication by thedependence among repeated observations and less controlof the circumstances [105] In our experiment the repeatedmeasurements were employed

Two basic methods of ANOVA are one-way between-groups and one-way within-groups A more complicatedmethod is two-way factorial ANOVA with one between-groups and one within-groups factor [106] The repeated-measures ANOVA can be considered as a special caseof two-way ANOVA [107] The formula of repeated mea-sures ANOVA with one within-groups factor (119882) and onebetween-groups factor (119861) is listed as [106]

119910 sim 119861 lowast119882 + Error(Subject119882

) (8)

For each HRV measure there are two ANOVA tables119882is the participant If the epoch is fixed 119861 is the rhythm Ifthe rhythm is fixed 119861 is the epoch Then the Tukey honestsignificant difference test was used to acquire pair-by-paircomparisons [108] for each between-groups pair

Finally the pairwise 119905-tests [109 110] were applied for C1and C2 to realize whether the HRV responses are influencedby some particular rhythm

4 Results

Table 2 collected the musical features factors of models andHRV data in this study Valence C1 and arousal C1 revealhow rhythmic features influenced the HRV responses whilelistening to music and valence C2 and arousal C2 revealhow rhythmic features influenced the HRV responses afterlistening tomusicThevalues ofmodified combinations of therhythmic and HRV features were illustrated in Figures 3(a)and 3(b) Furthermore (9) to (12) demonstrated the relation-ships

Fast tempo enhanced SDNN that is people prefer fastertempi

SDNN (C1) prop minus1

119879119901

(9)

High intensity and low complexity enhanced LFHF thephysiological arousal

LFHF

(C1) prop119864119892

119862119901

(10)

Advances in Electrical Engineering 7

Valence

0

5

10

15

20

1 2 3 4 5 6 7 8 9(ms)

RhythmSDNN

minus5

minus10

minus15

minus20

minus25

(a)

Arousal

0

05

1

15

2

1 2 3 4 5 6 7 8 9

RhythmLFHF

minus2

minus15

minus1

minus05

(b)

Figure 3 (a) The gray bars represent the averages of SDNN and the white bars represent the approached values of the model Items 1 to 4represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to 4 in C2 (b) gray bars represent the averages of LFHF and whitebars represent the approached values of the model Items 1 to 4 represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to4 in C2 (note related numeric data is presented in Table 2)

People prefer faster tempiHowever the sound should notbe too loud if we hope the effect can remain after the musicstops

SDNN (C2) prop119879119901

119864119892

(11)

If fast and loud music stops people feel relaxed

LFHF

(C2) prop minus119879119901 times 119864119892 (12)

Table 3 is the statistical resultsThe SDNN (C2) of the fourrhythms had a significance value 001 The LFHF of L3 had asignificance value 003 while and after listening to music

5 Discussion

51 Advantages of the Proposed Model There are four majoradvantages in this work Initially the psychological musicmodels are many [2 25ndash45] but the physiological musicmodels are rare [3] Secondly the psychologicalmusicmodels[2 25ndash45] and physiological musical experiments [1 111] aremany but studies of the psychological rhythmmodel [26] andthe physiological rhythmic experiment [112] are rareThirdlyalmost all studies concern the responses during themusic butdiscussions about the responses after the music are limited[113] Finally there aremanymusic studies aboutHRV [1 111]but there is no model Our work is the first one

52 Comparison of the Models There are six levels of musicalmodel illustrated in Figure 4 (a) themodel from the results ofthe psychological experiments (b) themodel of the perceivedemotion (c) the model of the felt emotion (d) the modelfrom the physiological results (e) the proposed physiolog-ical valencearousal model while listening to the musical

rhythms and (f) the proposed physiological valencearousalmodel after listening to the musical rhythms

The six levels of model need not reach a consensusThereare four reasons Initially the perceived emotion is not alwaysthe same as the felt emotion [10 54] These two emotionsare closer if the music clips are chosen by the participantsthemselves instead of others Secondly somemeasured phys-iological responses might not have been reported by self-report [55] Thirdly if the music clips are simplified to theform of rhythm or tempo the responses are changed [112]Finally the responses while and after listening to music neednot be the same obviously

Figure 4(a) shows that 119879119901 and 119864119892 dominate arousaland 119862119901 dominates valence in a survey of psychologicalexperiments (pp 383ndash392) [1] The effects of 119879119901 and 119864119892

on arousal are clear As for valence the regular and variedrhythms derive positive emotions and the irregular rhythmsderive negative emotions (pp 383ndash392) [1] Hence valence isnegatively correlated to 119862119901

Figure 4(b) shows that 119879119901 [26 31 45] and 119864119892 [26 2731 45] dominate arousal and 119879119901 [27] and 119862119901 [27] dominatevalence in the perceived models The results of arousal forthe perceived models are the same as the psychologicalexperiments But 119879119901 and 119862119901 are viewed as the valence-basedfeatures Whether they are positively correlated to valence ornot has not been highlighted in the paper [27]

Figure 4(c) shows that 119879119901 [35] and 119864119892 [35] dominatearousal and 119879119901 [35] dominates valence in the felt modelThe results of felt and perceived models are similar But thevalence-based feature is limited at 119879119901 in the felt model [35]

Figure 4(d) shows that 119879119901 [55 113 114] and 119864119892 [114]dominate arousal and 119879119901 [55] dominates valence in the phys-iological experiments Both psychological studies (Figures4(a) 4(b) and 4(c)) and physiological studies (Figure 4(d))have the same results about arousal But the faster tempi arefound to decrease SDNN the physiological valence [55]

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

4 Advances in Electrical Engineering

Table 1 Continued

Reference Year Acoustic features Emotion space MethodNumber Catalog Type A rhythm B pitch C timbre D

other

[44] 2005 4 B(4) B pitch (stability proximity directionmobility) T Perceived Formula

[45] 2004 6 A(3)B(3)A loudness (1) spectral centroid (1)tempo (1) B melodic (1) contour (1)texture (1)

Russell Perceived Regression

Emotion space V valence A arousal T tension D11 11 dimensionsMethod GMM Gaussian mixture model C5 classifiers 5 KNN K-nearest neighborhood C4 classifiers 4 MPW moving perceptual window ANOVAanalysis of variance CRF conditional random field SVM support vector machine NN neural network

strong thus LFHF could be used as a physiological indicatorof arousal levels [59]

24 Musical Features Employed in the Model The three mostimportant features of music are rhythm pitch and timbre[71ndash76] Although pitch [77] and timbre [78] have beenemployed to recognize emotion in music the role of rhythmseems more rudimentary In fact the rhythmic features [26]had been used in modeling the emotional responses andacquired a reasonable performance

An all-encompassing representation of the rhythm maybe the temporal organization of sound tightly connectedwith meter which refers to the periodic composition inmusic [23] Musical rhythm may have its roots in the motorrhythms controlling heart rate breathing and locomotion[79] dominated by brain stem the ancient structure [47]fast loud sounds produce an increased activation of the brain[47] and some evidence suggests that the musical rhythm canregulate emotion and motivational states [80]

Rhythm is less tractable than pitch and timbre in local-izing its specific neural substances [81] However manycultures emphasize the rhythm with two others playing a lesscrucial role [82] To study the neural basis of rhythm brainimaging psychophysical and computational approacheshave been employed [83] Beat and meter induction are thefundamental elements of cognitive mechanisms [84] whilethe representations of metric structure and neural markers ofexpectation of the beat have been found in both electroen-cephalography (EEG) and magnetoencephalography (MEG)[85] Beat perception is innate newborn infants expectdownbeats (onsets of rhythmic cycles) even unmarked bystress or the spectral features [86] Infants also engage in therhythmic movement to rhythmically regular sounds and thefaster movement tempo is associated with the faster auditorytempo [87] Although the beat perception is innate the abilityto detect rhythmic changes is more developed in adults thaninfants [88] and the trained musicians than the untrainedindividuals [89]

Of all rhythmic features experts suggest that tempo com-plexity (regularity) and energy (intensity strength dynamicloudness and volume) [15 27 90 91] are themost significant

3 Methods

31 Experiment

311 Participants There were 22 healthy subjects 15 malesand 7 females engaged in the experiment The average agewas 23 None of them had been professionally trained inmusic

312 Musical Stimuli There were four drum loops in thisstudy (L1 to L4) The parameters (tempo complexity andenergy) are listed in Table 2 and the rankings are illustratedin Figure 1

313 Apparatus TheECGsignalwas captured by a 3-channelportable device (MSI E3-80 FDA 510(k) K071085) at 500Hzsampling rate from the chest surface of the body Only thechannel-1 data was taken to be analyzed

314 Procedures All experiments were carried out in mod-erate temperature humidity and light with subjects sittingand wearing headphones (eyes closed) in a quiet room Eachsubject accepted 4 rounds of experiment in different daysEach round took 15 minutes separated as 5 stages as Figure 2illustrated Stage 1 had 5 minutes to let the subjects calmdown Stage 2 had 3 minutes of rest as the baseline for theresponses of the drum loops Stage 3 had 2 minutes as thebaseline for the responses after the drum loops Stage 4 had 3minutes of stimulus of some drum loop Stage 5 had 2minutesof rest The ECG signal from stage 2 to stage 5 was recordedand separated as epoch 1 (E1) to epoch 4 (E4) Comparison 1(C1) was the difference of E1 and E3 and comparison 2 (C2)was the difference of E2 and E4

32 Signal Processing

321 Musical Features Musical rhythm is expressed by thesuccessive notes that record the relating temporal informa-tion Tempo (119879119901) could be decided as the unit beats perminute (bpm) [92]

Advances in Electrical Engineering 5

Table 2 Rhythmic features model factors and HRV data collection of four drum loops (L1simL4)

Definition L1 L2 L3 L4Tp Tempo 8960 10530 13950 17650Cp Perceptual complexity 273 227 114 386Eg Energy sum(1198962)(1013) 199 102 383 156Valence (C1) minus1Tp minus418 minus223 050 226Valence (C2) TpEg minus233 448 minus334 564Arousal (C1) EgCp minus002 minus005 027 minus005Arousal (C2) minusTp times Eg minus015 minus010 minus042 minus023SDNN (C1) Standard deviation of all RR intervals of C1 minus501 plusmn695 minus081 plusmn1057 minus010 plusmn1163 227 plusmn1206SDNN (C2) Standard deviation of all RR intervals of C2 minus266 plusmn731 517 plusmn966 minus314 plusmn1552 508 plusmn1029LFHF (C1) Ratio of low frequency and high frequency of C1 minus004 plusmn167 minus002 plusmn081 028 plusmn089 minus006 plusmn066LFHF (C2) Ratio of low frequency and high frequency of C2 minus010 plusmn076 minus010 plusmn068 minus041 plusmn059 minus029 plusmn105

Ener

gy

4

4

35

3

3

25

2

2

15

1

1

05

0

Complexity

Tem

po

4

L4

3

L3

2

L2

1

L1

Figure 1 Four drum loop patterns (L1simL4) employed in this studywith ranked tempo (L4 gt L3 gt L2 gt L1) complexity (L4 gt L1 gt L2 gtL3) and energy (L3 gt L1 gt L4 gt L2)

The other characteristic perceptual complexity (119862119901)was obtained by asking the human subjects to judge thecomplexity of the rhythms they had listened to It was assessedby the subjects using a subjective rating of 1 to 4 on a LikertScale (4 being the most complex) [93]

The energy parameter (119864119892) was defined as Σ1198962 the

summation of square of 119896 where 119896 is the amplitude of thesignal [94 95]

322 HRV Features To acquire the measures of HRV fea-tures QRS detection was the first step [96ndash98] where 119877denotes the peak in a heartbeat signal After the abnor-mal beats were rejected [99] the mean of R-R intervals

5 8 10 13 150 (min)

Resting

C1

C2

Figure 2 The procedure of each experiment the epoch of 5minto 8min is the baseline of the physiological responses during musiclistening and the epoch of 8min to 10min is the baseline of thephysiological responses after music listening

(MRR) standard deviation of normal-to-normal R-R inter-vals (SDNN) and root ofmean of sumof square of differencesof adjacent R-R intervals (RMSSD) were measured in timedomain [59] After interpolation [100] (prepared for FFT)and detrending [101] (to filter the respiratory signal) theFFT [102] was applied to calculate the low- (LF) and high-frequency (HF) powers and their ratio (LFHF) The resultsof SDNN and LFHF are listed in Table 2 Four groups ofdata SDNN C1 SDNN C2 LFHF C1 and LFHF C2 wereobserved in our analysis

33 Algorithm For modeling the responses of some HRVmeasure and the related musical rhythms our algorithmincluded 3 steps Initially all possible combinations of therhythmic features were explored Secondly the values ofthe combinations were linearly transformed Finally thecoefficients of the linear transformationswere calculated suchthat the Euclidianmetric between the results of step 2 and therelated HRV responses is the minimum

The stimuli are 4 drum loops (rhythm) named as 1198771 to1198774 here

119877119903119894 119903119894 isin 1 2 3 4 (1)

Each rhythm relates to some HRV measure119867119903119894

119867119903119894 119903119894 isin 1 2 3 4 (2)

6 Advances in Electrical Engineering

For each rhythm there are three rhythmic features 119879119901119862119901 and 119864119892 named as 1198651 to 1198653

119865119891119894 119891119894 isin 1 2 3 (3)

Since a musical stimulus contains multiple combinationsand interactions of various features it is difficult to realizewhich feature is contributing to the perceived emotion [45]or physiological responses Our solution is considering allpossible combinations the influence of each feature 119865119891119894 islinear (order 1) of no effect (order 0) or inverse (orderminus1) Although the higher orders are plausible order one isstill the most suitable to construct a simplified model forunderstanding the relation of the musical rhythms to theirrelated HRV measures

119864119903119894 =

3

prod

119891119894=1

(119877119903119894119865119891119894)119891119894119901

119891119894119901 isin minus1 0 1 (4)

Linear transformation is necessary because the units ofrhythmic and HRV features are not uniform All we need arethe related correspondences Consider

119879119903119894 = 119909 (119864119903119894) + 119910 119909 isin 119877 119910 isin 119877 (5)

Thus for some subjectrsquos HRV responses 1198671 to 1198674 (egSDNN) to rhythms 1198771 to 1198774 some combination of rhythmicfeatures can model the relation if the metric between 119879119903119894 and119867119903119894 is the minimum Euclidean metric [27 103] is employedin this work

119863 = radic

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (6)

After squaring each side of (6) we can acquire thecoefficients of the linear transformation to get the minimummetric if the partial derivatives of 1198632 (with respect to 119909 and119910) are both zero

1198632=

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (7)

34 Statistics

341 Outliers The judgment and removal of outliers arefundamental in the processing of experimental data [104]In our study the experimental data was separated into threegroups by theminimummetric mentioned in Section 33 andthe group with larger metric was eliminated

The basic idea is to rank a group of data by the algorithmwe proposed and make the data with larger metric obsoleteThere were four groups of data SDNN (C1) SDNN (C2)LFHF (C1) and LFHF (C2) Each group had 22 records ofthe 22 participants Each record had four subrecords of drumloops L1 to L4 First we used the algorithm in each group andranked the 22 records as 1 to 22 by their minimum metricAfter summing the ranks of the four groups of these 22 par-ticipants these participants were partitioned into three levels

small medium and large metric (7 7 and 8 participants)We defined the predictable class with the small and mediummetrics and the nonpredictable class with the large metricThe nonpredictable class was excluded from our experi-mental data After removing the nonpredictable class therhythmic features and the four averages of these four groupsof predictable class were modeled by our algorithm again

342 ANOVA and 119905-Test Repeated measurements havefour major advantages obtaining the individual patterns ofchange less subjects serving as subjectsrsquo own controls andreliability the disadvantages are the complication by thedependence among repeated observations and less controlof the circumstances [105] In our experiment the repeatedmeasurements were employed

Two basic methods of ANOVA are one-way between-groups and one-way within-groups A more complicatedmethod is two-way factorial ANOVA with one between-groups and one within-groups factor [106] The repeated-measures ANOVA can be considered as a special caseof two-way ANOVA [107] The formula of repeated mea-sures ANOVA with one within-groups factor (119882) and onebetween-groups factor (119861) is listed as [106]

119910 sim 119861 lowast119882 + Error(Subject119882

) (8)

For each HRV measure there are two ANOVA tables119882is the participant If the epoch is fixed 119861 is the rhythm Ifthe rhythm is fixed 119861 is the epoch Then the Tukey honestsignificant difference test was used to acquire pair-by-paircomparisons [108] for each between-groups pair

Finally the pairwise 119905-tests [109 110] were applied for C1and C2 to realize whether the HRV responses are influencedby some particular rhythm

4 Results

Table 2 collected the musical features factors of models andHRV data in this study Valence C1 and arousal C1 revealhow rhythmic features influenced the HRV responses whilelistening to music and valence C2 and arousal C2 revealhow rhythmic features influenced the HRV responses afterlistening tomusicThevalues ofmodified combinations of therhythmic and HRV features were illustrated in Figures 3(a)and 3(b) Furthermore (9) to (12) demonstrated the relation-ships

Fast tempo enhanced SDNN that is people prefer fastertempi

SDNN (C1) prop minus1

119879119901

(9)

High intensity and low complexity enhanced LFHF thephysiological arousal

LFHF

(C1) prop119864119892

119862119901

(10)

Advances in Electrical Engineering 7

Valence

0

5

10

15

20

1 2 3 4 5 6 7 8 9(ms)

RhythmSDNN

minus5

minus10

minus15

minus20

minus25

(a)

Arousal

0

05

1

15

2

1 2 3 4 5 6 7 8 9

RhythmLFHF

minus2

minus15

minus1

minus05

(b)

Figure 3 (a) The gray bars represent the averages of SDNN and the white bars represent the approached values of the model Items 1 to 4represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to 4 in C2 (b) gray bars represent the averages of LFHF and whitebars represent the approached values of the model Items 1 to 4 represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to4 in C2 (note related numeric data is presented in Table 2)

People prefer faster tempiHowever the sound should notbe too loud if we hope the effect can remain after the musicstops

SDNN (C2) prop119879119901

119864119892

(11)

If fast and loud music stops people feel relaxed

LFHF

(C2) prop minus119879119901 times 119864119892 (12)

Table 3 is the statistical resultsThe SDNN (C2) of the fourrhythms had a significance value 001 The LFHF of L3 had asignificance value 003 while and after listening to music

5 Discussion

51 Advantages of the Proposed Model There are four majoradvantages in this work Initially the psychological musicmodels are many [2 25ndash45] but the physiological musicmodels are rare [3] Secondly the psychologicalmusicmodels[2 25ndash45] and physiological musical experiments [1 111] aremany but studies of the psychological rhythmmodel [26] andthe physiological rhythmic experiment [112] are rareThirdlyalmost all studies concern the responses during themusic butdiscussions about the responses after the music are limited[113] Finally there aremanymusic studies aboutHRV [1 111]but there is no model Our work is the first one

52 Comparison of the Models There are six levels of musicalmodel illustrated in Figure 4 (a) themodel from the results ofthe psychological experiments (b) themodel of the perceivedemotion (c) the model of the felt emotion (d) the modelfrom the physiological results (e) the proposed physiolog-ical valencearousal model while listening to the musical

rhythms and (f) the proposed physiological valencearousalmodel after listening to the musical rhythms

The six levels of model need not reach a consensusThereare four reasons Initially the perceived emotion is not alwaysthe same as the felt emotion [10 54] These two emotionsare closer if the music clips are chosen by the participantsthemselves instead of others Secondly somemeasured phys-iological responses might not have been reported by self-report [55] Thirdly if the music clips are simplified to theform of rhythm or tempo the responses are changed [112]Finally the responses while and after listening to music neednot be the same obviously

Figure 4(a) shows that 119879119901 and 119864119892 dominate arousaland 119862119901 dominates valence in a survey of psychologicalexperiments (pp 383ndash392) [1] The effects of 119879119901 and 119864119892

on arousal are clear As for valence the regular and variedrhythms derive positive emotions and the irregular rhythmsderive negative emotions (pp 383ndash392) [1] Hence valence isnegatively correlated to 119862119901

Figure 4(b) shows that 119879119901 [26 31 45] and 119864119892 [26 2731 45] dominate arousal and 119879119901 [27] and 119862119901 [27] dominatevalence in the perceived models The results of arousal forthe perceived models are the same as the psychologicalexperiments But 119879119901 and 119862119901 are viewed as the valence-basedfeatures Whether they are positively correlated to valence ornot has not been highlighted in the paper [27]

Figure 4(c) shows that 119879119901 [35] and 119864119892 [35] dominatearousal and 119879119901 [35] dominates valence in the felt modelThe results of felt and perceived models are similar But thevalence-based feature is limited at 119879119901 in the felt model [35]

Figure 4(d) shows that 119879119901 [55 113 114] and 119864119892 [114]dominate arousal and 119879119901 [55] dominates valence in the phys-iological experiments Both psychological studies (Figures4(a) 4(b) and 4(c)) and physiological studies (Figure 4(d))have the same results about arousal But the faster tempi arefound to decrease SDNN the physiological valence [55]

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

Advances in Electrical Engineering 5

Table 2 Rhythmic features model factors and HRV data collection of four drum loops (L1simL4)

Definition L1 L2 L3 L4Tp Tempo 8960 10530 13950 17650Cp Perceptual complexity 273 227 114 386Eg Energy sum(1198962)(1013) 199 102 383 156Valence (C1) minus1Tp minus418 minus223 050 226Valence (C2) TpEg minus233 448 minus334 564Arousal (C1) EgCp minus002 minus005 027 minus005Arousal (C2) minusTp times Eg minus015 minus010 minus042 minus023SDNN (C1) Standard deviation of all RR intervals of C1 minus501 plusmn695 minus081 plusmn1057 minus010 plusmn1163 227 plusmn1206SDNN (C2) Standard deviation of all RR intervals of C2 minus266 plusmn731 517 plusmn966 minus314 plusmn1552 508 plusmn1029LFHF (C1) Ratio of low frequency and high frequency of C1 minus004 plusmn167 minus002 plusmn081 028 plusmn089 minus006 plusmn066LFHF (C2) Ratio of low frequency and high frequency of C2 minus010 plusmn076 minus010 plusmn068 minus041 plusmn059 minus029 plusmn105

Ener

gy

4

4

35

3

3

25

2

2

15

1

1

05

0

Complexity

Tem

po

4

L4

3

L3

2

L2

1

L1

Figure 1 Four drum loop patterns (L1simL4) employed in this studywith ranked tempo (L4 gt L3 gt L2 gt L1) complexity (L4 gt L1 gt L2 gtL3) and energy (L3 gt L1 gt L4 gt L2)

The other characteristic perceptual complexity (119862119901)was obtained by asking the human subjects to judge thecomplexity of the rhythms they had listened to It was assessedby the subjects using a subjective rating of 1 to 4 on a LikertScale (4 being the most complex) [93]

The energy parameter (119864119892) was defined as Σ1198962 the

summation of square of 119896 where 119896 is the amplitude of thesignal [94 95]

322 HRV Features To acquire the measures of HRV fea-tures QRS detection was the first step [96ndash98] where 119877denotes the peak in a heartbeat signal After the abnor-mal beats were rejected [99] the mean of R-R intervals

5 8 10 13 150 (min)

Resting

C1

C2

Figure 2 The procedure of each experiment the epoch of 5minto 8min is the baseline of the physiological responses during musiclistening and the epoch of 8min to 10min is the baseline of thephysiological responses after music listening

(MRR) standard deviation of normal-to-normal R-R inter-vals (SDNN) and root ofmean of sumof square of differencesof adjacent R-R intervals (RMSSD) were measured in timedomain [59] After interpolation [100] (prepared for FFT)and detrending [101] (to filter the respiratory signal) theFFT [102] was applied to calculate the low- (LF) and high-frequency (HF) powers and their ratio (LFHF) The resultsof SDNN and LFHF are listed in Table 2 Four groups ofdata SDNN C1 SDNN C2 LFHF C1 and LFHF C2 wereobserved in our analysis

33 Algorithm For modeling the responses of some HRVmeasure and the related musical rhythms our algorithmincluded 3 steps Initially all possible combinations of therhythmic features were explored Secondly the values ofthe combinations were linearly transformed Finally thecoefficients of the linear transformationswere calculated suchthat the Euclidianmetric between the results of step 2 and therelated HRV responses is the minimum

The stimuli are 4 drum loops (rhythm) named as 1198771 to1198774 here

119877119903119894 119903119894 isin 1 2 3 4 (1)

Each rhythm relates to some HRV measure119867119903119894

119867119903119894 119903119894 isin 1 2 3 4 (2)

6 Advances in Electrical Engineering

For each rhythm there are three rhythmic features 119879119901119862119901 and 119864119892 named as 1198651 to 1198653

119865119891119894 119891119894 isin 1 2 3 (3)

Since a musical stimulus contains multiple combinationsand interactions of various features it is difficult to realizewhich feature is contributing to the perceived emotion [45]or physiological responses Our solution is considering allpossible combinations the influence of each feature 119865119891119894 islinear (order 1) of no effect (order 0) or inverse (orderminus1) Although the higher orders are plausible order one isstill the most suitable to construct a simplified model forunderstanding the relation of the musical rhythms to theirrelated HRV measures

119864119903119894 =

3

prod

119891119894=1

(119877119903119894119865119891119894)119891119894119901

119891119894119901 isin minus1 0 1 (4)

Linear transformation is necessary because the units ofrhythmic and HRV features are not uniform All we need arethe related correspondences Consider

119879119903119894 = 119909 (119864119903119894) + 119910 119909 isin 119877 119910 isin 119877 (5)

Thus for some subjectrsquos HRV responses 1198671 to 1198674 (egSDNN) to rhythms 1198771 to 1198774 some combination of rhythmicfeatures can model the relation if the metric between 119879119903119894 and119867119903119894 is the minimum Euclidean metric [27 103] is employedin this work

119863 = radic

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (6)

After squaring each side of (6) we can acquire thecoefficients of the linear transformation to get the minimummetric if the partial derivatives of 1198632 (with respect to 119909 and119910) are both zero

1198632=

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (7)

34 Statistics

341 Outliers The judgment and removal of outliers arefundamental in the processing of experimental data [104]In our study the experimental data was separated into threegroups by theminimummetric mentioned in Section 33 andthe group with larger metric was eliminated

The basic idea is to rank a group of data by the algorithmwe proposed and make the data with larger metric obsoleteThere were four groups of data SDNN (C1) SDNN (C2)LFHF (C1) and LFHF (C2) Each group had 22 records ofthe 22 participants Each record had four subrecords of drumloops L1 to L4 First we used the algorithm in each group andranked the 22 records as 1 to 22 by their minimum metricAfter summing the ranks of the four groups of these 22 par-ticipants these participants were partitioned into three levels

small medium and large metric (7 7 and 8 participants)We defined the predictable class with the small and mediummetrics and the nonpredictable class with the large metricThe nonpredictable class was excluded from our experi-mental data After removing the nonpredictable class therhythmic features and the four averages of these four groupsof predictable class were modeled by our algorithm again

342 ANOVA and 119905-Test Repeated measurements havefour major advantages obtaining the individual patterns ofchange less subjects serving as subjectsrsquo own controls andreliability the disadvantages are the complication by thedependence among repeated observations and less controlof the circumstances [105] In our experiment the repeatedmeasurements were employed

Two basic methods of ANOVA are one-way between-groups and one-way within-groups A more complicatedmethod is two-way factorial ANOVA with one between-groups and one within-groups factor [106] The repeated-measures ANOVA can be considered as a special caseof two-way ANOVA [107] The formula of repeated mea-sures ANOVA with one within-groups factor (119882) and onebetween-groups factor (119861) is listed as [106]

119910 sim 119861 lowast119882 + Error(Subject119882

) (8)

For each HRV measure there are two ANOVA tables119882is the participant If the epoch is fixed 119861 is the rhythm Ifthe rhythm is fixed 119861 is the epoch Then the Tukey honestsignificant difference test was used to acquire pair-by-paircomparisons [108] for each between-groups pair

Finally the pairwise 119905-tests [109 110] were applied for C1and C2 to realize whether the HRV responses are influencedby some particular rhythm

4 Results

Table 2 collected the musical features factors of models andHRV data in this study Valence C1 and arousal C1 revealhow rhythmic features influenced the HRV responses whilelistening to music and valence C2 and arousal C2 revealhow rhythmic features influenced the HRV responses afterlistening tomusicThevalues ofmodified combinations of therhythmic and HRV features were illustrated in Figures 3(a)and 3(b) Furthermore (9) to (12) demonstrated the relation-ships

Fast tempo enhanced SDNN that is people prefer fastertempi

SDNN (C1) prop minus1

119879119901

(9)

High intensity and low complexity enhanced LFHF thephysiological arousal

LFHF

(C1) prop119864119892

119862119901

(10)

Advances in Electrical Engineering 7

Valence

0

5

10

15

20

1 2 3 4 5 6 7 8 9(ms)

RhythmSDNN

minus5

minus10

minus15

minus20

minus25

(a)

Arousal

0

05

1

15

2

1 2 3 4 5 6 7 8 9

RhythmLFHF

minus2

minus15

minus1

minus05

(b)

Figure 3 (a) The gray bars represent the averages of SDNN and the white bars represent the approached values of the model Items 1 to 4represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to 4 in C2 (b) gray bars represent the averages of LFHF and whitebars represent the approached values of the model Items 1 to 4 represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to4 in C2 (note related numeric data is presented in Table 2)

People prefer faster tempiHowever the sound should notbe too loud if we hope the effect can remain after the musicstops

SDNN (C2) prop119879119901

119864119892

(11)

If fast and loud music stops people feel relaxed

LFHF

(C2) prop minus119879119901 times 119864119892 (12)

Table 3 is the statistical resultsThe SDNN (C2) of the fourrhythms had a significance value 001 The LFHF of L3 had asignificance value 003 while and after listening to music

5 Discussion

51 Advantages of the Proposed Model There are four majoradvantages in this work Initially the psychological musicmodels are many [2 25ndash45] but the physiological musicmodels are rare [3] Secondly the psychologicalmusicmodels[2 25ndash45] and physiological musical experiments [1 111] aremany but studies of the psychological rhythmmodel [26] andthe physiological rhythmic experiment [112] are rareThirdlyalmost all studies concern the responses during themusic butdiscussions about the responses after the music are limited[113] Finally there aremanymusic studies aboutHRV [1 111]but there is no model Our work is the first one

52 Comparison of the Models There are six levels of musicalmodel illustrated in Figure 4 (a) themodel from the results ofthe psychological experiments (b) themodel of the perceivedemotion (c) the model of the felt emotion (d) the modelfrom the physiological results (e) the proposed physiolog-ical valencearousal model while listening to the musical

rhythms and (f) the proposed physiological valencearousalmodel after listening to the musical rhythms

The six levels of model need not reach a consensusThereare four reasons Initially the perceived emotion is not alwaysthe same as the felt emotion [10 54] These two emotionsare closer if the music clips are chosen by the participantsthemselves instead of others Secondly somemeasured phys-iological responses might not have been reported by self-report [55] Thirdly if the music clips are simplified to theform of rhythm or tempo the responses are changed [112]Finally the responses while and after listening to music neednot be the same obviously

Figure 4(a) shows that 119879119901 and 119864119892 dominate arousaland 119862119901 dominates valence in a survey of psychologicalexperiments (pp 383ndash392) [1] The effects of 119879119901 and 119864119892

on arousal are clear As for valence the regular and variedrhythms derive positive emotions and the irregular rhythmsderive negative emotions (pp 383ndash392) [1] Hence valence isnegatively correlated to 119862119901

Figure 4(b) shows that 119879119901 [26 31 45] and 119864119892 [26 2731 45] dominate arousal and 119879119901 [27] and 119862119901 [27] dominatevalence in the perceived models The results of arousal forthe perceived models are the same as the psychologicalexperiments But 119879119901 and 119862119901 are viewed as the valence-basedfeatures Whether they are positively correlated to valence ornot has not been highlighted in the paper [27]

Figure 4(c) shows that 119879119901 [35] and 119864119892 [35] dominatearousal and 119879119901 [35] dominates valence in the felt modelThe results of felt and perceived models are similar But thevalence-based feature is limited at 119879119901 in the felt model [35]

Figure 4(d) shows that 119879119901 [55 113 114] and 119864119892 [114]dominate arousal and 119879119901 [55] dominates valence in the phys-iological experiments Both psychological studies (Figures4(a) 4(b) and 4(c)) and physiological studies (Figure 4(d))have the same results about arousal But the faster tempi arefound to decrease SDNN the physiological valence [55]

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

6 Advances in Electrical Engineering

For each rhythm there are three rhythmic features 119879119901119862119901 and 119864119892 named as 1198651 to 1198653

119865119891119894 119891119894 isin 1 2 3 (3)

Since a musical stimulus contains multiple combinationsand interactions of various features it is difficult to realizewhich feature is contributing to the perceived emotion [45]or physiological responses Our solution is considering allpossible combinations the influence of each feature 119865119891119894 islinear (order 1) of no effect (order 0) or inverse (orderminus1) Although the higher orders are plausible order one isstill the most suitable to construct a simplified model forunderstanding the relation of the musical rhythms to theirrelated HRV measures

119864119903119894 =

3

prod

119891119894=1

(119877119903119894119865119891119894)119891119894119901

119891119894119901 isin minus1 0 1 (4)

Linear transformation is necessary because the units ofrhythmic and HRV features are not uniform All we need arethe related correspondences Consider

119879119903119894 = 119909 (119864119903119894) + 119910 119909 isin 119877 119910 isin 119877 (5)

Thus for some subjectrsquos HRV responses 1198671 to 1198674 (egSDNN) to rhythms 1198771 to 1198774 some combination of rhythmicfeatures can model the relation if the metric between 119879119903119894 and119867119903119894 is the minimum Euclidean metric [27 103] is employedin this work

119863 = radic

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (6)

After squaring each side of (6) we can acquire thecoefficients of the linear transformation to get the minimummetric if the partial derivatives of 1198632 (with respect to 119909 and119910) are both zero

1198632=

4

sum

119903119894=1

(119879119903119894 minus 119867119903119894)2 (7)

34 Statistics

341 Outliers The judgment and removal of outliers arefundamental in the processing of experimental data [104]In our study the experimental data was separated into threegroups by theminimummetric mentioned in Section 33 andthe group with larger metric was eliminated

The basic idea is to rank a group of data by the algorithmwe proposed and make the data with larger metric obsoleteThere were four groups of data SDNN (C1) SDNN (C2)LFHF (C1) and LFHF (C2) Each group had 22 records ofthe 22 participants Each record had four subrecords of drumloops L1 to L4 First we used the algorithm in each group andranked the 22 records as 1 to 22 by their minimum metricAfter summing the ranks of the four groups of these 22 par-ticipants these participants were partitioned into three levels

small medium and large metric (7 7 and 8 participants)We defined the predictable class with the small and mediummetrics and the nonpredictable class with the large metricThe nonpredictable class was excluded from our experi-mental data After removing the nonpredictable class therhythmic features and the four averages of these four groupsof predictable class were modeled by our algorithm again

342 ANOVA and 119905-Test Repeated measurements havefour major advantages obtaining the individual patterns ofchange less subjects serving as subjectsrsquo own controls andreliability the disadvantages are the complication by thedependence among repeated observations and less controlof the circumstances [105] In our experiment the repeatedmeasurements were employed

Two basic methods of ANOVA are one-way between-groups and one-way within-groups A more complicatedmethod is two-way factorial ANOVA with one between-groups and one within-groups factor [106] The repeated-measures ANOVA can be considered as a special caseof two-way ANOVA [107] The formula of repeated mea-sures ANOVA with one within-groups factor (119882) and onebetween-groups factor (119861) is listed as [106]

119910 sim 119861 lowast119882 + Error(Subject119882

) (8)

For each HRV measure there are two ANOVA tables119882is the participant If the epoch is fixed 119861 is the rhythm Ifthe rhythm is fixed 119861 is the epoch Then the Tukey honestsignificant difference test was used to acquire pair-by-paircomparisons [108] for each between-groups pair

Finally the pairwise 119905-tests [109 110] were applied for C1and C2 to realize whether the HRV responses are influencedby some particular rhythm

4 Results

Table 2 collected the musical features factors of models andHRV data in this study Valence C1 and arousal C1 revealhow rhythmic features influenced the HRV responses whilelistening to music and valence C2 and arousal C2 revealhow rhythmic features influenced the HRV responses afterlistening tomusicThevalues ofmodified combinations of therhythmic and HRV features were illustrated in Figures 3(a)and 3(b) Furthermore (9) to (12) demonstrated the relation-ships

Fast tempo enhanced SDNN that is people prefer fastertempi

SDNN (C1) prop minus1

119879119901

(9)

High intensity and low complexity enhanced LFHF thephysiological arousal

LFHF

(C1) prop119864119892

119862119901

(10)

Advances in Electrical Engineering 7

Valence

0

5

10

15

20

1 2 3 4 5 6 7 8 9(ms)

RhythmSDNN

minus5

minus10

minus15

minus20

minus25

(a)

Arousal

0

05

1

15

2

1 2 3 4 5 6 7 8 9

RhythmLFHF

minus2

minus15

minus1

minus05

(b)

Figure 3 (a) The gray bars represent the averages of SDNN and the white bars represent the approached values of the model Items 1 to 4represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to 4 in C2 (b) gray bars represent the averages of LFHF and whitebars represent the approached values of the model Items 1 to 4 represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to4 in C2 (note related numeric data is presented in Table 2)

People prefer faster tempiHowever the sound should notbe too loud if we hope the effect can remain after the musicstops

SDNN (C2) prop119879119901

119864119892

(11)

If fast and loud music stops people feel relaxed

LFHF

(C2) prop minus119879119901 times 119864119892 (12)

Table 3 is the statistical resultsThe SDNN (C2) of the fourrhythms had a significance value 001 The LFHF of L3 had asignificance value 003 while and after listening to music

5 Discussion

51 Advantages of the Proposed Model There are four majoradvantages in this work Initially the psychological musicmodels are many [2 25ndash45] but the physiological musicmodels are rare [3] Secondly the psychologicalmusicmodels[2 25ndash45] and physiological musical experiments [1 111] aremany but studies of the psychological rhythmmodel [26] andthe physiological rhythmic experiment [112] are rareThirdlyalmost all studies concern the responses during themusic butdiscussions about the responses after the music are limited[113] Finally there aremanymusic studies aboutHRV [1 111]but there is no model Our work is the first one

52 Comparison of the Models There are six levels of musicalmodel illustrated in Figure 4 (a) themodel from the results ofthe psychological experiments (b) themodel of the perceivedemotion (c) the model of the felt emotion (d) the modelfrom the physiological results (e) the proposed physiolog-ical valencearousal model while listening to the musical

rhythms and (f) the proposed physiological valencearousalmodel after listening to the musical rhythms

The six levels of model need not reach a consensusThereare four reasons Initially the perceived emotion is not alwaysthe same as the felt emotion [10 54] These two emotionsare closer if the music clips are chosen by the participantsthemselves instead of others Secondly somemeasured phys-iological responses might not have been reported by self-report [55] Thirdly if the music clips are simplified to theform of rhythm or tempo the responses are changed [112]Finally the responses while and after listening to music neednot be the same obviously

Figure 4(a) shows that 119879119901 and 119864119892 dominate arousaland 119862119901 dominates valence in a survey of psychologicalexperiments (pp 383ndash392) [1] The effects of 119879119901 and 119864119892

on arousal are clear As for valence the regular and variedrhythms derive positive emotions and the irregular rhythmsderive negative emotions (pp 383ndash392) [1] Hence valence isnegatively correlated to 119862119901

Figure 4(b) shows that 119879119901 [26 31 45] and 119864119892 [26 2731 45] dominate arousal and 119879119901 [27] and 119862119901 [27] dominatevalence in the perceived models The results of arousal forthe perceived models are the same as the psychologicalexperiments But 119879119901 and 119862119901 are viewed as the valence-basedfeatures Whether they are positively correlated to valence ornot has not been highlighted in the paper [27]

Figure 4(c) shows that 119879119901 [35] and 119864119892 [35] dominatearousal and 119879119901 [35] dominates valence in the felt modelThe results of felt and perceived models are similar But thevalence-based feature is limited at 119879119901 in the felt model [35]

Figure 4(d) shows that 119879119901 [55 113 114] and 119864119892 [114]dominate arousal and 119879119901 [55] dominates valence in the phys-iological experiments Both psychological studies (Figures4(a) 4(b) and 4(c)) and physiological studies (Figure 4(d))have the same results about arousal But the faster tempi arefound to decrease SDNN the physiological valence [55]

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

Advances in Electrical Engineering 7

Valence

0

5

10

15

20

1 2 3 4 5 6 7 8 9(ms)

RhythmSDNN

minus5

minus10

minus15

minus20

minus25

(a)

Arousal

0

05

1

15

2

1 2 3 4 5 6 7 8 9

RhythmLFHF

minus2

minus15

minus1

minus05

(b)

Figure 3 (a) The gray bars represent the averages of SDNN and the white bars represent the approached values of the model Items 1 to 4represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to 4 in C2 (b) gray bars represent the averages of LFHF and whitebars represent the approached values of the model Items 1 to 4 represent drum loops 1 to 4 in C1 and items 6 to 9 represent drum loops 1 to4 in C2 (note related numeric data is presented in Table 2)

People prefer faster tempiHowever the sound should notbe too loud if we hope the effect can remain after the musicstops

SDNN (C2) prop119879119901

119864119892

(11)

If fast and loud music stops people feel relaxed

LFHF

(C2) prop minus119879119901 times 119864119892 (12)

Table 3 is the statistical resultsThe SDNN (C2) of the fourrhythms had a significance value 001 The LFHF of L3 had asignificance value 003 while and after listening to music

5 Discussion

51 Advantages of the Proposed Model There are four majoradvantages in this work Initially the psychological musicmodels are many [2 25ndash45] but the physiological musicmodels are rare [3] Secondly the psychologicalmusicmodels[2 25ndash45] and physiological musical experiments [1 111] aremany but studies of the psychological rhythmmodel [26] andthe physiological rhythmic experiment [112] are rareThirdlyalmost all studies concern the responses during themusic butdiscussions about the responses after the music are limited[113] Finally there aremanymusic studies aboutHRV [1 111]but there is no model Our work is the first one

52 Comparison of the Models There are six levels of musicalmodel illustrated in Figure 4 (a) themodel from the results ofthe psychological experiments (b) themodel of the perceivedemotion (c) the model of the felt emotion (d) the modelfrom the physiological results (e) the proposed physiolog-ical valencearousal model while listening to the musical

rhythms and (f) the proposed physiological valencearousalmodel after listening to the musical rhythms

The six levels of model need not reach a consensusThereare four reasons Initially the perceived emotion is not alwaysthe same as the felt emotion [10 54] These two emotionsare closer if the music clips are chosen by the participantsthemselves instead of others Secondly somemeasured phys-iological responses might not have been reported by self-report [55] Thirdly if the music clips are simplified to theform of rhythm or tempo the responses are changed [112]Finally the responses while and after listening to music neednot be the same obviously

Figure 4(a) shows that 119879119901 and 119864119892 dominate arousaland 119862119901 dominates valence in a survey of psychologicalexperiments (pp 383ndash392) [1] The effects of 119879119901 and 119864119892

on arousal are clear As for valence the regular and variedrhythms derive positive emotions and the irregular rhythmsderive negative emotions (pp 383ndash392) [1] Hence valence isnegatively correlated to 119862119901

Figure 4(b) shows that 119879119901 [26 31 45] and 119864119892 [26 2731 45] dominate arousal and 119879119901 [27] and 119862119901 [27] dominatevalence in the perceived models The results of arousal forthe perceived models are the same as the psychologicalexperiments But 119879119901 and 119862119901 are viewed as the valence-basedfeatures Whether they are positively correlated to valence ornot has not been highlighted in the paper [27]

Figure 4(c) shows that 119879119901 [35] and 119864119892 [35] dominatearousal and 119879119901 [35] dominates valence in the felt modelThe results of felt and perceived models are similar But thevalence-based feature is limited at 119879119901 in the felt model [35]

Figure 4(d) shows that 119879119901 [55 113 114] and 119864119892 [114]dominate arousal and 119879119901 [55] dominates valence in the phys-iological experiments Both psychological studies (Figures4(a) 4(b) and 4(c)) and physiological studies (Figure 4(d))have the same results about arousal But the faster tempi arefound to decrease SDNN the physiological valence [55]

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

8 Advances in Electrical Engineering

Table 3 Statistical P values for epochs and rhythms

(a) SDNN

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 100 100 100 100 100 057E2 040 040 063 100 100 100 015E3 100 100 100 096 100 100 047E4 100 100 100 100 100 100 077C1 064 100 100 100 100 055 026C2 014 014 014 100 100 014 001 lowastlowast

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 075 075 094 009 075 007 002 lowast 034L2 002 019 100 100 027 100 007 012L3 100 100 100 100 100 100 060 041L4 100 100 100 100 053 100 040 040

(b) LFHF

Epoch PairwiseL1-L2 L2-L3 L3-L4 L1ndashL3 L2ndashL4 L1ndashL4 L1simL4

E1 100 009 036 097 100 100 041E2 100 100 100 100 100 100 077E3 100 100 100 100 100 100 078E4 100 100 100 100 100 100 044C1 100 100 100 100 100 100 080C2 100 066 100 100 100 100 054

Rhythm PairwiseE1-E2 E2-E3 E3-E4 E1ndashE3 E2ndashE4 E1ndashE4 E1simE4 C1-C2

L1 100 100 100 100 100 100 087 090L2 027 074 007 100 100 027 009 055L3 042 100 045 082 014 100 012 003 lowast

L4 100 100 092 100 100 092 023 050The values of HRV measures in C1 are the comparisons (differences) of values of HRV measures in epochs E1 and E3The values of HRV measures in C2 are the comparisons (differences) of values of HRV measures in epochs E2 and E4L1-L2 drum loops 1 and 2 L1simL4 drum loops 1 to 4Significant codes 0 ldquolowastlowastlowastrdquo 0001 ldquolowastlowastrdquo 001 ldquolowastrdquo 005 ldquordquo 01 ldquo rdquo 1

Figure 4(e) shows that 119864119892 and 119862119901 dominate arousal and119879119901 dominates valence 119864119892 is directly proportional to arousaland 119862119901 is inversely proportional to arousal As for valencefaster tempi increase SDNN the physiological valence Thisrelation is also illustrated in the left part of Figure 3(a)

Figure 4(f) shows that 119879119901 and 119864119892 dominate both arousaland valence after the stimuli The arousal is negativelycorrelated to 119879119901 times 119864119892 Hence if a fast and loud rhythm isremoved the subjects will feel relaxed Fast tempi increasevalence in Figure 4(e) the responses duringmusicThis effectremains after the music stops if the intensity of the music islow

In the valence perspective 119879119901 and 119862119901 are considered astwo major rhythmic parameters [27] As for 119879119901 there existopposite comments 119879119901 is positively correlated with valencein the felt models [35] but 119879119901 is negatively correlated withphysiological valence since the faster tempo decreased SDNN

[55]The proper conclusion is that the role of119879119901 is dependenton the context both slow and fast tempi can derive differentemotions (pp 383ndash392) [1] For example both happy andangry music have fast tempi As for 119862119901 its contributionis not defined in the perceived model [27] Although 119862119901

is negatively correlated with valence in the psychologicalexperiments (pp 383ndash392) [1] the same definition has oppo-site results the firm rhythm derives positive valence (pp383ndash392) [1] and the firm rhythm derives negative valence[31] Instead of music if only the rhythm is consideredhigher SDNN is observed with faster tempi in our work asFigure 4(e) illustrated and119864119892 affects the responses of physio-logical valence after themusic stops as Figure 4(f) illustrated

In the arousal perspective 119879119901 and 119864119892 are two of the mostdominant features Our findings also show that if a fast loudsound (drum loop 3) was removed LFHF would decreaseas illustrated in the right part of Figure 3(b) Although 119879119901

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

Advances in Electrical Engineering 9

Tp Eg

minusCp

(a)

Tp Eg

(Tp Cp)

(b)

Tp Eg

Tp

(c)

Tp Eg

minusTp

(d)

EgCp

minus1Tp

(e)

TpEg

minusTp times Eg

(f)

Figure 4 Six levels of the relationships between the rhythmic features and the valencearousal model the horizontal axis is valence and thevertical axis is arousal (a) Model from the results of the psychological experiments (b) model of the perceived emotion (c) model of the feltemotion (d) model from the results of the physiological experiments (e) proposed physiological valencearousal model while listening tothe musical rhythm and (f) proposed physiological valencearousal model after listening to the musical rhythm

is usually considered as the most important factor in allfeatures (pp 383ndash392) [1] another research indicates 119864119892 ismore important than 119879119901 in arousal [45] our findings supportthis opinion as illustrated in Figure 4(e) 119862119901 is consideredas a valence parameter in Figures 4(a) and 4(b) But ourresults show that simple rhythms enhanced the physiologicalarousal

53 Statistical Results Table 3 collects all 119875 values and thereare two significant results the ANOVA of SDNN (C2) forL1simL4 and the 119905-test of LFHF (C1 C2) for L3 Although thesignificant value is more than 005 the left part of Figure 3(a)still shows that SDNN the physiological valence is positivelycorrelated with the tempi of the drum loops The right partof Figure 3(a) shows that SDNN is positively correlated with

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

10 Advances in Electrical Engineering

119879119901119864119892 (119875 lt 001) Both L2 and L4 have faster tempi andsmaller energies The left part of Figure 3(b) shows thatLFHF the physiological arousal is positively correlated with119864119892119862119901 L3 a fast drum loop with a very simple complexityenhances arousal in the largest degree The right part ofFigure 3(b) shows that after the music stops the fast and loudrhythms make the participants relaxed Although there is nostatistical significance in Figure 3(b) the 119905-test result of L3while and after listening to the music clip reaches statisticalsignificance (119875 lt 005)

54 Limitations

541 The Three Rhythmic Parameters Of the three elemen-tary rhythmic features tempo is the most definite [92] 119879119901 ispositively correlated with SDNN the physiological valence(Figure 4(e)) In the other way 119879119901 is negatively correlatedwith SDNN [55] To integrate these two opposite findingsthe psychological perspectives (Figures 4(a) 4(b) and 4(c))are more suitable to explain this dilemma That is 119879119901 is afeature of arousal both positive and negative valence can bemeasured it depends on the context (pp 383ndash392) [1]

To measure the complexity of a music or rhythm clipmathematical quantities are better than adjectives For exam-ple the firm rhythms can derive both positive (pp 383ndash392) [1] and negative [31] valence We can acquire a numberfrom Likert Scale [93] however it is not proper as anengineering reference Of the mathematical studies aboutrhythmic complexity [115ndash117] oddity [117] is the closestmethod for measuring the perceptual complexity Thereis a more fundamental issue two rhythms with differentresponses may have the same complexity measure

The last parameter the total energy of a waveform isdefined as Σ1198962 the summation of square of 119896 where 119896 is theamplitude of the signal [94 95] But the intensity within thewaveform may not be fixed large or small variations suggestfear or happiness the rapid or few changes in intensity maybe associated with pleading or sadness (pp 383ndash392) [1] Ifthe waveform is partitioned into small segments [118] theenergy feature will be not only an arousal but also a valenceparameter (pp 383ndash392) [1] And the analytic resolution ofthe musical emotion will increase

542 Nonlinearity of the Responses The linear and inversefunctions are simple intuitive and useful in a generaloverview However some studies assume that the liking (orvalence) of music is an inverted-U curve dependent onarousal (tempo and energy) [119ndash121] and complexity [119120] The inverted-U phenomenon may disappear with theprofessional musical expertise [122] There is also a plausibletransform of the inverted-U curve although 120 bpm maybe a preferred tempo for common people [123] the twin-peak curve was found in both a popular music database andphysiological experiments [124] Finally the more complexcurves are possible [3]

55 The Modified Model 119879119901 and 119864119892 correlate with arousalas Figures 4(a)ndash4(d) and Figure 4(f) illustrated but 119864119892 dom-inates the arousal [45] as Figure 4(e) showed To integrate

these models let 1198641015840119892be the average energy of all beats Thus

119864119892 is the product of 119879119901 and 1198641015840

119892 and (9)ndash(12) can be modified

as (13)ndash(16) This provides us with an advanced aspectThe valence of musical responses keeps the same

SDNN (C1) prop minus1

119879119901

(13)

Both tempo and energy are arousal parameters nowMoreover our model demonstrates that the complexity ofrhythm is also an arousal parameter

LFHF

(C1) prop 119879119901 times

1198641015840

119892

119862119901

(14)

The valence after the musical stimuli is influenced by theenergy

SDNN (C2) prop 1

1198641015840119892

(15)

Finally both tempo and energy influence the arousal aftermusical stimuli but tempo dominates the effect

LFHF

(C2) prop minus1198792

119901times 1198641015840

119892 (16)

To integrate all psychological and physiological musictheories we summarize briefly that tempo energy andcomplexity are both valence and arousal parameters Theinverted-U curve [119ndash121] is suggested for valence Con-sidering the valence after music people prefer the music oflow intensity For the responses after the musical stimuli thearousal correlates negatively with tempo and energy whereastempo is the dominant parameter

56 Criteria for Good Models There are four criteria for avaluable model [22] They are generalization (goodness-of-fit [21]) and predictive power simplicity and its relation toexisting theories In plain words a model should be able toexplain experimental data used in the model and not used inthe model for verification The parameters should be smallrelative to the amount of the data The model should beable to unify two formerly unrelated theories It will help tounderstand the underlying phenomenon instead of an input-output routine

Our proposed model satisfies generalization simplicityand the relation to other theories In our study the selectedmodel has the least metric (error) among all possible modelsIt has only three parameters (119879119901 119862119901 and 119864119892) And it fills thegap among the psychological and physiological experimentsand the models of music and rhythm The proposed modeldoes not have enough predictive power It was built forthe subjects whose responses are able to be modeled thenonpredictable class one-third of all subjects was excludedThere was no test data in this study However the modelderived from the experimental data can be integrated wellwith the literature in this field as Section 55 mentioned

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

Advances in Electrical Engineering 11

6 Conclusion and Future Work

A novel experiment a novel algorithm and a novel modelhave been proposed in this study The experiment exploredhow rhythm the fundamental feature of music influencedthe ANS and HRV And the relationships between musi-cal features and physiological features were derived fromthe algorithm Moreover the model demonstrated howthe rhythmic features were mapped to the physiologicalvalencearousal plane

The equations in our model could be the rules for musicgenerating systems [125ndash127] The model could also be fine-tuned if the rhythmic oddity [117] (for complexity) smallsegments [118] (for energy) and various curves [3 119ndash121124] are considered

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

The authors thank Mr Shih-Hsiang Lin heartily for hissupport in both program design and data analysis Theauthors also thank Professor Shao-Yi Chien and Dr Ming-Yie Jan for related discussion

References

[1] P N Juslin and J A Sloboda Eds Handbook of Music andEmotion Theory Research Applications Oxford UniversityPress 2010

[2] Y H Yang and H H Chen Music Emotion Recognition CRCPress 2011

[3] P Gomez and B Danuser ldquoRelationships between musicalstructure and psychophysiological measures of emotionrdquo Emo-tion vol 7 no 2 pp 377ndash387 2007

[4] G Cervellin and G Lippi ldquoFrom music-beat to heart-beat ajourney in the complex interactions between music brain andheartrdquo European Journal of Internal Medicine vol 22 no 4 pp371ndash374 2011

[5] S-H Lin Y-C Huang C-Y Chien L-C Chou S-C HuangandM-Y Jan ldquoA study of the relationship between twomusicalrhythm characteristics and heart rate variability (HRV)rdquo inProceedings of the IEEE International Conference on BioMedicalEngineering and Informatics pp 344ndash347 2008

[6] H-M Wang S-H Lin Y-C Huang et al ldquoA computationalmodel of the relationship between musical rhythm and heartrhythmrdquo in Proceeding of the IEEE International Symposium onCircuits and Systems (ISCAS 09) pp 3102ndash3105 Taipei TaiwanMay 2009

[7] H-M Wang Y-C Lee B S Yen C-Y Wang S-C Huangand K-T Tang ldquoA physiological valencearousal model frommusical rhythm to heart rhythmrdquo in Proceedings of the IEEEInternational Symposium of Circuits and Systems (ISCAS rsquo11) pp1013ndash1016 Rio de Janeiro Brazil May 2011

[8] C L Krumhansl ldquoMusic a link between cognition and emo-tionrdquo Current Directions in Psychological Science vol 11 no 2pp 45ndash50 2002

[9] M Pearce and M Rohrmeier ldquoMusic cognition and the cogni-tive sciencesrdquo Topics in Cognitive Science vol 4 no 4 pp 468ndash484 2012

[10] P Evans and E Schubert ldquoRelationships between expressed andfelt emotions in musicrdquoMusicae Scientiae vol 12 no 1 pp 75ndash99 2008

[11] V J Konecni ldquoDoes music induce emotion A theoretical andmethodological analysisrdquo Psychology of Aesthetics Creativityand the Arts vol 2 no 2 pp 115ndash129 2008

[12] M Zentner D Grandjean and K R Scherer ldquoEmotions evokedby the sound of music characterization classification andmeasurementrdquo Emotion vol 8 no 4 pp 494ndash521 2008

[13] J A Russell ldquoA circumplex model of affectrdquo Journal of Person-ality and Social Psychology vol 39 no 6 pp 1161ndash1178 1980

[14] R E Thayer The Biopsychology of Mood and Arousal OxfordUniversity Press on Demand 1989

[15] D Liu L Lu and H-J Zhang ldquoAutomatic mood detectionfrom acoustic music datardquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 81ndash87 2003

[16] Y E Kim E M Schmidt R Migneco et al ldquoMusic emotionrecognition a state of the art reviewrdquo in Proceedings of theInternational Symposium on Music Information Retrieval pp255ndash266 2010

[17] Y-H Yang and H H Chen ldquoMachine recognition of musicemotion a reviewrdquoACMTransactions on Intelligent Systems andTechnology vol 3 no 3 article 40 2012

[18] K Trohidis G Tsoumakas G Kalliris and I Vlahavas ldquoMulti-label classification of music into emotionsrdquo in Proceedings ofthe International SymposiumonMusic InformationRetrieval pp325ndash330 September 2008

[19] M A Rohrmeier and S Koelsch ldquoPredictive informationprocessing in music cognition A critical reviewrdquo InternationalJournal of Psychophysiology vol 83 no 2 pp 164ndash175 2012

[20] GWidmer andWGoebl ldquoComputationalmodels of expressivemusic performance the state of the artrdquo Journal of New MusicResearch vol 33 pp 203ndash216 2004

[21] H Honing ldquoComputational modeling of music cognition acase study on model selectionrdquoMusic Perception vol 23 no 5pp 365ndash376 2006

[22] H Purwins P Herrera M Grachten A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition I the perceptual and cognitive processing chainrdquoPhysics of Life Reviews vol 5 no 3 pp 151ndash168 2008

[23] H Purwins M Grachten P Herrera A Hazan R Marxerand X Serra ldquoComputational models of music perception andcognition II domain-specific music processingrdquo Physics of LifeReviews vol 5 no 3 pp 169ndash182 2008

[24] T Eerola ldquoModeling listenersrsquo emotional response to musicrdquoTopics in Cognitive Science vol 4 no 4 pp 607ndash624 2012

[25] J-C Wang Y-H Yang H-M Wang and S-K Jeng ldquoTheacoustic emotion gaussians model for emotion-based musicannotation and retrievalrdquo in Proceedings of the 20th ACMInternational Conference on Multimedia (MM rsquo12) pp 89ndash98November 2012

[26] J Cu R Cabredo R Legaspi and M T Suarez ldquoOn modellingemotional responses to rhythm featuresrdquo in Proceedings ofthe Trends in Artificial Intelligence (PRICAI rsquo12) pp 857ndash860Springer 2012

[27] J J Deng and C Leung ldquoMusic emotion retrieval based onacoustic featuresrdquo in Advances in Electric and Electronics vol155 of Lecture Notes in Electrical Engineering pp 169ndash177 2012

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

12 Advances in Electrical Engineering

[28] H G Avisado J V Cocjin J A Gaverza R Cabredo JCu and M Suarez ldquoAnalysis of music timbre features forthe construction of user-specific affect modelrdquo in Theory andPractice of Computation pp 28ndash35 Springer Berlin Germany2012

[29] M M Farbood ldquoA parametric temporal model of musicaltensionrdquoMusic Perception vol 29 no 4 pp 387ndash428 2012

[30] J Albrecht ldquoA model of perceived musical affect accuratelypredicts self-report ratingsrdquo in Proceedings of the InternationalConference on Music Perception pp 35ndash43 2012

[31] Y-H Yang and H H Chen ldquoPrediction of the distributionof perceived music emotions using discrete samplesrdquo IEEETransactions on Audio Speech and Language Processing vol 19pp 2184ndash2196 2011

[32] R Y Granot and Z Eitan ldquoMusical tension and the interactionof dynamic auditory parametersrdquoMusic Perception vol 28 no3 pp 219ndash246 2011

[33] E M Schmidt and Y E Kim ldquoModeling musical emotiondynamics with conditional random fieldsrdquo in Proceedings ofthe 12th International Society for Music Information RetrievalConference (ISMIR rsquo11) pp 777ndash782 October 2011

[34] B Schuller J Dorfner and G Rigoll ldquoDetermination of non-prototypical valence and arousal in popular music features andperformancesrdquo EURASIP Journal on Audio Speech and MusicProcessing vol 2010 Article ID 735854 2010

[35] E Coutinho and A Cangelosi ldquoThe use of spatio-temporalconnectionist models in psychological studies of musical emo-tionsrdquoMusic Perception vol 27 no 1 pp 1ndash15 2009

[36] T Eerola O Lartillot and P Toiviainen ldquoPrediction of multi-dimensional emotional ratings in music from audio using mul-tivariate regression modelsrdquo in Proceedings of the InternationalSymposium on Music Information Retrieval pp 621ndash626 2009

[37] T-L Wu and S-K Jeng ldquoProbabilistic estimation of a novelmusic emotionmodelrdquo inAdvances inMultimediaModeling pp487ndash497 Springer 2008

[38] G Luck P Toiviainen J Erkkila et al ldquoModelling the relation-ships between emotional responses to and musical content ofmusic therapy improvisationsrdquo Psychology of Music vol 36 no1 pp 25ndash45 2008

[39] Y-H Yang Y-C Lin Y-F Su and H H Chen ldquoA regressionapproach to music emotion recognitionrdquo IEEE Transactions onAudio Speech and Language Processing vol 16 no 2 pp 448ndash457 2008

[40] F Lerdahl and C L Krumhansl ldquoModeling tonal tensionrdquoMusic Perception vol 24 no 4 pp 329ndash366 2007

[41] Y-H Yang C-C Liu and H H Chen ldquoMusic emotionclassification a fuzzy approachrdquo in Proceeding of the AnnualACM International Conference onMultimedia (MM 06) pp 81ndash84 New York NY USA October 2006

[42] M D Korhonen D A Clausi and M E Jernigan ldquoModelingemotional content of music using system identificationrdquo IEEETransactions on Systems Man and Cybernetics B Cyberneticsvol 36 no 3 pp 588ndash599 2006

[43] M Leman V Vermeulen L de Voogdt D Moelants and MLesaffre ldquoPrediction of musical affect using a combination ofacoustic structural cuesrdquo Journal of NewMusic Research vol 34no 1 pp 39ndash67 2005

[44] E H Margulis ldquoA model of melodic expectationrdquo MusicPerception vol 22 no 4 pp 663ndash714 2005

[45] E Schubert ldquoModeling perceived emotion with continuousmusical featuresrdquoMusic Perception vol 21 pp 561ndash585 2004

[46] K R Scherer ldquoWhich emotions can be induced bymusicWhatare the underlying mechanisms And how can we measurethemrdquo Journal of New Music Research vol 33 pp 239ndash2512004

[47] P N Juslin and D Vastfjall ldquoEmotional responses to music theneed to consider underlyingmechanismsrdquoBehavioral andBrainSciences vol 31 no 5 pp 559ndash621 2008

[48] L-O Lundqvist F Carlsson P Hilmersson and P N JuslinldquoEmotional responses to music experience expression andphysiologyrdquo Psychology of Music vol 37 no 1 pp 61ndash90 2009

[49] N S Rickard ldquoIntense emotional responses to music a test ofthe physiological arousal hypothesisrdquo Psychology of Music vol32 no 4 pp 371ndash388 2004

[50] J Kim and E Andre ldquoEmotion recognition based on physiolog-ical changes in music listeningrdquo IEEE Transactions on PatternAnalysis andMachine Intelligence vol 30 no 12 pp 2067ndash20832008

[51] E Coutinho and A Cangelosi Computational and Psycho-Physiological Investigations of Musical Emotions University ofPlymouth 2008

[52] E Coutinho and A Cangelosi ldquoMusical emotions predictingsecond-by-second subjective feelings of emotion from low-level psychoacoustic features and physiological measurementsrdquoEmotion vol 11 no 4 pp 921ndash937 2011

[53] K Trochidis D Sears D L Tran and S McAdams ldquoPsy-chophysiological measures of emotional response to Romanticorchestral music and their musical and acoustic correlatesrdquo inProceedings of the International Symposium on Computer MusicModelling and Retrieval pp 45ndash52 2012

[54] V N Salimpoor M Benovoy G Longo J R Cooperstock andR J Zatorre ldquoThe rewarding aspects of music listening arerelated to degree of emotional arousalrdquo PLoS ONE vol 4 no10 Article ID e7487 2009

[55] M D van der Zwaag J H D M Westerink and E L vanden Broek ldquoEmotional and psychophysiological responses totempomode and percussivenessrdquoMusicae Scientiae vol 15 no2 pp 250ndash269 2011

[56] J R J Fontaine K R Scherer E B Roesch and P C EllsworthldquoThe world of emotions is not two-dimensionalrdquo PsychologicalScience vol 18 no 12 pp 1050ndash1057 2007

[57] T Eerola and J K Vuoskoski ldquoA comparison of the discrete anddimensional models of emotion in musicrdquo Psychology of Musicvol 39 no 1 pp 18ndash49 2011

[58] U R Acharya K P Joseph N Kannathal C M Lim and JS Suri ldquoHeart rate variability a reviewrdquoMedical and BiologicalEngineering and Computing vol 44 no 12 pp 1031ndash1051 2006

[59] ldquoHeart rate variability standards ofmeasurement physiologicalinterpretation and clinical use Task Force of the EuropeanSociety of Cardiology and the North American Society ofPacing and Electrophysiologyrdquo Circulation vol 93 no 5 pp1043ndash1065 1996

[60] R D Lane K McRae E M Reiman K Chen G L Ahern andJ F Thayer ldquoNeural correlates of heart rate variability duringemotionrdquo NeuroImage vol 44 no 1 pp 213ndash222 2009

[61] P Rainville A Bechara N Naqvi and A R Damasio ldquoBasicemotions are associated with distinct patterns of cardiorespira-tory activityrdquo International Journal of Psychophysiology vol 61no 1 pp 5ndash18 2006

[62] B M Appelhans and L J Luecken ldquoHeart rate variability asan index of regulated emotional respondingrdquo Review of GeneralPsychology vol 10 no 3 pp 229ndash240 2006

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

Advances in Electrical Engineering 13

[63] G H E Gendolla and J Krusken ldquoMood state task demandand effort-related cardiovascular responserdquoCognition and Emo-tion vol 16 no 5 pp 577ndash603 2002

[64] R McCraty M Atkinson W A Tiller G Rein and A DWatkins ldquoThe effects of emotions on short-term power spec-trum analysis of heart rate variabilityrdquoThe American Journal ofCardiology vol 76 no 14 pp 1089ndash1093 1995

[65] A H Kemp D S Quintana and M A Gray ldquoIs heartrate variability reduced in depression without cardiovasculardiseaserdquo Biological Psychiatry vol 69 no 4 pp e3ndashe4 2011

[66] C M M Licht E J C De Geus F G Zitman W J GHoogendijk R Van Dyck and BW J H Penninx ldquoAssociationbetween major depressive disorder and heart rate variabilityin the Netherlands study of depression and anxiety (NESDA)rdquoArchives of General Psychiatry vol 65 no 12 pp 1358ndash13672008

[67] R M Carney R D Saunders K E Freedland P Stein M WRich and A S Jaffe ldquoAssociation of depression with reducedheart rate variability in coronary artery diseaserdquoThe AmericanJournal of Cardiology vol 76 no 8 pp 562ndash564 1995

[68] F Geisler N Vennewald T Kubiak and HWeber ldquoThe impactof heart rate variability on subjective well-being is mediated byemotion regulationrdquo Personality and Individual Differences vol49 no 7 pp 723ndash728 2010

[69] S Koelsch A Remppis D Sammler et al ldquoA cardiac signatureof emotionalityrdquo European Journal of Neuroscience vol 26 no11 pp 3328ndash3338 2007

[70] R Krittayaphong W E Cascio K C Light et al ldquoHeart ratevariability in patients with coronary artery disease differencesin patients with higher and lower depression scoresrdquo Psychoso-matic Medicine vol 59 no 3 pp 231ndash235 1997

[71] M Muller D P W Ellis A Klapuri and G Richard ldquoSignalprocessing for music analysisrdquo IEEE Journal on Selected Topicsin Signal Processing vol 5 no 6 pp 1088ndash1110 2011

[72] K Jensen ldquoMultiple scale music segmentation using rhythmtimbre and harmonyrdquo Eurasip Journal on Advances in SignalProcessing vol 2007 Article ID 73205 2007

[73] N Scaringella G Zoia and D Mlynek ldquoAutomatic genreclassification of music content a surveyrdquo IEEE Signal ProcessingMagazine vol 23 no 2 pp 133ndash141 2006

[74] J-J Aucouturier and F Pachet ldquoRepresenting musical genre astate of the artrdquo Journal of New Music Research vol 32 pp 83ndash93 2003

[75] T Li and M Ogihara ldquoDetecting emotion in musicrdquo in Pro-ceedings of the International Symposium on Music InformationRetrieval pp 239ndash240 2003

[76] G Tzanetakis and P Cook ldquoMusical genre classification ofaudio signalsrdquo IEEE Transactions on Speech and Audio Process-ing vol 10 no 5 pp 293ndash302 2002

[77] L Jaquet B Danuser and P Gomez ldquoMusic and felt emotionshow systematic pitch level variations affect the experience ofpleasantness and arousalrdquo Psychology of Music 2012

[78] J C Hailstone R Omar S M D Henley C Frost M GKenward and J D Warren ldquoItrsquos not what you play itrsquos howyou play it timbre affects perception of emotion in musicrdquoTheQuarterly Journal of Experimental Psychology vol 62 no 11 pp2141ndash2155 2009

[79] L Trainor ldquoScience ampmusic the neural roots of musicrdquoNaturevol 453 no 7195 pp 598ndash599 2008

[80] J Bispham ldquoRhythm in Music what is it Who has it AndwhyrdquoMusic Perception vol 24 no 2 pp 125ndash134 2006

[81] J A Grahn ldquoNeuroscientific investigations of musical rhythmrecent advances and future challengesrdquo Contemporary MusicReview vol 28 no 3 pp 251ndash277 2009

[82] K Overy and R Turner ldquoThe rhythmic brainrdquo Cortex vol 45no 1 pp 1ndash3 2009

[83] J S Snyder EW Large andV Penhune ldquoRhythms in the brainbasic science and clinical perspectivesrdquo Annals of the New YorkAcademy of Sciences vol 1169 pp 13ndash14 2009

[84] H Honing ldquoWithout it no music beat induction as a fun-damental musical traitrdquo Annals of the New York Academy ofSciences vol 1252 no 1 pp 85ndash91 2012

[85] J A Grahn ldquoNeuralmechanisms of rhythmperception currentfindings and future perspectivesrdquo Topics in Cognitive Sciencevol 4 no 4 pp 585ndash606 2012

[86] I Winkler G P Haden O Ladinig I Sziller and H HoningldquoNewborn infants detect the beat in musicrdquo Proceedings of theNational Academy of Sciences vol 106 pp 2468ndash2471 2009

[87] M Zentner andT Eerola ldquoRhythmic engagementwithmusic ininfancyrdquo Proceedings of the National Academy of Sciences of theUnited States of America vol 107 no 13 pp 5768ndash5773 2010

[88] S E Trehub and E E Hannon ldquoConventional rhythms enhanceinfantsrsquo and adultsrsquo perception of musical patternsrdquo Cortex vol45 no 1 pp 110ndash118 2009

[89] E Geiser E Ziegler L Jancke and M Meyer ldquoEarly elec-trophysiological correlates of meter and rhythm processing inmusic perceptionrdquo Cortex vol 45 no 1 pp 93ndash102 2009

[90] C-H Yeh H-H Lin and H-T Chang ldquoAn efficient emotiondetection scheme for popular musicrdquo in Proceedings of the IEEEInternational Symposium on Circuits and Systems (ISCAS 09)pp 1799ndash1802 Taipei Taiwan May 2009

[91] L Lu D Liu and H J Zhang ldquoAutomatic mood detection andtracking of music audio signalsrdquo IEEE Transactions on AudioSpeech and Language Processing vol 14 no 1 pp 5ndash18 2006

[92] S Dixon ldquoAutomatic extraction of tempo and beat fromexpressive performancesrdquo Journal of New Music Research vol30 pp 39ndash58 2001

[93] R Likert ldquoA technique for the measurement of attitudesrdquoArchives of Psychology vol 22 no 140 pp 1ndash55 1932

[94] W A Sethares R D Morris and J C Sethares ldquoBeat trackingof musical performances using low-level audio featuresrdquo IEEETransactions on Speech and Audio Processing vol 13 no 2 pp275ndash285 2005

[95] J P Bello C Duxbury M Davies and M Sandler ldquoOn the useof phase and energy for musical onset detection in the complexdomainrdquo IEEE Signal Processing Letters vol 11 no 6 pp 553ndash556 2004

[96] WWei C Zhang andW Lin ldquoAQRSwave detection algorithmbased on complex wavelet transformrdquo Applied Mechanics andMaterials vol 239-240 pp 1284ndash1288 2013

[97] B U Kohler C Hennig and R Orglmeister ldquoThe principlesof software QRS detectionrdquo IEEE Engineering in Medicine andBiology Magazine vol 21 no 1 pp 42ndash57 2002

[98] J Pan and W J Tompkins ldquoA real-time QRS detection algo-rithmrdquo IEEE Transactions on Biomedical Engineering vol 32no 3 pp 230ndash236 1985

[99] J Mateo and P Laguna ldquoAnalysis of heart rate variability in thepresence of ectopic beats using the heart timing signalrdquo IEEETransactions on Biomedical Engineering vol 50 no 3 pp 334ndash343 2003

[100] S McKinley and M Levine ldquoCubic spline interpolationrdquohttponlineredwoodseduinstructdarnoldlaprojfall98sky-megprojpdf

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

14 Advances in Electrical Engineering

[101] M P Tarvainen P O Ranta-aho and P A KarjalainenldquoAn advanced detrending method with application to HRVanalysisrdquo IEEE Transactions on Biomedical Engineering vol 49no 2 pp 172ndash175 2002

[102] P Welch ldquoThe use of fast Fourier transform for the estimationof power spectra a method based on time averaging overshortmodified periodogramsrdquo IEEETransactions onAudio andElectroacoustics vol 15 pp 70ndash73 1967

[103] TWarren Liao ldquoClustering of time series data a surveyrdquoPatternRecognition vol 38 no 11 pp 1857ndash1874 2005

[104] V Barnett and T Lewis Outliers in Statistical Data vol 1 JohnWiley amp Sons 1984

[105] C S Davis Statistical Methods for the Analysis of RepeatedMeasurements Springer New York NY USA 2002

[106] R Kabacoff R in Action Manning Publications 2011[107] L PaceBeginning R An Introduction to Statistical Programming

Apress 2012[108] J Albert and M Rizzo R by Example Springer 2012[109] M Gardener Beginning R The Statistical Programming Lan-

guage Wrox 2012[110] J Adler and J Beyer R in a Nutshell OrsquoReilly Frankfurt

Germany 2010[111] S D Kreibig ldquoAutonomic nervous system activity in emotion

a reviewrdquo Biological Psychology vol 84 no 3 pp 394ndash421 2010[112] S Khalfa M Roy P Rainville S Dalla Bella and I Peretz ldquoRole

of tempo entrainment in psychophysiological differentiation ofhappy and sad musicrdquo International Journal of Psychophysiol-ogy vol 68 no 1 pp 17ndash26 2008

[113] L Bernardi C Porta and P Sleight ldquoCardiovascular cere-brovascular and respiratory changes induced by different typesof music in musicians and non-musicians the importance ofsilencerdquo Heart vol 92 no 4 pp 445ndash452 2006

[114] M Iwanaga A Kobayashi and C Kawasaki ldquoHeart rate vari-ability with repetitive exposure to musicrdquo Biological Psychologyvol 70 no 1 pp 61ndash66 2005

[115] E Thul and G T Toussaint ldquoAnalysis of musical rhythmcomplexity measures in a cultural contextrdquo in Proceedings of theC3S2E Conference (C3S2E rsquo08) pp 7ndash9 May 2008

[116] E Thul and G T Toussaint ldquoOn the relation between rhythmcomplexity measures and human rhythmic performancerdquo inProceeding of the C3S2E Conference (C3S2E rsquo08) pp 199ndash204May 2008

[117] E Thul and G T Toussaint ldquoRhythm complexity measures acomparison of mathematical models of human perception andperformancerdquo in Proceedings of the International Symposium onMusic Information Retrieval pp 14ndash18 2008

[118] Z Xiao E Dellandrea W Dou and L Chen ldquoWhat is the bestsegment duration for music mood analysis rdquo in Proceedingsof the International Workshop on Content-Based MultimediaIndexing (CBMI rsquo08) pp 17ndash24 London UK June 2008

[119] A Lamont and R Webb ldquoShort- and long-term musicalpreferences what makes a favourite piece of musicrdquo Psychologyof Music vol 38 no 2 pp 222ndash241 2010

[120] A C North and D J Hargreaves ldquoLiking arousal potentialand the emotions expressed by musicrdquo Scandinavian Journal ofPsychology vol 38 no 1 pp 45ndash53 1997

[121] S Abdallah and M Plumbley ldquoInformation dynamics patternsof expectation and surprise in the perception ofmusicrdquoConnec-tion Science vol 21 no 2-3 pp 89ndash117 2009

[122] M G Orr and S Ohlsson ldquoRelationship between complexityand liking as a function of expertiserdquoMusic Perception vol 22no 4 pp 583ndash611 2005

[123] D Moelants ldquoPreferred tempo reconsideredrdquo in Proceedings ofthe International Conference onMusic Perception and Cognitionpp 580ndash583 2002

[124] Y-S Shih W-C Zhang Y-C Lee et al ldquoTwin-peak effect inboth cardiac response and tempo of popular musicrdquo in Proceed-ings of the International Conference of the IEEE Engineering inMedicine and Biology Society pp 1705ndash1708 2011

[125] A P Oliveira and A Cardoso ldquoA musical system for emotionalexpressionrdquo Knowledge-Based Systems vol 23 no 8 pp 901ndash913 2010

[126] I Wallis T Ingalls and E Campana ldquoComputer-generatingemotional music the design of an affective music algorithmrdquo inProceedings of the 11th International Conference on Digital AudioEffects (DAFx rsquo08) pp 7ndash12 September 2008

[127] S R Livingstone R Muhlberger A R Brown and A LochldquoControlling musical emotionality an affective computationalarchitecture for influencing musical emotionsrdquo Digital Creativ-ity vol 18 no 1 pp 43ndash53 2007

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation httpwwwhindawicom

Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

DistributedSensor Networks

International Journal of