Mobbs, Grecicius,Abdel-Azim, Menon, Reiss - H Modulates the Mesolimbic Reward Centres.pdf - MA - References - s-abin-a

Neuron, Vol. 40, 1041–1048, December 4, 2003, Copyright

2003 by Cell Press

Humor Modulates the Mesolimbic Reward Centers

Dean Mobbs, 1,2 Michael D. Greicius, 1,2,3

modulate humor appreciation is of both social and clini-

Eiman Abdel-Azim, 1,2 Vinod Menon, 1,2,4,5

cal relevance.

and Allan L. Reiss 1,2,4,5, *

Recent advances in functional neuroimaging have en-

1 Stanford Psychiatry Neuroimaging Laboratory

abled researchers a clear avenue from which to explore

2 Department of Psychiatry and Behavioral Sciences

this critical human attribute. Figuring prominently in the

3 Department of Neurology and Neurological

semantic and linguistic aspects of humor comprehen-

Sciences

sion are the temporo-occipital junction, middle/inferior

4 Program in Neuroscience

temporal cortex, and inferior frontal gyrus (IFG), includ-

5 Stanford Brain Research Institute

ing Broca’s area (Goel and Dolan, 2001; Ozawa et al.,

Stanford University School of Medicine

2000). Moreover, stimuli that provoke laughter (i.e., the

Stanford, California 94305

motor response to humor) have been shown to modulate

activity in the supplementary motor area (SMA) proper,

a somatatopically mapped region involved in multiple

motor operations (e.g., Toyokura et al., 2002), including

Summary

motor components of expressive laughter (Iwase et al.,

2002; Osaka et al., 2003). One preliminary fMRI study of

Humor plays an essential role in many facets of human

joke-induced humor implicated the right medial ventral

life including psychological, social, and somatic func-

prefrontal cortex (MVPFC) in the amusing, or rewarding,

tioning. Recently, neuroimaging has been applied to

feeling that accompanies a joke, although this has yet

this critical human attribute, shedding light on the af-

to be replicated (Goel and Dolan, 2001). Despite this

fective, cognitive, and motor networks involved in hu-

sequence of discoveries, investigations have failed to

mor processing. To date, however, researchers have

conclusively demonstrate the subcortical correlates of

failed to demonstrate the subcortical correlates of the

the most fundamental feature of humor—reward.

most fundamental feature of humor—reward. In an

Important clues about the neurological systems in-

effort to elucidate the neurobiological substrate that

volved in regulating reward have come from a recent

subserves the reward components of humor, we un-

flurry of fMRI studies using a myriad of primary and

dertook a high-field (3 Tesla) event-related functional

secondary rewarding tasks, including monetary reward

MRI study. Here we demonstrate that humor modu-

paradigms, the perception of aesthetically attractive

lates activity in several cortical regions, and we pres-

faces, and objects signifying wealth/dominance (Aharon

ent new evidence that humor engages a network of

et al., 2001; Breiter et al., 2001; Erk et al., 2002; Knutson

subcortical regions including the nucleus accumbens,

et al., 2001). These studies have convergently docu-

a key component of the mesolimbic dopaminergic re-

mented increased hemodynamic signal in the mesolim-

ward system. Further, the degree of humor intensity

bic dopaminergic reward system, a system known to

was positively correlated with BOLD signal intensity

play a pivotal role in drug reward and motivational be-

in these regions. Together, these findings offer new

haviors (for review, see Schultz, 2002). This system en-

insight into the neural basis of salutary aspects of

compasses a variety of distinct, but interconnected,

humor.

dopamine-enriched structures, including the ventral stri-

atum/nucleus accumbens (NAcc), the ventral tegmental

area (VTA), and the amygdala. Although our understand-

Introduction

ing of the anatomical organization and function of the

mesolimbic dopaminergic reward system is relatively

Without humor, life would undeniably be less exhilarat-

advanced, our understanding of this system’s role in

ing. Indeed, the ability to comprehend and find a joke

humor, a powerful endogenous reward, remains remark-

funny plays a defining role in the human condition, es-

ably poor.

sentially helping us to communicate ideas, attract part-

In the experiment reported here, we used event-

ners, boost mood, and even cope in times of trauma

related fMRI (efMRI) to seek hemodynamic increases in

and stress (Dixon, 1980; Gavrilovic et al., 2003; Martin,

regions associated with cartoons considered to be

2001; Neuhoff and Schaefer, 2002; Nezlek and Derks,

funny. While in the scanner, each subject was presented

2001). These beneficial manifestations are compli-

with 42 cartoons previously rated, by a separate group

mented at the physiological level where humor (i.e., the

of matched subjects, as being funny and 42 nonfunny

perception that something is funny; McGhee, 1971) is

cartoons (i.e., cartoons with funny cues omitted). Sub-

thought to have numerous salutary effects, including

jects were explicitly asked to respond with a press of a

acting as a natural stress antagonist and possibly en-

button if they found the cartoon funny (Figure 1A) or not

hancing the cardiovascular, immune, and endocrine

(Figure 1B). Our rationale for the present efMRI design

systems (Bennett et al., 2003; Berk et al., 1989; Fred-

was 3-fold: (1) the unpredictable nature of random efMRI

rickson and Levenson, 1998; Fry, 1992; Lefcourt et al.,

designs allowed us to look at pure reward, rather than

1990). It is therefore apparent that developing a sophisti-

anticipatory rewards (cf. Braver and Brown, 2003); (2)

cated understanding of the discrete neural systems that

because of the subjective quality of humor appreciation,

we parsed out activation on a subject-by-subject and

cartoon-by-cartoon basis, thus allowing us to take into

*Correspondence: reiss@stanford.edu

Neuron

1042

Figure 1. Example of a Funny Cartoon and

the Same Cartoon with Funny Cues Omitted

(A) Funny cartoon. (B) Nonfunny cartoon.

Stimuli were presented in an event-related

fMRI paradigm, with each cartoon being pre-

sented in random order for 6000 ms. A jittered

interstimulus interval (ISI) was used, varying

randomly between 2000, 4000, and 6000 ms

and counterbalanced, a priori, across funny

and nonfunny events. Analysis was limited

to the blood-oxygenation level-dependent

(BOLD) signal acquired during stimulus pre-

sentation (Figure 1C). Data were collected in

one 15 min and 4 s session consisting of 84

events using a TR of 2000 ms (see Experimen-

tal Procedures for more details).

consideration individualistic differences in humor; and

ject and cartoon-by-cartoon basis) preferred funny car-

toons to those considered not funny (see Figures 1A

(3) ultimately, this design allowed us, using postscan

ratings by each volunteer, to parametrically examine the

and 1B). The primary voxel-based analysis revealed a

network of cortical and subcortical regions involved in

association between humor intensity (and presumably

the degree of reward) and blood-oxygen level-depen-

humor appreciation (Figure 2). Significantly higher BOLD

dent (BOLD) signal magnitude.

signal was identified in three cortical areas. One cluster

In accordance with previous neuroimaging studies of

was centered in the left temporo-occipital junction ex-

humor, laughter, and reward, we hypothesized that

tending into the fusiform gyrus (Brodmann area [BA] 37).

funny cartoons, in comparison to nonfunny cartoons,

A second cluster was observed in Broca’s area of the

would elicit increased activation in several language and

left lateral IFG (BA 44/45). This cluster also extended

semantic regions, including the left anterior and poste-

ventrally to include a subcluster in the temporal pole

rior temporal regions and IFG, including Broca’s area.

(BA 38). The third cortical cluster was observed in the

We also predicted that motor aspects of humor would

SMA proper (BA 6) contiguously extending to the pre-

be expressed in the SMA (i.e., laughter and smiling).

SMA and dorsal anterior cingulate (dACC; BA 32). A

Ultimately, we hypothesized, several structures within

significant activation cluster was also found encapsulat-

the mesolimbic dopaminergic reward system, including

ing the anterior thalamus, ventral striatum/NAcc, ventral

the NAcc, would become active as subjects interpreted

tegmental area (VTA), hypothalamus, and amygdala.

cartoons they subjectively considered funny.

These results are summarized in Table 1.

A post hoc covariate analysis examining the associa-

Results

tion between humor intensity (i.e., degree of funniness

as rated by each experimental subject) and BOLD signal

Behavioral Results

magnitude revealed a striking concordance with whole-

Examination of response latencies showed a robust,

brain activation. This analysis showed humor intensity

albeit nonsignificant, trend [ t (15) 1.8, p .093]

to be associated with increased activation in several

for subjects to respond faster to nonfunny cartoons

regions also detected in our primary analysis including

(mean

standard deviation: 3645.1

691.1) than to

the left temporo-occipital junction, IFG, temporal pole,

funny cartoons (3859.1

438.4). Of the funny cartoons,

SMA proper, and the mesolimbic dopaminergic reward

subjects rated an average of 61.5%

6.9% as subjec-

system (see Figure 3 and Experimental Procedures for

tively funny. These findings parallel those of a prior ef-

more details).

MRI study of humor appreciation (Goel and Dolan, 2001).

fMRI Results

Time-Series Analysis: Nucleus Accumbens

We used a random-effects model to identify residual

activation patterns for subjectively (i.e., subject-by-sub-

To further probe the hemodynamic response of the NAcc

to humor, we raised the height threshold and isolated

Humor and Reward

1043

0.05) corrected at the whole-brain level. Results revealed activation in the left temporo-occipital junction

(Brodmann area [BA] 37), inferior frontal gyrus (IFG; BA 44/45) extending ventrally to include a subcluster in the temporal pole (BA 38), and

supplementary motor area (SMA proper; BA 6/32) extending to the pre-SMA and dorsal anterior cingulate (dACC). A subcortical cluster also

was observed encompassing the ventral striatum/NAcc, anterior thalamus, ventral tegmental area (VTA), hypothalamus, and amygdala (see

Results and Experimental Procedures for more details).

the caudal aspect of this structure as a 10 voxel subclus-

nonfunny cartoons activated the temporo-occipital

junction, IFG/temporal pole, and SMA/dACC, all in the

ter (at p

0.0001; peak stereotaxic coordinate at: 6, 2,

4) within the larger subcortical cluster. Using this 10

left hemisphere. This distinct pattern of left-lateralization

has been observed in monetary and video-game reward

voxel region of interest (ROI), we extracted the average

time courses for funny and nonfunny cartoons across

tasks, in addition to event-related potential (ERP) and

clinicopathological studies of humor appreciation, signi-

all 16 subjects (Figure 4). This analysis emphasizes the

prominent increase in BOLD signal in the NAcc during

fying that this hemisphere plays a distinct role in the

processing of reward and positive emotional stimuli

funny cartoons, compared to negligible BOLD signal

decreases in response to those that were not funny.

(Breiter et al., 2001; Coulson and Kutas, 2001; Gardner

et al., 1975; Koepp et al., 1998).

Activation of the temporo-occipital junction (BA 37),

Discussion

a division of the ventral-stream of visual cortical pro-

cessing (Ungerleider and Haxby, 1994), has previously

The results reported here provide the first evidence that

humor engages a network of subcortical structures, in-

been implicated in the semantic processing of jokes

needing holistic coherence (Goel and Dolan, 2001), de-

cluding the VTA, NAcc, and amygdala—key components

of the mesolimbic dopaminergic reward system. With

tection of incongruity (Iwase et al., 2002), and the identifi-

cation of emotionally important visual cues (Geday et al.,

respect to cortical components of humor appreciation,

our results are in good agreement with previous studies.

2003). Intriguingly, this activation cluster encompassed

the left fusiform gyrus, a region which, when electrically

We found that funny cartoons when contrasted with

Table 1. Brain Areas in Which Stimuli-Related BOLD Signal Was Significant for Funny Cartoons over and above Nonfunny Cartoons

Cluster Size

Talairach Coordinates

Regions

Right/Left

Brodmann’s Area

p Value

(Voxel)

z Score

(x, y, z)

Temporo-occipital junction a

0.001

587

5.63

(

44,

60, 12)

IFG, a temporal pole

44/45/38

0.001

1457

4.73

( 50, 16, 20)

SMA proper, a pre-SMA, dACC

6/32

0.017

218

3.47

( 2, 5, 62)

Ventral striatum, a NAcc,

L/R

–

0.001

1406

4.71

( 10, 2, 4)

anterior thalamus, VTA,

hypothalamus, and amygdala

Only clusters with an extent threshold of p 0.05 corrected for whole brain are reported. Stereotaxic coordinates and Brodmann areas as

in Talairach and Tournoux (1988) atlas space.

Abbreviations: SMA, supplementary motor area; NAcc, nucleus accumbens; VTA, ventral tegmentum area; dACC, dorsal anterior cingulate

cortex; IFG, inferior frontal gyrus.

a Denotes peak activation.

Figure 2. Functional Topographical Map of Funny Minus Nonfunny Cartoons

Activation clusters were superimposed on Talairach normalized brain using MRIcro (http://www.psychology.nottingham.ac.uk/staff/cr1/

mricro.html). Significant clusters of activation were determined using the joint expected probability distribution (Poline et al., 1997) with height

0.01) and extent threshold (p

Neuron

1044

Figure 3. Composite Montage Showing Activated Regions Covarying with Degree of Humor Intensity

Significant clusters of activation were determined using the joint expected probability distribution (Poline et al., 1997), with height (p 0.05)

and extent threshold (p 0.01) corrected at the whole-brain level. Results showed activation in the left temporo-occipital junction: peak

Talairach coordinates: x, y, z; 46, 65, 14; Z 5.73, BA 19/37; left IFG: 51, 9, 27; Z 4.92; BA 9/44/45. Again this cluster extended

ventrally to the temporal pole (TP) (BA 38). A cluster was also found in the medial SMA proper: 2, 18, 51; Z 4.22; BA 6/8. A final cluster

was also observed encompassing the mesolimbic doperminergic system: 8, 33, 7; Z 4.01. This clusters also extended to the right IFG

(BA 44).

stimulated, induces laughter accompanied by a feeling

cognitive functions somewhat of an enigma (Cabeza and

Nyberg, 2000), it is thought that this region is involved in

of mirth (i.e., positive emotion) (Arroyo et al., 1993). In

view of these observations, this region may be involved

lexical retrieval and is a major storehouse for semantic

knowledge (Damasio et al., 1996; Mummery et al., 2000).

in the incongruent, or surprising (Brownell et al., 1983),

elements of a joke, and thus may play a pivotal role in

One interpretation is that these regions may constitute

a frontal-temporal network involved in integrating lan-

the early stages of the humor network.

The largest area of cortical activation occurred in the

guage and long-term memory (Goel, 2003). Such a net-

work would presumably be needed to comprehend and

left lateral IFG (BA 44), including Broca’s area, possibly

reflecting the language-based decoding of the stimuli.

find the cartoon jokes funny.

Collectively, the temporo-occipital junction, IFG, and

The vast amount of literature has implicated the IFG in

word perception and production (e.g., Price et al., 1996),

temporal pole are of particular theoretical interest as

they fit well with Suls’ influential incongruity-resolution

although it is now commonly believed that the IFG is a

polymodal language region, involved in numerous as-

model of joke appreciation, which posits that a cartoon

is found funny via a two-stage process. First, the joke

pects of language processing, including semantic and

sentence processing (for review, see Gernsbacher and

recipient finds their expectation is incongruous with the

cartoon caption. Second, the joke recipient revises their

Kaschak, 2003). This cluster also proceeded ventrally

to the temporal pole (BA 38). Although the temporal pole

initial interpretation to accommodate the caption and

the rest of the cartoon, thus establishing coherence

is highly prone to susceptibility artifact and resultant

signal loss (Ojemann et al., 1997), leaving its role in

(Suls, 1972). It is an appealing conceptualization that

Figure 4. Averaged Time Series for Funny

Compared to Nonfunny Activity in the Right

NAcc

Stereotaxic coordinates: x, y, z; 6, 2, 4;

p 0.0001.

Humor and Reward

1045

the temporo-occipital junction detects incongruence as

induced euphoria in humans (Breiter and Rosen, 1999;

suggested in previous studies of humor and laughter

Drevets et al., 2001), it is reasonable to conclude that

(Goel and Dolan, 2001; Iwase et al., 2002), while more

the NAcc activation observed in the present study re-

anterior regions, including Broca’s area and the tempo-

flects the hedonic feeling that accompanies humor. Fur-

ral pole, ascertain linguistic coherence.

ther investigations, however, are needed to unravel the

Engagement of the left SMA proper (BA 6) and pre-

discrete nexus between NAcc activation and rewarding

SMA are likely to reflect motor aspects of expressive

aspects of humor.

laughter. Intraoperative electrical stimulation of the left

The presence of left amygdala activation also pre-

pre-SMA has been shown to elicit smiles and laughter

sents a compelling finding. The amygdala is an integral

(Fried et al., 1998). Recent neuroimaging studies also

component of the dopamine reward system, providing

have shown increased activation in the bilateral SMA

excitatory innervation to the NAcc (Price and Amaral,

proper to be correlated with laughter (Iwase et al., 2002;

1981). Comparative studies have demonstrated that dis-

Osaka et al., 2003). In the present study, the SMA proper

crete ablation of the amygdala produces conspicuous

cluster also extended to the adjacent dACC, a multifac-

impairments in stimulus-reward learning (for review, see

eted structure implicated in reward-based decision

Baxter and Murray, 2002). In humans, the amygdala,

making, attention allocation, and laughter (Arroyo et al.,

while classically associated with negative emotions, has

1993; Bush et al., 2002; Osaka et al., 2003; Procyk et

also been implicated in reward magnitude (Pratt and

al., 2000). Particularly compelling is that ictal laughter

Mizumori, 1998), laughter (Iwase et al., 2002), and pro-

(i.e., gelastic seizures) has been shown to arise from a

cessing of pleasurable emotions (Yang et al., 2002). Fur-

circumscribed region encompassing the SMA and dACC

thermore, the finding of amygdala activation is of clinical

(Chassagnon et al., 2003). It is also intriguing to note

interest, as this region has been implicated in the patho-

that both SMA proper and dACC receive rich dopamine

logical features of many affective disorders. Diminished

input via ascending mesocortical projections from the

dopaminergic tone in the amygdala has been implicated

ventral striatum (Bates and Goldman-Rakic, 1993; Dum

in the emotional memory dysfunction and anhedonia

and Strick, 1993), suggesting that these regions play

observed in depression (Nestler et al., 2002) and the

an extended role in the dopaminergic reward network

“affective flattening” seen in Parkinson’s disease (Tessi-

associated with humor appreciation.

tore et al., 2002). Conceptually, connections between

A novel finding of this study relates to the detection

the amygdala and ventral striatum may provide new

of a humor-specific cluster that encompassed several

insight into the symptomatology of psychiatric disorders

subcortical structures, including the amygdala, ventral

with hypodopaminergic underpinnings.

striatum/NAcc, ventral tegmental area (VTA), anterior

In summary, our results provide compelling new evi-

thalamus, and the subadjacent hypothalamus (see Table

dence that subcortical, dopaminergic structures may

1 and Figure 2). These regions constitute the core of the

play a key role in the hedonic aspects of humor. We

subcortical dopaminergic reward network, beginning in

also, in part, replicate previous findings related to the

the VTA, where cell bodies of dopamine neurons are

cortical, presumably cognitive and motor, aspects of

located, and projecting rostrally to striatal, limbic, and

humor and laughter. These findings also make modest

frontal lobe terminal fields (Schultz, 2000). Functional

steps toward elucidating the neural basis of salutary

connectivity within this network of subcortical regions

aspects of humor that may lead to a better understand-

has been demonstrated in oral amphetamine (Devous et

ing of the putative psychological and physiological ben-

al., 2001) and cocaine (Breiter and Rosen, 1999) infusion

efits of a good sense of humor.

studies, reflecting the prominent role of dopaminergic

signaling in drug rewards.

Experimental Procedures

Of these several components of the reward system,

the NAcc has been consistently implicated in psycho-

Subjects

logically and psychopharmalogically driven rewards

Sixteen young, healthy, adult subjects (mean age, 22.4 1.8; range,

20–26 years; 7 males, 9 females) participated in this study. All sub-

(Breiter et al., 2001; Breiter and Rosen, 1999; Knutson

jects spoke native English, were right-handed (as measured by the

et al., 2001). In the present study, the time series analysis

Edinburgh Handedness Inventory; Oldfield, 1971), and were

revealed a pronounced increase in activation during

screened for history of psychiatric or neurological problems using

funny cartoons, when compared to nonfunny cartoons

the Symptom Checklist-90-R (Derogatis, 1977). Subjects were

(see Figure 4). Modulation of the NAcc by funny cartoons

deemed eligible only if scores were within one standard deviation

of the mean normative standardized sample. All experimental proce-

was also confirmed in the post hoc covariate analysis

showing that activity in this region increases with the

dures complied with the guidelines of the human subjects committee

degree of humor intensity (see Figure 3). In addition,

at Stanford University School of Medicine. Written informed consent

was obtained from each subject.

humor-elicited NAcc activation converges with findings

from fMRI studies across a number of psychologically

Stimuli: Rating of Cartoons

rewarding tasks, suggesting that this structure is in-

Subjects, similar in age and background to the experimental sub-

volved in the processing of a diverse number of stimuli

jects, chose 42 of the funniest cartoons from a portfolio of approxi-

mately 130 cartoons. In addition, each cartoon was rated for simplic-

with rewarding characteristics (Aharon et al., 2001;

Breiter et al., 2001; Breiter and Rosen, 1999; Erk et al.,

ity (i.e., how easy the jokes were to comprehend) and visual clarity.

2002; Goel and Dolan, 2001; Rilling et al., 2002). Although

Of the final 42 cartoons, 36 of the funny cartoons were captioned,

compared to 37 of the nonfunny stimuli. The majority of the cartoons

we cannot exclude other intervening factors (e.g., nov-

were of the violation-of-expectation type (cf. Herzog and Larwin,

elty), in light of prior fMRI and physiological studies

1988). The final 42 funny cartoons were compared to 42 nonfunny

implicating NAcc modulation in self-reported happiness

cartoons (i.e., funny cues omitted) in the scanner. Nonfunny car-

toons were also matched to funny cartoons for luminance and visual

(Knutson et al., 2001) and cocaine/amphetamine-

Mobbs, Grecicius,Abdel-Azim, Menon, Reiss - H Modulates the Mesolimbic Reward Centres.pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: