DOI: 10.1212/01.wnl.0000203123.68747.02 2006;66;887-893 Neurology

B. Wild Privatdozentin, F. A. Rodden, A. Rapp, M. Erb, W. Grodd and W. Ruch volitional components

Humor and smiling: Cortical regions selective for cognitive, affective, and

This information is current as of April 4, 2006

http://www.neurology.org/cgi/content/full/66/6/887located on the World Wide Web at:

The online version of this article, along with updated information and services, is

Print ISSN: 0028-3878. Online ISSN: 1526-632X. published continuously since 1951. Copyright © 2006 by AAN Enterprises, Inc. All rights reserved. Neurology is the official journal of AAN Enterprises, Inc. A bi-monthly publication, it has been

at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

Humor and smilingCortical regions selective for cognitive, affective, and

volitional componentsB. Wild, Privatdozentin Dr Med; F.A. Rodden, PhD; A. Rapp, Dr Med; M. Erb, Dr Rer Nat;

W. Grodd, Prof Dr Med; and W. Ruch, Prof Dr Phil

Abstract—Background: The interrelationships among humor, smiling, and grinning have fascinated philosophers formillennia and neurologists for over a century. A functional dissociation between emotional facial expressions and thoseunder voluntary control was suggested decades ago. Recent functional imaging studies, however, have been somewhat atodds with older studies with respect to the role of the right frontal cortex in the perception of humor. Methods: Bloodoxygen level–dependent (BOLD) activity was measured in 13 subjects during the presentation of “funny” vs “nonfunny”versions of essentially the same cartoons and compared with BOLD activity associated with “merely grinning” at similarnonfunny cartoons via fMRI. Results: Humor perception was correlated with BOLD activity in the left temporo-occipitoparietal junction and left prefrontal cortex and humor-associated smiling (recorded with an MR-compatible videocamera) with bilateral activity in the basal temporal lobes. Unexpectedly, both conditions were also accompanied by adecrease in BOLD activity in the right orbitofrontal cortex. Voluntary “grinning” in the absence of humorous stimuli wasaccompanied by bilateral activity in the facial motor regions. Conclusions: These results confirm the clinically derivedhypothesis of separate cortical regions responsible for the production of emotionally driven vs voluntary facial expressions.The right orbitofrontal decrease reconciles inconsistencies between clinical and functional imaging findings and mayreflect a disinhibition of facial emotional expression.

NEUROLOGY 2006;66:887–893

Humor, smiling, and laughter are integral compo-nents of humanity. Whereas they have been ana-lyzed for over two millennia within the contexts ofphilosophy, psychology, theology, and philology,1 ourknowledge of the neurophysiologic processes that un-derlie these phenomena is thin.2

The existence of a clinical dissociation betweenvoluntary and emotionally driven facial expressionshas been known for over 150 years.2 Lesions induc-ing a paresis of voluntary smiling were found in theventral mesencephalic pons and the internal capsuleor, if bilateral, in the operculum. Emotionally drivensmiling, on the other hand, was compromised by le-sions in the tegmental brainstem, the frontal cortex,the internal capsule and striatum, the basal ganglia,and the posterior thalamus.

More recent studies of patients whose “senses ofhumor” have been impaired owing to brain damage2-4

suggested that the right frontal cortex is necessaryfor humor perception. This region, however, was notactivated in the functional imaging studies publishedto date on humor perception (table 1). None of these

studies separated the effects of humor perceptionfrom the reactions to humor (i.e., smile).

In this event-related fMRI study, we sought toseparate 1) regions associated with purely voluntarysmiling and those associated with emotionally asso-ciated smiling and 2) regions related to humor per-ception and those related to humor-induced smilingwith particular interest in the right frontal cortex.

Methods. Overview. Thirteen right-handed men (ages 18 to25) were paid to participate in the study. All were healthy andfree of neurologic or psychiatric symptoms. Written, informed con-sent was obtained from each subject, and the experiment wasapproved by the Medical Ethics Committee of the University ofTubingen.

Stimuli consisted of nonverbal (i.e., no captions) cartoons by asingle cartoonist (Gary Larson) projected onto a screen that wasvisible to the subjects from within the MR apparatus. These“funny” cartoons were presented at random intervals during abaseline state in which the subject was shown very similar “non-funny” cartoons. These were created by digitally altering “funny”cartoons, so that the “visual punch line” was blanked out, butwithout otherwise changing the content of the picture, or by omit-ting the caption of verbal cartoons. To induce a baseline of a“looking-at-this-particular-Gary-Larson-cartoon” state, we usedsets of three variations of each cartoon. The first two were always“not funny” but the third was sometimes the original funny car-toon. Stimulus duration was 2 seconds each, without intervals.

The experiment consisted of two conceptually closely relatedsegments: The first was designed to examine humor-associated

Additional material related to this article can be found on the NeurologyWeb site. Go to www.neurology.org and scroll down the Table of Con-tents for the March 28 issue to find the title link for this article.

From the Department of Psychiatry (B.W., A.R.) and Section for Experimental Magnetic Resonance of the CNS (B.W., F.A.R., A.R., M.E.,W.G.), Departmentof Neuroradiology, University of Tubingen, Germany; and Department of Psychology (W.R.), University of Zurich, Switzerland.Supported by a grant from Fortune (University of Tubingen).Disclosure: The authors report no conflicts of interest.Received May 13, 2005. Accepted in final form November 21, 2005.Address correspondence and reprint requests to Dr. B. Wild, Zwingerweg 1, 72202 Nagold, Germany; e-mail: bawild@med.uni-tuebingen.de

Copyright © 2006 by AAN Enterprises, Inc. 887 at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

states and the second to examine states associated with affect-free“grinning.” Due to limitations in information storage capacities,the part of the experiment intended to induce the humor-associated states (“humor perception” and “humor-associatedsmiling”) was divided into three equivalent sessions. The partstudying “grinning” formed a fourth session. Sessions were shownin random order with approximately 3 minutes of rest in between.

To evoke humor-associated states, subjects were shown 63 setswith 2 “nonfunny” and one “funny” cartoon in random order inter-spersed between 72 sets with 3 “nonfunny” cartoon variations (insum 63 “funny” and 342 “nonfunny” cartoons). To evoke humorlessgrinning, a separate series of 45 sets of 3 “nonfunny” cartoons wasused. In 15 of those sets, the last was enclosed within a thin circle.This was the signal for the subject to “smile as though this cartoonwere funny.”

Monitoring of facial movements. Subjects were requested toavoid head movements during the fMRI measurements but weretold that they need not suppress their facial movements. To mini-mize head movements, subjects’ heads were restrained in a framewith bilateral padding. A specially designed nonferrous video cam-era5 monitored facial expressions (smiles).

The events of smiling were determined from the video data bytwo independent raters (interrater agreement with respect to theinstant that each smile began and ended was �95%). Smiles dur-ing the humorless grinning part and in response to the visualcommand to smile (encircled “nonfunny” cartoons) were defined asevents of “humorless grinning.” Smiles during the humor seg-ment, irrespective of whether the subject saw a cartoon deemed“funny” by the experimenters or not, were defined as events of“humor-associated smiling,” because interest here was in the sub-jects’ subjective, spontaneous reactions.

Humor perception. Humor perception events were deter-mined without regard to the video data. Immediately upon exitingfrom the MR apparatus, each subject was again shown each of thecartoons he had just seen: “nonfunny,” “funny,” and “circled,” andasked to rate it on a “funniness scale” of 0 to 5 (0 � not funny; 5 �very funny). No consideration was given to the subjects’ priorexposure to the funny cartoons inasmuch as the subjects were toldto rate the funniness they experienced at the time of scanning, notthe funniness “per se.” When the data were later evaluated, theinstants when the subject had begun viewing a cartoon that helater listed as having been at least “slightly funny” (i.e., a score of1 or more) were defined as events of humor perception.

fMRI scanning. With use of a 1.5 T scanner (SONATA; Sie-mens, Erlangen, Germany), an anatomic T1-weighted data setwas acquired for each subject with a three-dimensional gradientecho sequence (MPRAGE, 128 slices with 256 � 256 pixels insagittal orientation, 1 � 1 � 1.5 mm3, repetition time [TR] � 9.7milliseconds, echo time [TE] � 4 milliseconds, inversion time �300 milliseconds, acquisition time � 397.6 seconds, flip angle �8°). Then, in four blocks of 95 measurements, separated by periodsof rest amounting to approximately 3 minutes, 380 T2*-weightedecho-planar measurements were made, which covered the wholebrain (28 slices of 4-mm thickness in axial orientation, 1-mm gap,64 � 64 pixels, 3 � 3 mm2 in-plane resolution, TE � 39 millisec-onds, TR � 3 seconds, flip angle � 90°).

Statistical analysis. Image processing and statistical analysiswere performed using statistical parametric mapping (SPM99;Wellcome Department of Cognitive Neurology, London, UK). Allvolumes were realigned to the first image as correction for inter-scan movements by means of a rigid body transformation with sixvariables (three rotations and three translations). A coregistrationof the structural volume to the first T2* volume of each subjectensured that the functional and anatomic data were in the samecoordinate system. The structural image was normalized into astandard stereotactic space6 by estimating the numbers for anappropriate nonlinear transformation. With use of a transforma-tion with the same numbers, the T2* data were resampled to 3 �3 � 3–mm3 voxels. The functional scans were smoothed using aGaussian filter with 10-mm full width at half-maximum.

Stimulus presentation and fMRI measurements were time-linked by the presentation program. Via a split screen technique,the stimulus and the subject’s face could be monitored simulta-neously on the video film. Smile latency was measured by count-ing the number of video frames (each 40-millisecond duration)between presentation of the stimulus and the beginning of themovement.

Data were analyzed by modeling the events as stick functions,representing stimulus onset, convolved with the synthetic canoni-cal hemodynamic response function as supplied by SPM99. Thisresulted in using the times when the subjects did not smile asbaseline for the “voluntary smile” and the “humorous smile”events. The times during which the subjects were watching car-toons they later did not judge to be funny served as baseline forthe “humor perception” events, irrespectively whether these wereamong the nonfunny versions of funny cartoons or the other non-

Table 1 Synopsis of existing studies

Ref.Imagingmethod

Facialexpressionsmeasured Paradigm Activated brain regions

20 fMRI No Listening to humorous/philosophical/boring sentences

Humorous sentences vs other: Broca area, left middle gyrus

12 fMRI No Listening to phonologic andsemantic joke vs nonjokes

Phonologic jokes: left posterior inferior temporal gyrus, leftinferior frontal gyrus; semantic jokes: left posteriormiddle temporal gyrus, right posterior middle temporalgyrus, cerebellum. Covarying with subjective funniness:medial ventral prefrontal cortex

13 PET Yes(EMG)

Smile evoked by funny films vsvoluntary smile

During humor-related laughter/smile—degree of funninessnot rated: bilaterally SMA, left putamen

14 fMRI No Funny vs nonfunny verbal andnonverbal cartoons

Nucleus accumbens, left temporo-occipital junction, Brocaarea, SMA, dorsal anterior cingulum, anterior thalamus,ventral striatum, hypothalamus, and amygdalae

15 fMRI No Funny films Humor detection: anterior and posterior left temporal lobe,left inferior frontal lobe, right posterior middle temporallobe, right cerebellum. Humor appreciation: insula,amygdalae bilaterally, left lateral parietal cortex, lefthippocampus, right inferior frontal lobe

29 fMRI No Amusing/sad/neutral films Uniquely active during amusement: right caudate, rightputamen, left globus pallidus, and right hippocampus

EMG � electromyography; SMA � supplementary motor area.

888 NEUROLOGY 66 March (2 of 2) 2006 at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

funny cartoons. If cartoons immediately following each other werejudged as funny, these were modeled as one event because thepostexperiment ratings of funniness were potentially problematic:When seen after the experiment, a nonfunny version of a funnycartoon might bring recollections of the funny version, and thismight generate a rating of mild funniness.

Group results were calculated with a fixed effect model, as fora random effect model, the within-subject variance has to be equalfor each subject. Because of the widely varying number of eventsamong the different subjects (humor perception events: mean 35,SD 21.5; humorous smile events: mean 15.4, SD 10.6; voluntarygrinning events: mean 14.7, SD 0.8), this assumption was notfulfilled. This, however, makes the data nongeneralizable beyondthe cohort studied. In line with established functional imagingconventions, we report blood oxygenation level– dependent(BOLD) responses seen at an uncorrected threshold of p � 0.001for regions about which we had an a priori hypotheses. We alsoreport effects in other regions if they survived a more stringentthreshold of p � 0.05, corrected for multiple comparisons acrossthe whole-brain volume. We also used the masking procedure asdefined by SPM99. This results in isolating regions that wereactivated in one contrast (i.e., humor perception) and not by theother contrast (i.e., humorous smile). As level of significance formasking, we used p � 0.05 corrected for the first contrast and p �0.05 uncorrected for the mask.

An additional parametric analysis was performed using thefunniness ratings to determine which regions showed a pattern ofBOLD activity modulated with the degree of funniness. We con-sidered the seven subjects who had generated at least 31 events(i.e., subjects who found more than 30 cartoons as at least slightlyfunny). This selection ensured a sufficient number of events andvariances of such a degree of similarity that an estimate of para-metric variations was possible. A parametric analysis for the smil-ing events was not feasible because although there were multiplesmiling events, these events exhibited only minute differences intheir intensities. BOLD responses seen at a threshold of p � 0.05corrected for multiple comparisons with an extent of at least 6voxels are reported.

To estimate the correlation between head movements and thehemodynamic activity, we calculated the correlation between eachstimulus regressor and the realignment regressors as calculatedby SPM. We calculated this for each session in the six directions(translation x, y, z and rotation x, y, z). To average these data andto search for significant differences, we performed a Fisher Ztransformation. The mean correlations in the six directions wereas follows: tx: �0.0544; ty: �0.1163; tz: 0.0580; rx: �0.0246; ry:0.1035; rz: �0.0682.

Results. On average, subjects exhibited 17.6 spontane-ous smiles in response to the set of 135 cartoons, of which21 were the nonaltered originals and thus designed to be“funny.” Subjects averaged 14.5 voluntary “grins” in re-sponse to the 15 circled cartoons, thus showing a 97%compliance with the instructions to “smile as though thiscartoon were funny” when these cartoons were shown. Theaverage number of cartoons the subjects rated as “funny”(i.e., during which they had experienced some degree of“humor perception”) was 40.1. Thus, only 43.9% of thecartoons (i.e., 17.6/40.1) perceived as funny were funnyenough to elicit a smile. This is within the range reportedin other humor studies outside the scanner.7 The “funny”ratings for cartoons with smiles were higher than thosewithout (3.6 [SD 1.4] vs 2.1 [SD 1.4]; p � 0.01). Butthere was also a time difference between the “smileevents” and the “funny events.” The “funny events” pre-ceded the “smile events” on average by 0.6 scan (SD 0.6),that is, 1.8 seconds. Some of the “nonfunny” cartoonswere rated as “funny” too. The reasons the subjects gavefor this unexpected “perceived funniness” varied fromsuch statements as “I thought that the way the cartoon-ist drew the dog was funny” to “The cartoon remindedme of something funny.”

As was to be expected, during “grinning” on command,that is, a predominantly motor task, the predominantBOLD changes occurred in the primary sensorimotor corti-ces bilaterally (more markedly on the left). Activity wasalso seen in the supplementary motor area and in thefrontal operculae (Brodmann area [BA] 47, also morepronounced on the left), the cuneus, the right basal tem-poral lobe, and two regions in the cerebellum (figure 1A;also see table E-1 on the Neurology Web site; go towww.neurology.org).

During humor-induced smiling, in contrast, prominentregions of activation were seen in the left frontal cortexanterior to Broca area (BA 47), bilaterally in the temporalpoles, in the hippocampus and amygdala (all more pro-nounced on the right), in lateral regions of the cerebellum(more on the left), the thalamus bilaterally, the left poste-rior temporal lobe (BA 39), and the border between thesuperior and inferior parietal lobules (BA 7). As opposed tothe volitional smile condition, the primary sensorimotorcortices exhibited only a very small cluster of activation onthe right (figure 1B; see also table E-1).

Humor perception events correlated significantly withpositive BOLD activity in a large region over the lefttemporo-occipitoparietal junction, in the temporal poles(predominantly right), the left anterior prefrontal cortex,the right frontal operculum, the pallidum and putamenbilaterally, the mesencephalon, the right occipitotemporaljunction (BA 39), and small clusters bilaterally in the sen-sorimotor cortices and the right cerebellum (figure 1C; seealso table E-1).

Masking regions of activation during “humorous smile”by “humor perception” revealed regions presumably moreinvolved with the facial expression of humor-induced smil-ing than with the perception of humor (figure 2A; table 2).This showed activations in both basal temporal poles, theleft and right hippocampus, amygdala and parahippocam-pal gyrus, the left frontal operculum, the thalamus bilater-ally, and the lateral cerebellum (right and left). By usingthe opposite masking procedure, regions presumably moreinvolved in the perception of humor than in its expressionwere displayed. This revealed predominantly activation inthe left temporoparieto-occipital junction with local max-ima in BA 5/7/40/39 (figure 2B; table 2). Additional regionsof activation were in the left inferior frontal gyrus (BA45/46), the left medial frontal gyrus (BA 6), and the rightmedial temporal gyrus.

The parametric analysis, searching for regions whereactivation during “humor perception” was positively cor-related with the “amount of funniness” as rated by thesubjects afterward showed a significant parametric modu-lation again in the left frontolateral cortex (BA 10/11/44),in the right temporal pole (BA 20/21/38), in the left tempo-ral lobe (�57, �66, 6; BA 37/39), and in the left temporo-occipital junction (BA 7) (figure 2C).

Damage to the right frontal cerebral cortex resulted inan impairment of patients’ abilities to perceive or react tohumor.3,4 As expected, this region exhibited no BOLD ac-tivity during volitional grinning. Surprisingly, however, italso exhibited no positive BOLD changes either duringhumor perception or humor-induced smiling. This regiondid, however, show an unexpected negative BOLD effectduring “humorous smile” and, to a lesser degree, during“humor perception.” Differentiating those regions with sig-

March (2 of 2) 2006 NEUROLOGY 66 889 at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

nificant (p � 0.05 corrected) deactivation during “humor-ous smile” that were not significantly deactivated during“humor perception” by masking revealed BOLD changes inthe right G. frontalis medius (p � 0.004; 39, 57, 0; BA 10;figure 2D). During volitional grinning, no negative BOLDchange occurred.

Discussion. This study revealed discrete, charac-teristic patterns of cerebral BOLD activity duringnonhumorous voluntary smiles (“grinning”), duringhumor-induced smiles, and during humor percep-tion. It adds support for the existence of a dissocia-tion between neural systems for voluntary and foremotionally motivated facial expressions in healthysubjects, as previously postulated on the basis ofclinical observations in neurologic patients.2,8-10 Vol-untary grinning, as expected, seems to be controlledby cortical regions also responsible for other volun-tary facial movements. These findings also confirmthe postulated role of the operculae as premotor ar-eas for facial movements and are consistent with theclinical observations of isolated voluntary facial pa-reses after bilateral opercular lesions (the so-called“Foix–Chavany–Marie syndrome” or “anterior oper-cular syndrome”).9,11

Before discussing the differences between humor-induced smiles and humor perception in detail,methodologic issues should be considered. The con-trast for humorous smiling was based on fewerevents than for humorous perception. This couldmake it more difficult for activation during humor-induced smiles to reach the same level of significanceand would cause the maps of activation to appear inthe same general regions, but to be smaller. On the

other hand, the events of smiling occurred with car-toons with higher ratings of funniness. This couldrender the comparison between both a kind of simpleparametric analysis with larger activation expectedfor the humor-induced smile than for the humor per-ception. Both interpretations, however, do not ex-plain why there were regions activated with humorperception but not with humorous smiling (prefron-tal bilaterally, at the left temporoparietal junction,right temporal, putamen and pallidum, and the mes-encephalon; see table E-1) and also regions only acti-vated with humorous smiling (left parietal andlimbic: amygdale, hippocampus, and parahippocam-pal). Furthermore, the humor perception events oc-curred slightly earlier than the humorous smilingevents. This was to be expected: First one perceives ajoke and than reacts with facial movements. Thismeans that the two activation maps show not just aparametric difference but also a timing difference.Therefore, a better explanation, and fitting with datafrom other studies as detailed below, is that the twoshow overlapping but differing parts of a networkengaged in reactions to humorous stimuli.

Increased cerebral activity in the temporoparieto-occipital junction associated with humor perceptionhas also been described in several previousstudies.12-15 and in one study interpreted as beingrelated to the verbal nature of the humorous stimu-li.12 That these regions were activated by the nonver-bal “funny” material in the our study and others,13,14

however, suggests a different, possibly humor-associated function for this region. Other studieshave shown increased activation in nearby regions

Figure 1. Activation maps for “nonhu-morous grin” (A), “humorous smile”(B), and “humor perception” (C). Eventscompared with baseline as obtained bySPM99 as maximal intensity projectiononto representations of standard stereo-tactic space in (from left) sagittal, coro-nal (viewed from behind), andtransverse (viewed from above) view.All maps were created with a thresholdof p � 0.05, corrected for multiplecomparisons.

890 NEUROLOGY 66 March (2 of 2) 2006 at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

during such activities as viewing emotionally ladenvisual stimuli,16 the attribution of intentions to oth-ers,17 or the perception of causality.18 Just as therecognition of music depends on hierarchical synthe-ses of phenomena from purely acoustic tonal andrhythmic elements through chordal structures toemotional associations,19 the perception of humorpossibly depends on similar emergent constructions.We propose, therefore, that when confronted withthe raw material of humor, the temporoparieto-occipital region may be responsible for the task ofdetecting the incongruity necessary for the synthesisof the perception of humor—and possibly also in the

resolution of that incongruity in later stages of hu-mor perception.

The left lateral prefrontal cortex (BA 44/45/47)also has been shown to be active during humor per-ception in almost all of the existing studies, indepen-dent of the protocol employed.12,14,15,20 An increasingnumber of imaging studies confirms the key role thisbrain region plays in directing semantic search21 andin the resolution of incongruity.22,23 The activationfound in our study may likewise reflect this process.

The basal temporal activation seen in this study ismost likely related to emotional components of theexhilarative reaction to humor24 and to the facialexpression of emotion,25 inasmuch as it was mostprominent when “humorous smile” was masked by“humor perception.” This interpretation correspondswell with clinical observations of patients with foci inthe temporal poles or the hippocampal region whosmile or laugh during epileptic seizures26-28 and withthe mirth-associated laughter observed during para-hippocampal intraoperative electrical stimulation.28

It is also consistent with the reported activation inthis region including the amygdala for humor appre-ciation.14,15 This region overlaps with the mesolimbicreward system (as defined previously)14 and the re-ported activation in the basal ganglia and the hip-pocampus in amusing vs neutral films.29

Surprisingly, in most of the reported imagingstudies, no evidence was found for the participationof the right frontal lobe. This lack of evidence for anassociation between humor perception and frontallobe activity seemed inconsistent with studies in pa-tients, almost all of which indicated a prominent in-volvement of this region.3,30 Only one study12

reported activation of the medial ventral prefrontalcortex bilaterally. The proposition15 that the rightfrontal lobe may only be needed if an “immediateappraisal” of the funniness is needed, as was thecase in the patient studies, is inconsistent with theresults of the study14 in which subjects had to imme-diately judge the degree of funniness in their stimuli.None of the published studies, however, reportedhaving searched for regions with a negative correla-tion with humor detection or appreciation. A finalinterpretation of the local decreases in oxygenatedblood observed in this study must await further clar-ification of the neural concomitants of the negativeBOLD effect. This BOLD negativity, however, hasbeen suggested to reflect a reduction in neuronalactivity.31

The classic view of the right frontal cortex’s role inhumor reactions, that is, integration of complex con-textual linguistic or ideational information,3,30 hasalready been challenged.4 Those authors pointed outthat it has functions important not only for humorappreciation but also for emotional responsiveness.In particular, they stressed the anatomic connectionsbetween area 10 with the multimodal superior tem-poral sulcus and its role in facial grimacing. Thistemporal region was activated in our study as well.They conclude that “the right frontal damage is rele-

Figure 2. Comparison between humor-induced smiling (A),humor perception (B), areas with parametric modulationrelated to “humor perception” (C), and negative blood oxy-genation level–dependent (BOLD) effect during humoroussmile (D). Activation maps superimposed on a surface-rendered T1 template brain after masking (A) humoroussmile by humor perception and (B) humor perception byhumorous smile, both at p� 0.05 (view from the left [top]and below [bottom]). (C) Regions with parametric modula-tion by the individually rated degree of funniness. (D)Negative BOLD effect during humorous smile (top: viewfrom the right side, bottom: from below; threshold set atp� 0.05 corrected).

March (2 of 2) 2006 NEUROLOGY 66 891 at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

vant to producing a dissociation between cognitiveand affective responses.” Interestingly, a recent verydetailed patient study32 provided another argumentfor this. They used the Gardner “Right HemisphereCommunication Battery,” which includes all the hu-mor tests on a large number of patients with ana-tomically very well classified lesions. They found noevidence for a selective involvement of the right fron-tal cortex in humor appreciation.

We submit the hypothesis that the frontal deacti-vation is related to the disinhibition of emotionalexpression following the feeling of exhilaration, or,as Gowers (1887)33 commented on laughter manyyears ago: “The will is needed not to effect it, but torestrain it.” There have been several reports of pa-tients with prefrontal lesions who exhibited involun-tary laughter,34,35 and those patients have also beendescribed as showing inadequate emotional expres-sions. Another observation congruent with our inter-pretation is that of the right frontal decrease ofneural activation described during transient happi-ness in healthy volunteers using H2

15O PET.36

Laughing gas, an N-methyl-d-aspartate antagonist,has been suggested to exert its influence via inhibi-tion of neurons in the premotor and motor cortex.37

Goel and Dolan instructed their subjects “not tolaugh out loud” (V. Goel, personal communication,2002) during the perception of humor, whereas our

subjects, probably in contrast to those in all theother studies except one,13 were permitted to smilewhen they saw something funny. Thus, in most ex-periments, facial reactions may have been sup-pressed and a disinhibition could not be observed.

We propose a neural network associated with re-actions to humorous material in which the lefttemporo-occipitoparietal junction and the left lateralprefrontal lobe are involved in its perception andcognitive processing, the basal temporal lobes in thegeneration of facial reactions to humor, and the or-bitofrontal cortex in the disinhibition of spontaneousfacial movements. The latter two regions also arelikely candidates for regions involved in the feelingof exhilaration.

The current study represents another step in dis-tinguishing among those regions of the brain in-volved in the perception of humor, those involved inits display, and those involved in its enjoyment. Fur-ther studies will be needed to validate the functionalrelevance of these regions in the presence such con-founding factors as surprise, problem solving, detec-tion of incongruity, and presence of various moods. Adeeper understanding of the cerebral correlates ofhumor will also depend on further refinements inoperational definitions of such key concepts as“funny,” “humor,” “exhilaration,” and “mirth” andthe relationships of these concepts with the brains of

Table 2 Regions with significant change of blood oxygenation level–dependent signal

Region/side Coordinates Brodmann area

Humor perception masked by humorous smileSensorimotor cortex, R 54, �6, 39*† 6Frontal operculum, L �54, 15, �6 47Supplementary motor area, R � L 0, 3, 66*† 6Temporal poles

R 27, 6, �45 36/3842, 18, �33 38

L �36, 6, �42 36/38Amygdala/hippocampus/parahippocampus

R 18, �12, �15*†L �18, �12, �15

Thalamus, R � L 0, �18, 9*Cerebellum

R 27, �66, �21*L �42, �60, �30*

Humorous smile masked by humor perceptionFrontal operculum, L �54, 30, 15* 45/46Prefrontal, L �39, 6, 54* 6Temporo-occipitoparietal junction, L �57, �63, 30 39

�63, �39, 36 40�42, �48, 60 5/7

Temporal, R 57, �60, 9* 21/37

The table shows all clusters of activation with p � 0.05 corrected for multiple comparisons, giving the x, y, and z coordinates in theMNI space provided by SPM99.

* A priori hypotheses predicting the activation of a region (based, e.g., on observations from previous studies), significant at the lessstringent level of p � 0.001 uncorrected.

† The cluster was �10 voxels.

892 NEUROLOGY 66 March (2 of 2) 2006 at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

subjects from a broad spectrum of demographicgroups.

AcknowledgmentThe authors thank Silke Anders for help with statistics, FranziskaHosl and Hans-Jorg Mast for technical assistance, and MathiasBartels for providing a stimulating and helpful atmosphere forresearch.

References1. Martin R. Approaches to the sense of humor: a historical review. In:

Ruch W, ed. The sense of humor. New York: Mouton de Gryter, 1998:16–60.

2. Wild B, Rodden FA, Grodd W, Ruch W. Neural correlates of laughterand humour: a review. Brain 2003;126:2121–2138.

3. Gardner H, Ling PK, Flamm L, Silverman J. Comprehension and ap-preciation of humorous material following brain damage. Brain 1975;98:399–412.

4. Shammi P, Stuss DT. Humour appreciation: a role of the right frontallobe. Brain 1999;122:657–666.

5. Wild B, Erb M, Lemke N, Scholz P, Bartels M, Grodd W. Video cameraand light system for application in magnetic resonance scanners. MagnRes Imag 2000;18:893–896.

6. Collins DL, Neelin P, Peters TM, Evans AC. Automatic 3D intersubjectregistration of MR volumetric data in standardized Talairach space.J Comput Assist Tomogr 1994;18:192–205.

7. Ruch W. Die Emotion Erheiterung: Ausdrucksformen und Bedingun-gen. Department of Psychology, University of Dusseldorf, 1990.

8. Magnus A. Fall von Aufhebung des Willenseinflusses auf einige Hirn-nerven. In: Muller J, ed. Archiv fur Anatomie, Physiologie und Wissen-schaftliche Medicin. Berlin: Verlag von W. Thome, 1837:258–266.

9. Hopf HC, Muller FW, Hopf NJ. Localization of emotional and volitionalfacial paresis. Neurology 1992;42:1918–1923.

10. Poeck K. Pathophysiology of emotional disorders associated with braindamage. In: Vinken, PJ, Bruyn GW, eds. Handbook of clinical neurol-ogy. Amsterdam: Elsevier, 1985:343–367.

11. Weller M. Anterior opercular cortex lesions cause dissociated lowercranial nerve palsies and anarthria but no aphasia: Foix–Chavany–Marie syndrome and “automatic voluntary dissociation” revisited. JNeurol 1993;240:199–208.

12. Goel V, Dolan RJ. The functional anatomy of humor: segregating cogni-tive and affective components. Nat Neurosci 2001;4:237–238.

13. Iwase M, Ouchi Y, Okada H, et al. Neural substrates of human facialexpression of pleasant emotion induced by comic films: a PET study.Neuroimage 2002;17:758–768.

14. Mobbs D, Greicius MD, Abdel-Azim E, Menon V, Reiss AL. Humormodulates the mesolimbic reward centers. Neuron 2003;40:1041–1048.

15. Moran JM, Wig GS, Adams RB, Janata P, Kelley WM. Neural corre-lates of humor detection and appreciation. Neuroimage 2004;21:1055–1060.

16. Karama S, Roch Lecours A, Leroux J-M, et al. Areas of brain activationin males and females during viewing of erotic film excerpts. Hum BrainMap 2002;16:1–13.

17. Brunet E, Sarfati Y, Hardy-Bayle M-C, Decety J. A PET investigationof the attribution of intentions with a nonverbal task. Neuroimage2000;11:157–166.

18. Blakemore SJ, Fonlupt P, Pachot Clouard M, et al. How the brainperceives causality: an event-related fMRI study. Neuroreport 2001;12:3741–3746.

19. Peretz I, Zatorre RJ. The cognitive science of music. Oxford: OxfordUniversity Press, 2003.

20. Ozawa F, Matsuo K, Kato C, et al. The effects of listening comprehen-sion of various genres of literature on response in the linguistic area: anfMRI study. Neuroreport 2000;11:1141–1143.

21. Bookheimer S. Functional MRI of language: new approaches to under-standing the cortical organization of semantic processing. Annu RevNeurosci 2002;25:151–188.

22. Martin A, Chao L. Semantic memory and the brain: structure andprocesses. Curr Opin Neurobiol 2001;11:194–211.

23. Rapp A, Leube D, Erb M, Bartels M, Grodd W, Kircher T. Neuralcorrelates of metaphor processing. Cogn Brain Res 2004;20:395–402.

24. Phan KL, Wager T, Taylor SF, Liberzon I. Functional neuroanatomy ofemotion: a meta-analysis of emotion activation studies in PET andfMRI. Neuroimage 2002;16:331–348.

25. Wild B, Erb M, Eyb M, Bartels M, Grodd W. Why are smiles conta-gious? An fMRI study of the interaction between perception of facialaffect and facial movements. Psychiatry Res Neuroimag 2003;123:17–36.

26. Molinuevo JL, Arroyo S. Ictal smile. Epilepsia 1998;39:1357–1360.27. Coria F, Bahillo ME, Moral BM, Garcia GP, Ortiz SdSMR. Late-onset

isolated gelastic epilepsy secondary to entrapment of the right temporalhorn. Neurologia 2000;15:204–207.

28. Arroyo S, Lesser RP, Gordon B, et al. Mirth, laughter and gelasticseizures. Brain 1993;116:757–780.

29. Goldin PR, Hutcherson CAC, Ochsner KN, et al. The neural basementof amusement and sadness: a comparison of block contrast and subject-specific emotion intensity regression approaches. Neuroimage 2005;27:25–36.

30. Wapner W, Hamby S, Gardner H. The role of the right hemisphere inthe apprehension of complex linguistic material. Brain Lang 1981;14:15–33.

31. Harel N, Lee S-P, Nagaoka T, Kim D-S, Kim S-G. Origin of negativeblood oxygenation level-dependent fMRI signals. J Cereb Blood FlowMetab 2002;22:908–917.

32. Zaidel E, Kasher A, Soroker N, Batori E. Effects of right and lefthemisphere damage on performance of the “Right Hemisphere Commu-nication Battery.” Brain Lang 2002;80:510–535.

33. Ironside R. Disorders of laughter due to brain lesions. Brain 1956;79:589–609.

34. Zeilig G, Drubach DA, Katz ZM, Karatinos J. Pathological laughter andcrying in patients with closed traumatic brain injury. Brain Injury1996;10:591–597.

35. Mendez MF, Nakawatase TV, Brown CV. Involuntary laughter andinappropriate hilarity. J Neuropsychiatry Clin Neurosci 1999;11:253–258.

36. George MS, Ketter TA, Parekh PI, Horwitz B, Herscovitch P, Post RM.Brain activity during transient sadness and happiness in healthywomen. Am J Psychiatry 1995;152:341–351.

37. Franks NP, Lieb WR. A serious target for laughing gas. Nat Med1998;4:383–384.

ALERT: NEUROLOGY NOW USING ONLINE PEER REVIEW ANDMANUSCRIPT SUBMISSION SYSTEM

Neurology is now using an online peer review and manuscript submission system called Bench�Press.

Authors should upload all original submissions via the Neurology website (www.submit.neurology.org). TheInstructions to Authors detail the submission process and adjusted specifications.

March (2 of 2) 2006 NEUROLOGY 66 893 at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from

DOI: 10.1212/01.wnl.0000203123.68747.02 2006;66;887-893 Neurology

B. Wild Privatdozentin, F. A. Rodden, A. Rapp, M. Erb, W. Grodd and W. Ruch volitional components

Humor and smiling: Cortical regions selective for cognitive, affective, and

This information is current as of April 4, 2006

& ServicesUpdated Information

http://www.neurology.org/cgi/content/full/66/6/887including high-resolution figures, can be found at:

Supplementary Material http://www.neurology.org/cgi/content/full/66/6/887/DC1

Supplementary material can be found at:

Subspecialty Collections

smenthttp://www.neurology.org/cgi/collection/neuropsychological_asses

Neuropsychological assessment aviorhttp://www.neurology.org/cgi/collection/all_neuropsychology_beh

All Neuropsychology/Behaviorfollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.neurology.org/misc/Permissions.shtmlor in its entirety can be found online at: Information about reproducing this article in parts (figures, tables)

Reprints http://www.neurology.org/misc/reprints.shtml

Information about ordering reprints can be found online:

at Universitaet Tuebingen on April 4, 2006 www.neurology.orgDownloaded from