Affective Valence and Arousal in ADHD and Normal BoysDuring a Startle Habituation Experiment

EDWARD M. ORNITZ. M.D., PATRIK GABIKIAN. ANDREWT. RUSSELL. M.D.. DONALD GUTHRIE. PH.D..CHIAKI HIRANO, M.D.•AND JEAN-GUIDO GEHRICKE, M.A.

ABSTRACT

ObJective: To measure two dimensions of emotion (affective valence and arousal) in 29 boys with attention-deficit hyper­

activity disorder (ADHD) and 32 normal boys. Method: After a startle habituation experiment during which these subjects

heard 40 startling sounds while watching a silent interesting movie, they were asked 12 questions (categorized a priori

into questions relating to affective valence and to arousal) about their emotional reactions to these putatively unpleasant

and pleasurable stimuli. Responses were recorded for the two dimensions of emotion, using two cartoon strips in each

of which five expressions of a cartoon character varied linearly from happy to unhappy (affective valence dimension) and

calm to excited (arousal dimension) . Results: Factor analyses of the 12 responses revealed four factors in which the

highest loadings were for affective valence to the startle responses, affective valence to the silent movie, arousal, and

scary feelings. Relative to the normal group. the responses of the ADHD group were significantly biased toward pleasura­

ble valence to the startling stimuli and to the silent movie. with a trend toward hypoarousal. Startle magnitude and

habituation were similar in both groups. The normal tonic heart rate acceleration throughout the experimental session

was not sustained in the ADHD group. Conclusions: The self-reports of affective valence biased in the direction of

pleasure and away from displeasure and the trend toward hypoarousal suggest an emotional dysfunction in ADHD. J.

Am. Acad. Child Ado/esc. Psychiatry, 1997, 36(12) :1698-1705. Key Words: attention-deficit hyperactivity disorder,

emotion, heart rate. startle. habituation.

Since the seminal work of Douglas (1972), the behaviorof children who had been described as hyperkinetic,hyperactive, or suffering from "minimal brain dysfunc­tion" has been attributed to an underlying disorder ofthe capacity to sustain attention. With the publicationof DSM-I/I, this conceptualization became embodiedin the nomenclature as attention deficit disorder. In thisand subsequent editions (DSM-I/I-R, DSM-IV) of theDiagnostic and Statistical Manual ofMental Disorders,the major conceptual changes have involved the extent

AcupttdJalluary 30. 1997.Drs. Umitz; Rusull, and (iuthri« arrftculty membersand Mr. Grhric!u is 011

th« rtaff of the Department of Psychiatry. Division of Child and Adolesren:Prychiatry. UCLA, Lor Allgrlrr. Dr. Ornitz is a member ofthr Brain Rrrrarchlnstitut« and Dr. Guthri« is a member ofthr Department ofBiortatirticr, UCLA.Mr. Gabilriall is all ulldrrgraduatt student at lICLA. Dr. Hirano is 011 th« rtaff

ofNagahama Rrd Cross Hospital. Nagahama. [apan.

This illllm igatioll /t'm mpporttd ky NICHHD gr,"JI HD -/4/93 and theAlic« andJulius Kantor Charitabk Trust.

Reprint rrqurm to Dr. Umitz; 27-384 NPI . UCLA. 760 Wrrru'ood Plaza,LorAngrkr. 01 90024.

0890-8S67/97/3612-16<)8/S0.3()()/M' I,)<)7 by the American Academy ofChild and Adolescent Psychiatry,

to which hyperactivity and impulsivity are consideredseparate or integral features of the attentional disorder(Morgan et al., 1996).

None of these conceptualizations of attention-deficithyperactivity disorder (ADHD) have explicitly consid­ered the presence of an emotional dysfunction. How­ever, Douglas' view of the syndrome included, alongwith deficits of attention and impulse control, a strongmotivational dysfunction evidenced by a need for con­tinuous reenforcement (Douglas and Parry, 1983; Parryand Douglas, 1983). These and other studies have led tothe concept of a reward system dysfunction in ADHDmanifest as abnormal response to positive emotionalstimuli, excessive dependence on continuous reward,and elevated reward threshold (Becker er al., 1993;Douglas and Parry, 1983; Haenlein and Caul, 1987).Emotional dysfunction in ADHD is also suggested byRorschach responses indicating an avoidance of affect­laden stimuli (Cotugno, 1995). There is some evidencefor a deficit in the recognition of affect and the use ofclues to emotional information in children with ADHDcharacteristics, depending on age and the modality in

1698 J. AM. ACAD . CHILD ADOl.E SC . PSYCHIATRY. Jr.:12, DECEMBER 1997

which emotional clues are presented (Harding-Roundy,1989; Mackenberg, 1992; Shapiro et aI., 1993; Voeller,1986). Reconceptualizations ofADHD both as a motiva­tion deficit disorder (Barkley, 1990) and a temperrnent­based disorder (Carey, 1992) have been suggested.

Dysfunctions of motivation, temperament, affectrecognition, affect avoidance, and response to rewardssuggest but do not define an emotional component inADHD. To do so requires a definition of the structureof emotion. Utilizing this concept of bidirectionaliry(approach or avoidance) of emotions (Hebb, 1949),augmented by a second dimension, the vigor or intensitywith which approach or avoidance occurs, Lang and hiscolleagues (Lang, 1995; Lang et al., 1990) have definedemotion along the dimensions of affictive valence (thedisposition toward appetitive or defensive behavior, i.e.,the degree to which pleasurable things are approachedor un pleasurable things are avoided) and arousal (theintensity with which such behavior occurs) . Researchhas demonstrated strong correlations between combina­tions of these two dimensions of emotion and ratings ofdiscrete emotions such that negative affective valence(displeasure) combined with high arousal significantlycorrelates with fear and anger while positive affectivevalence (pleasure) combined with low arousal correlateswith feelings of relaxation. Conversely, positive affectivevalence combined with high arousal correlates with joyand excitement while negative affective valencecombined with low arousal correlates with sadness anddepression (Bradley and Lang, 1994; Lang, 1995). Inthis study, the focus is not on discrete emotions butrather on the continuum of affective response frompleasure to displeasure and the intensity with whichsuch response is experienced.

The conceptualization of arousal as a dimensionof emotion may be relevant to ADHD. One recent(Shibagaki et aI., 1993) and some older (Satterfield andDawson, 1971; Satterfield et aI., 1974) psychophysi­ological studies have suggested hypoarousal in childrenwho have ADHD. Some clinical studies have described"a sluggish cognitive tempo" (Lahey et al., 1987) orhypovigilance (Weinberg and Harper, 1993) in sub­groups of children with ADHD, and impaired vigilanceon continuous performance tasks is suggestive of hypo­arousal (Voeller, 1991) .

Lang et aI. (1990) have shown that the startle blinkreflex, as a component of the startle response, a defen­sive or avoidance response to a sudden threat, is aug-

EMOTIONAL RESPONSES TO STARTLE IN ADHD

mented when a startling stimulus is presented while thesubject is in an emotional state characterized by negativeaffective valence, e.g., anger or fear. There is a "match"between the direction of affective valence and thedefensive reflex. Conversely, the startle blink reflex isinhibited when there is a "mismatch," i.e., when thesubject is in an emotional state characterized by positiveaffective valence, e.g., joy or lust (Lang et aI., 1990).This differential effect occurs only when the arousalcomponent of emotion is sufficiently high (Cuthbertet al., 1996; Lang, 1995).

In this study, ADHD and normal boys were exposedto 40 startling stimuli presented in a simple habituationparadigm. Ornitz et aI. (1996) have described a signi­ficant tonic increase in heart rate, suggesting increasingarousal, accompanying habituation of the startle blinkreflex in response to this type of repetitive stimulationin normal boys. Predicated on the assumption that thesesudden loud sounds are inherently aversive, unpleasant,prone to evoke avoidance or fear and are also arousing(Lang et aI., 1990), the children were queried on theiremotional reactions to the startling stimuli. Their atten­tion was drawn to a silent movie, appropriate to school­age boys, presented continuously before and duringexposure to the auditory stimuli, and they were alsoasked to rate their emotional responses to the relativelypleasant movie. Ratings of emotional reactions to boththe startling sounds and the silent movie were facilitatedby asking the children to point to the picture on each oftwo five-panel cartoon strips that best reflected theirfeelings or state of arousal. The cartoon strips are takenfrom the Self-Assessment Manikin (SAM), a pictorialinstrument developed for the measurement of emo­tional responses to pictures, images, and sounds(Bradley and Lang , 1994; Hodes et aI., 1985 ; Lang,1980). SAM is advantageous for use with childrenbecause the five-picture scales are graphic and permitnonverbal responses. Affective valence is depicted onone cartoon strip in which SAM ranges from a smiling,happy to a frowning, unhappy figure; arousal isdepicted on a second strip in which SAM ranges from arelaxed to an excited figure (Bradley and Lang, 1994).

The first purpose of this investigation is to evaluatethe emotional reactions (affective valence and arousal)in boys with ADHD (relative to normal boys) to star­tling stimuli that on face value are noxious and shouldevoke both arousal and negative affect. Second, themagnitude and habituation of the startle responses and

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 36:12, D ECEMBER 1997 1699

ORNITZ ET AL.

the tonic changes in heart rate (indexing arousal) thataccompany startle habituation (Ornitz et al., 1996) arecompared in the ADHD and normal boys.

METHOD

Subjects

Seven- to II -year-old boys were recruited from the generalpopulation. Thirty-two normal boys and 29 boys who met all DSM­Ill-R criteria for ADHD were entered into the study.

Exclusion criteria were a Full Scale IQ less than 85. evidence ofhearing or visual impairment. or any medical. neurological. or psy­chiatric (particularly conduct disorder) disorder. and for normalboys more than 3 of the 14 symptoms of ADHD listed in DSM-lll­R (American Psychiatric Association. 1987). Developmental dis­orders were also exclusion criteria for the normal boys but not forthe ADHD boys. in whom certain comorbid disorders are expected.Hence, nine ADHD boys had primary nocturnal enuresis. two hadsecondary nocturnal enuresis. six met criteria for oppositionaldefiant disorder. and six had specificdevelopmental disorders. Onlyseven ADHD boys had been taking stimulant medication; all wereoff medication for at least 1 week prior to the study. The mean ageof the normal boys was 115.5 months (SD 18.4 months. range 86 to143 months) and that of the ADHD group was 108.2 months (SD17.5 months, range 84 to 137 months). The 61 subjects who wereentered into the current study had been asked the seriesof questionsconcerning their reactions to the startling sounds and the silentmovie (see "Procedures," below). Startle-response habituation datawere available from all 61 subjects. Data on heart rate changesaccompanying startle habituation were available from only 34 ofthese subjects (17 normal and 17 ADHD boys). To obtain sufficientheart rate data, additional cohorts of 23 normal and 22 ADHDboys (who had not been queried about their emotional reactions)from a larger pool of 7- to l l-year-olds were added to these data.creating a separate database of 40 normal and 39 ADHD boys foranalyses of heart rate changes.

Diagnostic Evaluation and Instruments

Each child and at least one parent were interviewed with theDiagnostic Interview for Children and Adolescents (DICA-C andDICA-P) (Welner et al.• 1987) by a child psychiatrist (A.T.R.).TheWISC-R (Wechsler. 1974) was administered to each subject. Aparent of each subject completed the Child Behavior Checklist(Achenbach and Edelbrock, 1983) and the Werry-Weiss-PetersRating Scale (Routh er al., 1974). A teacher completed the Teacher'sReport Form (Achenbach and Edelbrock, 1986) and the ADDHComprehensive Teacher's Rating Scale (ACTeRS) (Ullmann er al.,1985). These reports were available for all but one subject.

Startle Stimuli and Startle Responses

The startle stimuli (SS) were forty 104-dB, 50-millisecond noisebursts presented through earphones every 23 to 25 seconds. Thestartle response was derived from orbicularis oculi electromyogram.Details concerning startle stimulus presentation and startle responserecording. processing, and measurement can be found in the articleby Ornitz et al. (1996).

Heart Rate

The heart rate was derived from the interbeat intervals of theelectrocardiogram during three time periods: a 5-second period

(baseline) preceding each of the 40 SS; a 5-second periodcommencing 0.5 second after each SS (poststimulus 1); and a 5­second period immediately after each poststimulus 1 period (post­stimulus 2).

Procedures

Informed consent was obtained after the nature of the study wasfully explained to both the child and his parent. This included asingle exposure to the SS. presented during an adaptation sessionthat included familiarization with all aspects of the experimentexcept the repetitive presentation of the SS. All subjects passed ahearing test.

Subjects watched silent TV movies of their choice while hearingthe 40 SS. No instruction was given regarding the sounds other thanthat they would be presented repetitively. Immediately after thepresentation of the 40 startling stimuli, each subject was asked sixquestions about his personal reactions to hearing the loud soundsand another six questions about his reactions to the silent movie.Each set of six questions included four questions about affectivereactions to the respective stimuli (loud sounds or silent movie)followed by two questions about arousal to the respective stimuli(Table 1). The subject was asked to indicate his answer to eachspoken question by pointing to one of five pictures in two cartoonstrips, each of which shows five expressions of a cartoon characternamed SAM (Self-Assessment Mannikin) (Hodes et al., 1985). Theexpressions ranged from happy to unhappy for the affectivevalencequestions, and calm to excited for the arousal questions. The rangeof expressions were given the values 1 through 5; 3 indexed neutralreactions. and 1 and 5 indexed the extremes of pleasurable andunpleasurable or least aroused and most aroused reactions,respectively. Both Lang (1980) and Bradleyand Lang (1994) providedata validating the use of SAM in measurement of affectivevalenceand arousal.

Data Analysis

Startle Responses. Ten successive blocks of four trials were averagedand analyzed with a repeated-measures analysis of variance usingblocks as intrasubject factors and diagnostic groups as between-sub­ject factors.

HeartRate Changts. Heart rates were computed as the mean inter­beat interval during each of the 40 baseline and posrstimulus 1 and2 time periods. The 40 sets of baseline and poststimulus heart rateperiods were divided into 10 successive blocks of four trials eachand were analyzed with repeated-measures analysesof variance usingblocks as intrasubject factors and diagnostic groups as between-sub­ject factors. Orthogonal polynomial components were used to studylinear, quadratic, and higher-order rates of change.

Responses to the 12 QutstionsAbout EmotionalValenc« and Arousalto the Stimuli. The 12 responses from each child were subjected to aprincipal-components factor analysis with varimax rotation, usingthe responses from all normal and ADHD boys for validation of thea priori distinctions between affective valence and arousal andbetween reactions to the startling sounds and reactions to the silentmovie. Two-sample t tests compared the derived factor scores of thenormal and ADHD groups. Additional t tests were used to comparethe actual responses of the two groups to each of the 12 questions.and multivariate analyses of variance (Hotelling's T 2) were used toevaluate group differences in the clusters of responses identified bythe factor analysis and in the clusters of responses that had beendefined a priori (as classified in Table 1).

1700 J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY. 36:12. DECEMBER 1997

TABLE 1Twelve Questions Used to Elicit AffectiveValence and Degree of

Arousal to Startling Sounds and Silent Movie

Affective Valence to Startling Sounds

I. Look at this guy. His name is SAM. Please take a moment andthink about the loud sounds you just heard while you look atthese pictures of SAM. Point to the picture of SAM that showshow much you liktd or disliktd the loud sounds .

2. How much did you enjoy the loud sounds or how much didthey bother or annoy you? Point to the picture of SAM thatshows how you felt.

3. How did you fitl about the loud sounds? Were they fUn to hearor were they upsttting? Point to the picture of SAM that showshow you felt.

4. Were the loud sounds scary? Point to the picture of SAM thatshows how scary they were.

Arousal to Startling Sounds

5. Now look at these pictures of SAM and point to the one thatshows how excitingor how relaxingthe loud sounds were.

6. Did you gel tensedup or did you stay calm when you heard theloud sounds? Point to the picture of SAM that shows how tenseor how calm you felt.

Affective Valence to Movie

7. Remember the movie (cartoon) that you were watching? Pleasetake a moment and think about that movie (cartoon). Point tothe picture of SAM that shows how much you liktd or disliktdthe movie (cartoon) .

8. How much did you enjoythe movie (cartoon) or how much didit botheror annoy you? Point to the picture of SAM that showshow you felt.

9. How did you fitl about the movie (cartoon)? Was it fUn towatch or was it upsttting? Point to the picture of SAM thatshows how you felt.

10. Was the movie (cartoon) scary? Point to the picture of SAMthat shows how scary it was.

Arousal to Movie

II. Now look at these pictures of SAM and point to the one thatshows how excitingor how relaxing the movie (cartoon) was.

12. Did you get tensedup or did you stay calm while watching themovie (cartoon)? Point to the picture of SAM that shows howtense or how calm you felt.

EMOTIONAL RESPONSES TO STARTLE IN ADHD

RESULTS

Responses to Questions About Emotional Reactions to

Startling Sounds and Movie

The factor analysis revealed four factors. The highestloadings on the first factor corresponded to the responsesto questions 7, 8, and 9 (Table 1); on the second factorto the responses to questions 1, 2, 3, and 6; on the thirdfactor to the responses to questions 5, 11, and 12; andon the fourth factor to the responses to questions 4 and10. We labeled these four factors affect-movie, affect­startle, arousal, and scary, respectively. The ADHD boysshowed more pleasurable affective valence to both 55and the movie and a trend in the same direction forarousal (Table 2).

Two-sample t tests on the responses to each of the 12individual questions in Table 1 revealed significantdifferences between the ADHD and normal subjects forthe responses to questions 1, 4, and 6 (reactions to star­tling stimuli) and 7, 8, and 12 (reactions to the silentmovie). For all 12 responses, the mean response was less(more in the direction of pleasurable affective valenceand less arousal) in the ADHD than the normal group(Table 3). Multivariate analyses of variance evaluatedgroup differences for relevant clusters of these responses.For response clusters based on the highest factor load­ings, the results are shown in the upper half of Table 4.For those response clusters that were defined a priori(see categories in Table 1), the results are shown in thelower half of Table 4. The significant differences indi­cate more pleasurable affective valence and less arousalin the ADHD than in the normal group.

For both the normal and ADHD groups, the meanresponses in Table 3 to questions 1 through 3 weresignificantly greater than the responses to questions 7through 9, respectively (t ~ 4.47, P = .0001), indicatingmore negative affective valence to the startling soundsthan to the silent movie. Likewise, for both groups, themean responses to question 4 were significantly greater

TABLE 2Factor Scores for 29 ADHD and 32 Normal Boys (Mean :I: SO)

ADHD NormalFactor

Affect-movieAffect-startleArousalScary

-0.29:1: 0.61-0.26:1: 0.99-0.27:1: 0.94-0.19 ± 1.01

0.28 ± 1.210.29 ± 0.910.21 ± 1.010.20:1:0.97

-2.35-2.27-1.911-1.548

df P

46.6 .02359.0 .02759.0 .06159.0 NS

Note: ADHD = attention-deficit hyperactivity disorder; NS = not significant.

J. AM . ACAD. CHILD ADOLESC. PSYCHIATRY. 36:12, DECEMBER 1997 1701

ORNITZ ET AL.

TABLE 3Responses From ADHD and Normal Boys to the 12 Questions in Table I Concerning Affective Valence and Arousal

to Startling Sounds and Silent Movie (Mean ± SD)

Responses Concerning Startl ing Stimuli

Question ADHD Normal t df P

I 2.17 ± 1.14 2.91 ± 0.86 -2.87 59.02 2.45 ± 1.09 2.84 ± 1.05 -1.44 59.03 2.38 ± 1.29 2.78 ± 1.13 -1.30 59.04 1.76 ± 1.15 2.59 ± 1.34 -2.59 59.05 3.14 ± 1.55 3.41 ± 1.16 -0.77 59.06 2.00 ± 1.51 2.94 ± 1.24 -2.66 59.0

Responses Concerning Silent Movie

Question ADHD Normal t df

7 1.14 ± 0.35 1.53 ± 0.67 -2.90 47.78 1.24 ± 0.51 1.69 ± 1.06 -2.12 45.69 1.21 ±0.56 1.56 ± 0.88 -1.90 53.210 1.24 ± 0.58 1.59 ± 0.95 -1.77 52.0II 2.59 ± 1.76 3.22 ± 1.47 -1.52 59.012 1.90 ± 1.34 2.72 ± 1.40 -2.34 59.0

.006NSNS.012NS.010

p

.005

.039NSNSNS

.023

Note: Responses are on a scale from I (positive affect or low arousal) to 5 (negative affect or high arousal); 3 represents anemotionally neutral response. ADHD =attention-deficit hyperactivity disorder; NS =not significant.

Note: ADHD =attention-deficit hyperactivity disorder .

TABLE 4Group Differences Between ADHD and Normal Boys for Clusters

of Responses to the 12 Questions in Table I

Responses With Highest Factor Loadings

Responses toFactor Questions HoteHing's T2 p

I. Affect-movie 7. 8. 9 8.41 .0532. Affect-startle I, 2. 3. 6 11.0 .0453. Arousal 5. 11 . 12 7.28 .0824. Scary 4.10 7.59 .030

.012

.037

.074

.0359.517.25

14.837.07

Response Clusters Defined a Priori

1.2.3.45.6

7.8.9.1011.12

Responses toQuestions Hotelling's T 2 p

Affective valence tostartle

Arousal to startleAffective valence to

movieArousal to movie

Cluster

Startle Amplitude and Its Habituation

Startle amplitude habituated significantly (F[9 ,531]= 29.9, corrected p < .OOOI) over the 10 trial blocks.There were no group differences in amplitude and nogroup differences in rate of startle habituation.

than those to question 10 (t ~ 2.64, P :5 .013),indicating that both groups found the loud soundsmore "scary" than the silent movie. In contrast, bothgroups reported similar responses to questions 5 and 6and 11 and 12. respectively. (t:5 1.25, not significant),indicating. for both groups. similar self-reports ofarousal to the startling sounds and the movie.

Effect of Ageon Emotional Reactions to Startling

Sounds and Movie

Multivariate analyses of variance with both diagnosisand age as between-subject factors were carried out onthe a priori defined response clusters (Table 5). Agegroupings consisted of 11 normal and 14 ADHD boysbetween 84 and 105 months of age and 21 normal andIS ADHD boys between 108 and 143 months of age.The significant differences between the ADHD andnormal children (Table 4, bottom) persisted, with nointeraction with age for emotional reactions to startle(Table 5). The significant age effect for affective valenceto startle (Table 5) reflected the tendency of theyounger children in both groups to report more pleasur­able reactions to the loud sounds. The significant age bydiagnosis interaction for affective valence to the moviereflected more pleasurable reactions from the youngerADHD than the older normal children.

1702 J. AM. ACAD . CHILD ADOLES C. PSYCHIATRY. 36:12. DECEMBER 1997

EMOTIONAL RESPONSES TO STARTLE IN ADHD

TABLE 5Age and Diagnostic Effects for a Priori Defined Clusters of Responses to the 12 Questions in Table 1

T2 P

1.60 NS4.06 NS

10.97 .0460.85 NS

.002NSNSNS

20.021.543.774.50

.024

.026NS

.017

12.97.925.988.88

Affective valence to startleArousal to startleAffective valence to movieArousal to movie

Age by DiagnosisADHD vs. Normal Younger vs. Older Interaction

Cluster T2 p T2 P

Note : ADHD =attention-deficit hyperactivity disorder; NS =not significant .

72....L.--,--,---,-----,---r----,---,---..,--....--,-9 '087

• Baseline• Post-Stimulus 1

o Post-Stimulus 2

6543

ADHD 7-" Year Old Boys (N = 39)

Normal 7-" Year Old Boys (N =40)

2

73

72....L.--,--,---,-----,r----r----,---,---..,--....--,-

79

79

78

73

S"77Q.

e. 76

faa: 75

174

78

S"77Q.III- 76

i 75i~ 74

sion, i.e., less dislike, annoyance, upset. and fear (ormore pleasure, enjoyment, and fun), to the putativelypleasurable movie than to the putatively aversive star­tling sounds. That both groups gave similar self-reportsof arousal to the movie and the startling sounds suggeststhat these two affectively divergent stimuli inducedequal activation and interest.

The major finding of this investigation is that theboys with ADHD diverged from the normative reac-

Heart Rate Changes Across the 10 Trial Blocks

The heart rate data from the 40 normal subjects havebeen reported previously in relation to startle habitu­ation (Ornitz et al., 1996). Briefly, significant heart rateincreases across all 10 trial blocks during the baseline,poststirnulus 1, and poststimulus 2 periods (F[9,35I]~ 2.75. corrected p = .024) were mainly linear (F[1,39]~ 6.6, P~ .014) (Fig. 1. top panel). The only quadratictrend occurred during the baseline period (F[ 1,39] =

4.2, P = .047).The heart rates in the 39 ADHD subjects also

increased across the 10 trial blocks during all three timeperiods (F[9,342] ~ 3.58, corrected p -s .0032), withstrong linear components (F[1,38] ~ 7.60, P~ .009) .In contrast to the normal group. the ADHD groupshowed not only a quadratic trend during baseline(F[1,38] = 3.49, P = .069) but also strong quadraticcomponents during poststimulus 1 (F[ 1,38] = 12.0, P =

.00l) and poststimulus 2 (F[I.38] = 33.1, P < .000l).These results are shown in Figure 1 (bottom panel),where it can be seen that the heart rates for the ADHDgroup peak during trial block 7 and then decrease (incontrast to the continuing increases across all 10 trialblocks in the normal group). .

The repeated-measures analyses of variance wererepeated for all 79 subjects (40 normal and 39 ADHD)using diagnosis as a grouping factor. A significant trialblock by group effect for the quadratic componentduring the poststimulus 2 period (F[I.77] = 4.17 . P =

.045) confirmed the relative deficit in sustained arousalin the ADHD group.

9 '0832 456 7Trial Block

Fig. 1 Comparison of tonic heart rate changes in normal boys and boyswith attention -deficit hyperactivity d isorder (ADHD) during 5-secondperiods before (baseline) and after (post-stimulus 1 and post-stimulus 2) eachstartling stimulus. averaged evety 4 successive trials and presented as 10 tr ialblocks. BPM = beats per minute.

DISCUSSION

The self-reports of both the normal and ADHD boysappear to be valid and realistic representations of theirfeelings, as both groups reported significantly less aver-

J. AM . ACAD. CHILD ADOLE SC . PSYCHIATRY, 36:12, DECEMB ER 1997 1703

ORNITZ ET AL.

tions by reporting more pleasurable affective valence toboth the startling sounds and the interesting silentmovie . Data analyses confirmed the significance ofthese differences for their reactions to the startlingsounds whether based on factor scores, the cluster ofhighest factor loadings, or an a priori grouping ofresponses representing affective valence. For thereactions to the silent movie, the affectivevalence of theADHD group was also significantly more in thedirection of pleasurable reaction than that of the con­trols when factor scores were compared, and there weretrends in the same direction when the clusters ofhighest factor loadings or the a priori response clusterswere compared. The most significant differencesbetween the two groups were the specific responses toquestions 1 and 7 (seeTable 1 and Table 3) that simplyasked the boys how much they liked or disliked theloud sounds or the movie, respectively. It is noteworthythat these group differences occurred even though thenormal group reponed essentially neutral affectivevalence to the sudden loud sounds . Even though thesounds were loud enough to induce startling, they werenot aversive enough to evoke consistent reports of dis­pleasure in the normal boys. Hence, the emotionalresponses of the ADHD group on the dimension ofaffective valence were, relative to the normal boys,biased toward an inappropriate extreme of pleasurableresponse.

Neither the magnitude nor the rate of habituation ofthe startle responses differed between the ADHD andnormal groups. It should be noted that the startlingstimuli did not evoke self-reports of high arousal in thenormal boys. It is possible that had the startling stimulibeen louder or of longer duration, hence more noxious,larger startle responses might have been evoked in thenormal group, and differences between the ADHD andnormal group might have occurred. The manifest effectof affective valence on startle depends on a sufficientlyhigh state of arousal (Cuthbert et aI., 1996).

While the factor analysis provided distinctive factorsfor affective valence toward the startling sounds and themovie, arousal seems to have been a more generalreaction, with the arousal factor including responseswith high factor loadings from both reactions to thestartling sounds and the movie. On the arousal dimen­sion, group differences were not distinctive; there wereonly trends in the factor scores and the clusters of high­est factor loadings toward less arousal in the ADHD

group. When arousal to the startling sounds and themovie were defined a priori as separate response clusters,then the ADHD group did show significantly reducedarousal. It must be noted, however, that the responses toquestion 6, treated a priori as indicative of arousal tostartle, factored together with the responses that wereindicative of affective valence to startle (seeTable 4). Itseems, in retrospect, that for the startling sounds, thewording "tensed up" in question 6 may have beenpoorly chosen, evoking a sense of affective directionrather than arousal.

On the other hand, the tonic heart rates (indexingarousal) of the boys with ADHD failed to show the sus­tained increase throughout the experimental sessionthat was typical of the normal boys (Fig. 1). The theory(Groves and Thompson, 1970; Thompson et al., 1979)underlying the concurrence of startle habituation andton ic heart rate increase in normal boys is discussed byOrnitz et al. (1996). A deficient capacity to sustainarousal, i.e., impaired vigilance, could underlie the defi­ciency in sustained attention that is characteristic ofADHD (e.g., Voeller, 1991; Weinberg and Harper,1993). Alternatively, inattention to the emotionalsignificance, including intensity, of stimuli, may contri­bute to aberrant emotional responses. Further researchwill be required to tease apart the relative contributionsof dysfunctions of attention and arousal in ADHD.

The self-reports of affective valence biased in thedirection of pleasure and away from displeasure suggestan emotional dysfunction in ADHD. Hebb's concept(1949) of emotion as a bidirectional process fluctuatingbetween the tendency to "increase the original stimu­lating conditions (pleasurable or integrative emotions)"and the tendency to "abolish or decrease the stimulus,"i.e., displeasure, suggests that the child who has ADHDmight be more stimulus-bound or, more specifically,reward-bound (Douglas and Parry, 1983; Parry andDouglas, 1983), resulting in stimulus-driven behavior,i.e., hyperactivity and impulsivity. The bias towardpleasure suggests the need for frequent and consistentpositive reinforcement, a strategy well known to teachersand parents who work with ADHD children; negativestimuli such as a loud, firm voice or direct punishmentmay not be effective. This research suggests that chil­dren with ADHD have a tendency to react to such stim­uli with relatively more pleasure than other children,leading to effects other than intended. Also, childrenwith ADHD may ignore (because of a combination of

1704 J. AM . ACA D . C H I LD AD O L ESC . PSYCHIATRY, 36 :12, DE C EMBER 1997

hypoarousal and impaired attention) important stimulileading to impulsive and even dangerous behavior.

In summary, this study suggests an emotional com­ponent, consisting of a bias toward pleasurable reactionsand underarousal to putatively unpleasurable stimuli, tothe behavior of boys with ADHD. The limitations ofthe study include the subjective reports of the boys, theirinterpretation of the wording of the questions concern­ing their feelings, the specific experimental context inwhich the stimulation occurred, the intensity of thestimuli, and the conceptual issue as to whether hypo­arousal might lead to inattention or inattention mightinduce hypoarousal. Common speech underscores thelast issue: we "pay attention" to things, suggesting a com­ponent of effort or intensity (arousal?) in the attentionalprocess. The findings suggest further investigation tobetter understand the relationship between the emotionaland attentional dysfunction. It will be also important toexamine the predictive nature of this dysfunction: arehypoarousal and positive affective valence in the face ofnegative stimuli risk factors for the development of con­duct disorder in children with ADHD? Finally, mea­sures of emotional dysfunction should be included inADHD research along with the more typical relianceon attentional and cognitive instruments.

REFERENCESAchenbach TM . Edelbrock CS (1983). Manual for the Child Behavior

Chukliu and Rtviud Child Behavior Profil«. Burlington : UniversityAssociates in Psychiatry

Achenbach TM. Edelbrock CS (1986) . Manual fOrtb« Teachers Report Formand Teacher ),irsion ofth« Child BehaviorProfile. Burlington: University ofVermont Deparrrnent of Psychiatry

Amer ican Psychiatric Association (1987) , Diagnostic and Statistical Manual ofMental Disorders, 3rd edition-revised (DSM-Ill-R). Washington. DC:American Psychiatric Association

Barkley RA (1990). Nature and diagnos is. part I. In: Attention-DeficitHypaactiviry Disorder: A HandbookfOr Diagnosis and Treatment. NewYork: Guilford

Becker D F. Doane JA. Wexler BE (1993), Effects of emotion on perceptualasymmetry in adolescent inpatients with attention-deficit hyperact ivirydisorder, J Am Acad ChildAdole« Psychiatry32:318-321

Bradley MM. Lang PJ (1994). Measuring emot ion: the Self-AssessmentManikin and the sem antic differential. J Beha u Ther Exp Psychiatry25:4 9-59

Carey WB (1992). Temperament issues in the school -aged ch ild. PediatrClinNorth Am 39:569-584

Cotugno AJ (1995) , Personaliry attributes of attention deficit hyperactivityd isorder (AD H D ) using the Ror schach inkblot test. J Clin Psychol51:554-561

Cuthbert BN. Bradley MM , Lang PJ (1996). Probing picture perception:activation and emotion . Psychophysiology 33:103-111

Douglas VI (1972). Stop . look and listen : the problem of sustained attentio nand impulse control in hyperactive and normal children. CanJ BebavSci4:259-282

EMOTIONAL R ESPONSES TO STARTLE IN ADHD

Douglas VI. Parry PA (1983). Effects of reward on delayed reaction time taskperformance of hyperactive ch ildren .J Abnorm Child Psycholll :313- 326

Groves PM. Thompson RF (1970). Habituation: a dual-process theory.Psychol Rtv 77:4 19- 450

Haenlein M, Caul WF (1987). Attention deficit disorder with hypera criviry:a specific hypothesis of reward dysfunction. J Am Acad Child Adoles«Psychiatry 26:356-362

Harding-Roundy JL (1989). Attention-deficit hyperactiviry disorder inrelation to visuosparial processing. affect recogn it ion , and receptivelanguage. Diu Abstr Int B Sci Eng 49 :55 18- 55 19

Hebb D (1949), The Organization ofBthavior. New York: W ileyHodes RL. Cook EW III. Lang PJ (1985). Individual d ifferences in

autonomic response: conditioned association or conditioned fear ?Psychophysiology 22:545-560

Lahey BB. Schaughency EA. Hynd GW. Carlson CL. Nieves N (1987).Attention deficit disorder with and without hyperactivity: comparison ofbehavioral characteristics of clinic-referred children. JAm Acad ChildAdolesc Psychiatry 26:718-723

Lang PJ (1980). Behavioral treatment and bio-behavioral assessment:computer applications. In: Technology in Mental Health Cart DeliverySystems, Sidowski JB , Johnson JH. Williams TA, eds . Norwood. NJ :Ablex Publishing Corporation, pp 119-137

Lang PJ (1995) , The emotion probe : studies of motivation and attention.Am PsychoI50:372-385

Lang PJ, Bradley MM , Cuthbert BN (1990). Emotion. attention. and thestartle reflex. Psychol Rro97:377-395

Mackenberg H (1992), Recogn izing feelings in children with minimal braindysfun ction diagnosis. PraxKindrrpsycholKintkrpsychiatr41 :9- 16

Morgan AE. Hynd GW; Riccio CA . Hall J (1996). Validiry of DSM-IVAD H D predominantly inattentive and combined rypes: relationship toprevious DSM diagnoses/subrype differences . JAm Acad Child AdolescPsychiatry 35:325-333

Ornitz EM, Russell AT, Yuan H , Liu M (1996) , Autonomic, electroence­phalographic. and myogenic act ivity accompanying startle and itshab ituation during mid-childhood. Psychophysiology 33:507-513

Parry PA, Douglas VI (1983) . Effects of reinforcement on concept identifi­cation in hyperactive children. J Abnorm Child Psycholll :327- 340

Routh DK . Schroeder CS . O'Tuama L (1974) . Development of act ivity levelin children. Deu PsychollO:163-168

Satte rfield JH. Cantwell DP, Satterfield BT (1974) . Pathophysiology of thehyperactive child syndrome. Arch Gen Psychiatry 31 :839- 844

Satterfield JH, Dawson ME (1971) . Electrodermal correlates of hyperactiviryin children. Psychophysiology 8:191-197

Shapiro EG. Hughes SJ, August G) , Bloomquist ML (1993). Processing ofemotional information in children with attention-deficit hyperactivitydisorder. DeuNturopsychoI9:207-224

Shibagaki M, Yamanaka T, Furuya T (1993) , Attention state in electrodermalactiviry during auditory stimulation of children with attention-deficithyperactiviry disorder. Percept Mot Skills77:331-338

Thompson RF. Berry SD, Rinaldi PC , BergerTW (1979). Habituation andthe orienting reflex: the dual-process theory revisited. In : The GrientingRtfkx in Humans. Kimmel HD, van Olsr EH . Orlebeke )F, eds. NewYork: Wiley. pp 21-60

Ullmann RK. Slearor EK. Sprague RL (1985 ). Introduction to the use ofACTeRS . Psychopharmacol Bull 21:9 15- 920

Voeller KKS (1986) , Right-hemisphere deficit syndrome in children. Am JPsychiatry 143:1004-1009

Voeller KKS (1991). What can neurological models of attention. intention .and arousal tell us about attention-deficit hyperactivity di sorder? JN~ropsychiatry Clin Neurosci 3:209-216

Wechsler D (1974), Wtchskr Inulligmu Scale fOr Children-Revised. NewYork: Psychological Corporation

Weinberg WA. Harper CR (1993). Vigilance and its disorders. Neural Clin11:59-78

Welner Z. Herjanic B. Jung K. Amado H (1987). Reliabiliry. valid iry. andchild agreement studies of the Diagnostic Interview for Children andAdolescents (DICA). JAm Acad Child Adoksc Psychiatry44:649-653

J. AM. ACAD. CHILD ADOL ESC. PSYCHIATRY. 36:12. DE C EMB ER 19 9 7 1705