Neural Correlates of Sexual Arousal in Homosexual and - Reberlab

12 Pages·2007·0.52 MB·English

BehavioralNeuroscience Copyright2007bytheAmericanPsychologicalAssociation
2007,Vol.121,No.2,237–248 0735-7044/07/$12.00 DOI:10.1037/0735-7044.121.2.237
Neural Correlates of Sexual Arousal in Homosexual and Heterosexual Men
Adam Safron, Bennett Barch, J. Michael Bailey, Darren R. Gitelman, Todd B. Parrish, and Paul J. Reber
NorthwesternUniversity
Men exhibit much higher levels of genital and subjective arousal to sexual stimuli containing their
preferredsexthantheydotostimulicontainingonlythenonpreferredsex.Thisstudyusedevent-related
functional magnetic resonance imaging to investigate how this category-specific pattern would be
reflectedinthebrainsofhomosexual(n(cid:1)11)andheterosexual(n(cid:1)11)men.Comparisonsofactivation
to preferred sexual stimuli, nonpreferred sexual stimuli, and sports stimuli revealed large networks
correlatedwithsexualarousal,spanningmultiplecorticalandsubcorticalareas.Bothhomosexualand
heterosexual men exhibited category-specific arousal in brain activity. Within the amygdala, greater
preference-relatedactivitywasobservedinhomosexualmen,butitisunclearwhetherthisisacauseor
aconsequenceoftheirsexuality.Inasubsequentanalysisofregionshypothesizedtosupportarousal,
both participant groups demonstrated widespread increases in evoked activity for preferred stimuli.
Aggregatedatafromtheseregionsproducedsignificantdifferencesbetweenstimulustypesin16outof
22participants.Significantactivationaldifferencesmatchedreportedsexualorientationin15ofthese16
participants,representinganadvanceinpsychophysiologicalmeasuresofarousal.
Keywords:fMRI,event-related,sexualarousal,categoryspecificity,sexualorientation
Sexual arousal is a highly coordinated set of reactions that relevant mating psychology, men of both orientations reported
prepare an organism for reproductive behavior. It is a dynamic similarinterestsinuncommittedsexandvisualsexualstimuli,and
process characterized by widespread changes in an organism’s they valued physical attractiveness to similar degrees (Bailey,
neurophysiologicalstatesuchthatadaptiveresponsesareachieved. Gaulin, Agyei, & Gladue, 1994). Furthermore, as noted above,
Attentive, affective, and motivational systems are optimized for men of both orientations exhibit category-specific sexual arousal
thesuccessfulselectionandengagementofsexualstimuli. tovisualsexualstimuli.Thesedatasuggestthatthefactorsleading
Menshowcategory-specificgenitalandself-reportedsubjective todifferencesinsexualorientationmaynotproducedifferencesin
sexual arousal in response to visual sexual stimuli, and their other aspects of sexual behavior. The current study explored the
greatest sexual arousal is to the categories of people with whom neural mechanisms of sexual arousal in homosexual and hetero-
they prefer to have sex (Chivers, Rieger, Latty, & Bailey, 2004; sexualmen.Weexaminedbothsimilaritiesanddifferencesintheir
Freund, 1963). Heterosexual men experience higher genital and patterns of neural activation to preferred versus nonpreferred vi-
subjectivearousaltowomenthantomen(thus,theypreferfemale sualsexualstimuli.
sexual stimuli), but homosexual men show the opposite pattern
(andthusprefermalesexualstimuli).Themostinfluentialbiolog- Previous Neuroimaging Studies
icaltheoryofsexualorientationisthathormonesinfluence(largely
prenatally)thedevelopmentofneuralstructuresthatregulatesex- Several imaging studies have identified neural correlates of
ualbehavior(Ellis&Ames,1987;LeVay,1997).Bythismodel, sexual arousal (Arnow et al., 2002; Beauregard, Levesque, &
certain aspects of neurohormonal development proceed in a sex- Bourgouin, 2001; Ferretti et al., 2005; Gizewski et al., 2006;
atypical manner in homosexual individuals, resulting in psycho- Hagemannetal.,2003;Hamann,Herman,Nolan,&Wallen,2004;
logicaldifferencessuchasatypicaltriggersforsexualarousal. Holstege et al., 2003; Karama et al., 2002; Park et al., 2001;
Althoughhomosexualandheterosexualmenbecomearousedto Ponseti et al., 2006; Redoute et al., 2000, 2005; Sabatinelli,
different sexual stimuli, there are also substantial similarities in Flaisch, Bradley, Fitzsimmons, & Lang, 2004). These studies
theirsexualpsychologyandbehavior.Instudiesofevolutionarily depict sexual arousal as a composite psychophysiological state
correlated with functional changes in several brain regions (Red-
outeetal.,2000).
Fewer studies have compared sexual arousal in heterosexual
AdamSafron,BennettBarch,J.MichaelBailey,andToddB.Parrish, men and women. When participants viewed erotic film excerpts,
DepartmentofPsychology,NorthwesternUniversity;DarrenR.Gitelman, Karama et al. (2002) found activation in the thalamus and hypo-
Department of Neurology and Department of Radiology, Northwestern
thalamusinmenbutnotinwomen.Men,however,reportedhigher
University; Paul J. Reber, Department of Psychology and the Cognitive
perceived sexual arousal compared with women, thus making
NeurologyandAlzheimer’sDiseaseCenter,NorthwesternUniversity.
interpretations of gender differences problematic. A magnetoen-
WethankJoanLinsenmeierforhercommentsonanearlierdraftofthis
cephalographic (MEG) study by Costa, Braun, and Birbaumer
article.
Correspondence concerning this article should be addressed to J. Mi- (2003)foundthatthemagnitudesofcontingentnegativevariation
chael Bailey, Department of Psychology, Northwestern University, 2029 and visual-evoked magnetic fields were higher for preferred than
SheridanRoad,Evanston,IL60208.E-mail:[email protected] fornonpreferredsexualstimuliinmenandwomen.Differencesin
237
238 SAFRONETAL.
magnitudesweregreaterinmenthanwomenforthefirstcompo- terns differ between homosexual and heterosexual participants.
nent of the visual-evoked magnetic fields at 126 ms. More re- Regions that figure robustly in prior imaging studies of arousal–
cently,Hamannetal.(2004)foundamygdaladifferencesbetween affect–rewardwereidealtotestwhetheractivationisincreased(or
men and women in response to visual sexual stimuli. Although decreased) to preferred or nonpreferred sexual stimuli. Regions
they attempted to control for differential arousal through the se- with differing morphologies or functionalities between men and
lectionofstimuliwithsimilarsubjectiveratings,weakcorrelations women were of particular interest because the neurohormonal
between subjective and genital arousal in women make this ap- hypothesis of sexual orientation suggests that these areas may
proach uncertain (Chivers et al., 2004). Gizewski et al. (2006) differbetweenhomosexualandheterosexualmen.
foundgreatervisualsexualstimuliinducedactivationsinmenthan
womeninthethalamus,amygdala,anteriorcingulate,orbitofrontal Method
cortex, parahippocampal, and insular cortices. Interpretations of
groupdifferencesremainchallengingbecauseitisunclearwhether Participants
observed differences represent different levels of arousal or dif-
Twenty-fourright-handed,malevolunteers(12heterosexual,12
ferentprocessingduringsimilarlevelsofarousal. homosexual)betweentheagesof20and26(heterosexualM(cid:1)21,
Three neuroimaging studies have explored sexual arousal in ho- homosexualM(cid:1)21)wererecruitedfromtheNorthwesternUni-
mosexual men. Savic, Berglund, and Lindstrom (2005) found that
versity community and screened for compatibility with MRI. All
homosexual men and heterosexual women displayed activation of
participants were screened to ensure experience (and presumably
sexuallydimorphichypothalamicnucleiinresponsetoatestosterone
comfort) with sexual images. Participants provided written in-
derivative found in male sweat, whereas heterosexual men did not.
formed consent for participation in the research study, following
KranzandIshai(2006)foundthatheterosexualmenandhomosexual
the procedures of the Northwestern University Institutional Re-
women responded more to female faces and homosexual men and
view Board. Sexual orientation was assessed using Kinsey scale
heterosexual women responded more to male faces in the medial
reportsofsexualactivityandfeelings(Kinsey,Pomeroy,&Mar-
dorsalnucleusofthethalamusandmedialorbitofrontalcortex.Most
tin, 1948/2003). All participants indicated exclusive or nearly
recently, Ponseti et al. (2006) studied arousal in homosexual and
exclusive sexual activity and feelings for either women (hetero-
heterosexual women and men using an event-related design with
sexual) or men (homosexual) during adulthood. Each participant
images of naked male or female trunks displaying signs of genital
was in good health and free from neurological and psychiatric
arousal.Acrossgroups,theventralstriatum,centromedianthalamus,
problems. All data from 2 participants (1 heterosexual, 1 homo-
and ventral premotor cortex showed stronger neuronal responses to
sexual),andrunsfrom5participants(3heterosexual,2homosex-
preferred relative to nonpreferred stimuli. Together, these studies
ual) were not used because of poor signal-to-noise ratio (SNR)
provide strong evidence that category-specific arousal patterns are
duringfMRIdataacquisition(seebelow).
reflectedinbrainactivity.
An event-related design was used to study sexual arousal in
Stimuli
homosexual and heterosexual men. Most previous studies used
blocks over 20 s in length for eliciting mental states. Although VisualstimuliwerecollectedfromavarietyofWebsitesonthe
longer periods of stimulation may allow for deeper states of WorldWideWeb.Sexualstimuliconsistedofphotographsdepict-
arousal, they are prone to expectancy biases, habituation effects, ing male and female nudity and sexual activity. To ensure that
and distracting thought processes. Through the use of shorter arousalwasassociatedwithonlyonesexatatime,onlymale–male
stimulus presentations (3,500 ms), the current investigation is orfemale–femalesexualinteractionswereused.Thismethodhas
designedtoexamineearlierstagesofthearousalprocessandavoid beenvalidatedinphallometricstudiesofmalesexualarousal(e.g.,
confounds associated with longer blocks. Our stimulus presenta- Chiversetal.,2004).Sexualstimuliwereprescreenedtomaximize
tiontimewaschosentomaximizearousalwhilestillprovidingthe appetitive and minimize aversive responses. Nonsexual stimuli
advantagesofevent-relateddesign. consisted of photographs of male and female sports activity. As
with the sexual stimuli, males and females were never depicted
Regions of Interest (ROIs) together within a single image. All images had identical dimen-
sions.
Inourinvestigation,brainregionsimplicatedinpreviousstudies Toexaminewhereandtowhatextentthebrainshowscategory-
were organized a priori into a hypothesized network supporting specificactivation,wedisplayedsexualstimulioftheparticipants’
sexual arousal and sexually dimorphic behavior (see Table 1). preferred and nonpreferred sexes. (The preferred sexual stimuli
Regionstargetedforexaminationincludedthebasolateral–medial depictedmalesforhomosexualmenandfemalesforheterosexual
amygdala,sublenticularextendedamygdala,midbrain,hippocam- men; the nonpreferred sexual stimuli depicted females for homo-
pal complex, medial orbitofrontal cortex, nucleus accumbens– sexual men and male stimuli for heterosexual men.) To further
subcallosal cortex, rostral anterior cingulate, medial dorsal tha- characterize the networks underlying sexual arousal, we used
lamic nucleus, hypothalamus, and visual cortex (Brodmann areas neutralsportsimagesasabaselineconditiontocontrolforpossible
[BAs] 17 and 18). It is possible that other locations may also aversivereactionstothenonpreferredstimuli.
exhibitincreasedactivityinresponsetovisualsexualstimuliwhen
measured with functional magnetic resonance imaging (fMRI);
Procedure
however, we focused on the aforementioned areas in order to
conductaprioristatisticaltestsofwhetherthesebrainareasreflect Stimuliwerepresentedinanevent-relateddesignconsistingof
category-specific sexual arousal and whether their activation pat- four runs containing 100 stimuli each. Each stimulus was pre-
NEURALCORRELATESOFSEXUALAROUSAL 239
Table1
HypothesizedArousalNetworkBasedonaReviewofthePreviousLiterature
Region,Talairachcoordinates,
radius Functionalsignificance/background Arousal/rewardliterature
AC(BA24);(0,36,3);9mm Extensiveconnectionswiththeamygdalaandprojectsto Sexualarousal(Arnowetal.,2002;Ferrettietal.,
autonomicbrainstemnucleiandforebrainregions 2005;Karamaetal.,2002;Ponsetietal.,2006;
controllingautonomicfunctions(Devinsky,Morrell,& Redouteetal.,2000).Cocainecraving(Risinger
Vogt,1995). etal.,2005).
Amy;((cid:2)23,(cid:3)5,(cid:3)15);9mm Receivessensoryinformationfromthethalamus, Sexualarousal(Beauregardetal.,2001;Ferrettiet
hippocampus,andcortexandthenactivatesormodulates al.,2005).Cocainecraving(Bonsonetal.,
synaptictransmissionintargetareasappropriateforthe 2002).Morestronglyactivatedinmenthanin
reinforcementsignalwithwhichthesensoryinformation womenwhenviewingsexualstimuli(Hamannet
hasbeenassociated(Davis&Whalen,2001).Medial al.,2004).Sex-relatedhemispheric
nucleiareparticularlysensitivetogonadalsteroid lateralizationforemotionallyinfluenced
hormonesandarealikelysiteforregulationofsexually memoryanddifferingfunctionalactivityduring
dimorphicsocialbehavior(Cooke&Woolley,2005). rest(Cahilletal.,2004;Kilpatricketal.,
Lesionleadstoabnormalsexualbehaviorinprimates 2006).
(Kluver&Bucy,1939/1997).
Hip;((cid:2)30,(cid:3)24,(cid:3)9);10mm Memoryformation;encodingcanbemodulatedby Increasedactivityforsexuallyarousingstimuli
amygdalaractivity(McIntyre,Marriott,&Gold,2003). (Hamann,Ely,Grafton,&Kilts,1999).
Showsactivityforretrievalofautobiographicalmemories
(Addis,Moscovitch,Crawley,&McAndrews,2004).
Hyp;(0,2,(cid:3)7);6mm Adiversesetofnucleiregulatingemotional,autonomic,and Sexualarousal(Arnowetal.,2002;Beauregardet
endocrinefunctioningwithmorphologythatvariesasa al.,2001;Ferrettietal.,2005;Karamaetal.,
functionofsexandorientation(Byneetal.,2001;LeVay, 2002;Redouteetal.,2000).Expectancyand
1991).Responsivetoolfactorysignalsofasexualnature experienceofmonetarygains(Breiteretal.,
(Ferrisetal.,2001). 2001).Morestronglyactivatedinmenthanin
womenwhenviewingsexualstimuli(Karamaet
al.,2002).
MDTN;((cid:2)9,(cid:3)16,8);7mm Connectslimbicstructureswithcingulateandprefrontal Morestronglyactivatedinmenthaninwomen
cortices(Berridge,2003).Inhamsters,lesionsresultin (trend)whenviewingsexualstimuli(Karamaet
inappropriateandinefficientsexualbehavior(Sewards& al.,2002).Category-specificactivationfor
Sewards,2003). viewingoffacesinhomosexualand
heterosexualwomenandmen(Kranzetal.,
2005).Sexualarousal(Ponsetietal.,2006).
Midbrain;(0,(cid:3)16,(cid:3)10);9mm Containsdiffuseneuromodulatorysystemscapableof Self-stimulationsitenecessaryforpositiveaffect
releasingdopaminergicrewardsignalsthroughoutthe andliking(Berridge,2003).Eatingchocolate
mesolimbicreinforcementsystem(Morgane,Galler,& (Smalletal.,2001).Expectancyandexperience
Mokler,2005),loweringthresholdsinsensorysystems ofmonetarygains(Breiteretal.,2001).
throughacetylcholine,andincreasingvigilancewith
norepinephrine(Davis&Whalen,2001).Necessaryfor
sexualdesireintherat(Sewards&Sewards,2003).
mOFC;((cid:2)9,36,(cid:3)12);9mm Innonhumanprimates,thisregionhasextensiveconnections Cocainecraving(Bonsonetal.,2002;Risingeret
withthehypothalamus,whichallowsfortop-down al.,2005).Expectancyandexperienceof
regulationoflimbicactivity(O’Dohertyetal.,2003). monetarygains(Breiteretal.,2001).Sexual
arousal(Karamaetal.,2002;Ponsetietal.,
2006;Redouteetal.,2000).Attractivefaces
(O’Dohertyetal.,2003).
NAc;((cid:2)9,8,(cid:3)8);6mm Involvedinprocessingrewardvalueformultipletypesof Passiveviewingofbeautifulfaces(Aharonetal.,
stimuli;associateddopaminergicactivityinavarietyof 2001).Expectancyandexperienceofmonetary
addictive-compulsivebehaviors(Salamone,Correa, gains(Breiteretal.,2001).Cocainecraving
Mingote,&Weber,2003). (Breiteretal.,1997;Risingeretal.,2005).
SLEA;((cid:2)14,4,(cid:3)8);6mm Anextensiveforebraincontinuumestablishingspecific Expectancyandexperienceofmonetarygains
neuronalcircuitswiththemedialprefrontal-orbitofrontal (Breiteretal.,2001).Positiveandnegative
cortexandmedialtemporallobe,characterizedbya emotionallysalientstimuli(Liberzon,Phan,
systemofintrinsicassociationfibers,andavarietyof Decker,&Taylor,2003;Phanetal.,2003).
downstreamprojectionstothehypothalamusand
brainstem;ideallystructuredtogenerateendocrine,
autonomic,andsomatomotoraspectsofemotionaland
motivationalstates(Heimer,2003;Heimer,Harlan,
Alheid,Garcia,&deOlmos,1997).
Visual((cid:2)13,(cid:3)85,4);10mm Diffuseneuromodulatorysystemsmaylowerneuronal Sexualarousal(Ferrisetal.,2001;Karamaetal.,
thresholdstoincreasevisualinformationprocessing 2002;Mourasetal.,2003;Sabatinellietal.,
(Davis&Whalen,2001). 2004).Activatedinmotivatedattention
(Bradleyetal.,2003).
Note. Regionswerechosenbasedonhavingfunctionalsignificancethatwouldberelevanttothearousalprocess,orbasedonshowingsexuallydimorphic
activation between men and women (italics). AC (cid:1) rostral anterior cingulate; BA (cid:1) Brodmann area; Amy (cid:1) basolateral/medial amygdala; Hip (cid:1)
hippocampal complex; Hyp (cid:1) hypothalamus; MDTN (cid:1) medial dorsal thalamic nucleus; mOFC (cid:1) medial orbitofrontal cortex; NAc (cid:1) nucleus
accumbens/subcallosalcortex;SLEA(cid:1)sublenticularextendedamygdala.
240 SAFRONETAL.
sentedfor3,500mswitha1,500-msintertrialinterval.Inaddition, stimulus onset (to account for hemodynamic delay). Differences
50fixationperiodsof5,000mseachwerepseudorandomlyinter- between trial types were estimated for each participant individu-
spersedamongthestimulitofacilitatethedeconvolutionofevoked ally and then combined in a second-pass random-effects analysis
responsestoeachtrialtype.Oneachtrial,participantsindicateda thatidentifieddifferencesinevokedresponsesthatwereconsistent
preferenceratingbypressingoneoffourbuttons(lefthand,middle across participants. The reliability threshold for the whole-brain
finger (cid:1) strongly dislike; left hand, index finger (cid:1) dislike; right analysis was set by identifying the false positive rate in Monte
hand, index finger (cid:1) like; right hand, middle finger (cid:1) strongly Carlo analysis of matched noise data (i.e., simulated datasets
like). Practice trials familiarized participants with the rating pro- composed of random numbers matched to the observed data in
cedureandminimizedanyinitialstartleeffectsfromstimuli. meanandvarianceonavoxelbyvoxelbasis).Thatanalysisfound
a whole-brain false positive rate of less than .05 for clusters in
whicheachvoxelexhibitedreliableactivityatt(cid:5)4.25(p(cid:6).001,
Imaging
uncorrected) in a cluster of at least 340 mm3 in volume. These
Stimuliwereprojectedontoarear-projectionscreenandviewed valueswereusedasthecutoffcriteriaforallcontrasts.
throughamirror.ASiemensTrio3Tmagnetandradio-frequency
head coil were used to collect T2*-weighted gradient-recalled
ROI Analysis
echoplanar images (EPI) from the whole brain (44 3-mm slices;
returntime(cid:1)2,500ms,echotime(cid:1)20ms,flipangle(cid:1)90°,field To compare activity in the hypothesized arousal network for
ofview(cid:1)22cm).Sliceswereorientedalongtheplaneconnecting heterosexualandhomosexualmen,weidentified10ROIsanatom-
the anterior and posterior commissures (slightly oblique from ically. For each ROI, a sphere was centered on the loci with a
transverse)witharesolutionof3.44mm(cid:4)3.44mm(cid:4)3.00mm. radiuschosenappropriatetotheanatomicalregion.Sphereswere
Ineachrun,310whole-brainvolumeswerecollected(fourinitial drawnontheaveragedbrainthatwereslightlylargerthanthesize
volumestoallowforsignalsaturation,andsixadditionalvolumes of the region to encompass all of the relevant neural tissue and
to observe the final hemodynamic responses). For anatomical makeupforvariationsinalignmentofregionsduringnormaliza-
localization, T1-weighted images (160 1-mm axial slices; return tion.TheROIsselectedforanalysisandrationalefortheiruseare
time (cid:1) 2.1 ms, echo time (cid:1) 4.38 ms, flip angle (cid:1) 15°, field of describedinTable1.
view(cid:1)220mm;256(cid:4)192matrix)wereacquiredafterthetesting The SNR in each ROI for each participant was assessed to
runs. ensurethatadequatesignalwasobservedtosupportthepreference
analysis. Although the ROIs did vary from each other (e.g., the
medialorbitofrontalcortexhadloweraverageSNRthantheante-
Whole Brain Analysis
riorcingulatecortexorvisualcortexROIs),therewasnosystem-
Whole-brain SNR values were calculated for all runs and all aticvariationbetweengroups.
participants.AnyrunwithanSNRvaluelowerthanthemidpoint To test the discriminative validity of the cognitive model, we
between the highest and lowest observed SNRs of all runs was combined the activation from all regions and determined the
removed from the analysis. Five individual runs, each from dif- average evoked responses to different trial types. Peak hemody-
ferentparticipants,wereeliminatedbecauseoflowbrainwideSNR namic response from 5,000–10,000 ms from sexual female trial
for that specific run. In addition, 2 participants were removed typeswassubtractedfromcorrespondingresponsetosexualmale
entirely because the average of their four runs was less than the trial types. Negative scores indicated greater responses to female
cutoffvalue. sexualstimuliandwereconsideredtoreflectaheterosexualpattern
Images were coregistered through time using a three- ofbrainactivity,andpositivescoresindicatedgreaterresponsesto
dimensional registration algorithm (Cox, 1996). Functional vol- male sexual stimuli and were considered to reflect a homosexual
umeswerespatiallysmoothedwitha6.88-mm(twovoxels)full- pattern. Predicted orientation was then compared with self-report
width half-maximum Gaussian kernel to improve the SNR and todeterminethesensitivityandspecificityofthenetwork.
accommodate residual anatomical differences across participants. Additional individual tests were performed to determine
Withineachrun,voxelswereeliminatediftheSNRwaslessthan whetherthenetworkwouldvaryonthebasisofsubjectivereport.
20acrossthescan(i.e.,themeanEPIsignalwasnotatleast20(cid:4) A subset of trials was chosen contingent on the condition of
greaterthanthevariance).Eachoftheremainingrunswastrans- preferred stimuli being strongly liked (SL preferred) and nonpre-
formed (Collins, Neelin, Peters, & Evans, 1994) to conform ap- ferredstimulibeingstronglydisliked(SDnonpreferred).Consid-
proximately to the atlas of Talairach and Tournoux (1988; using eringthatROIswerechosenonthebasisofattentional,affective,
theMNI-305referencemodel)withafinalresolutionof2.5mm3. and motivational significance, the use of extreme ratings on the
Allavailablerunsoffunctionaldata(usuallyfour)wereconcate- liking scale should help control for variations in liking across
nated into a single time series for each participant. The average participants and may enhance differential processing in those
response to each trial type was estimated using a general linear areas.
modelanalysis(Ward,2006)thatincludedtheonsetofeachtrial WeaveragedthedataobtainedwithintheROIintoasingletime
typeandseveralcontrolvariables;themeanandlineardriftinthe course and estimated the peak response to each stimulus type to
EPIsignalwereestimatedineachvoxelforeachofthefourruns, determineresponsivenesstodifferenttrialtypes.Wethenanalyzed
andestimatesofcorrectedmotionforeachtimepointwereusedto theseestimateswith2(cid:4)2analysesofvariance(ANOVAs)using
removesignalchangescorrelatedwithhead–brainmotion.Differ- stimulussex(preferred,nonpreferred)andorientation(heterosex-
encesbetweentrialtypeswereestimatedbycontrastingtheaver- ual, homosexual) as factors. To test whether the inclusion of
age peak response within the window of 5,000–10,000 ms after subjective data would yield different neural responses, we per-
NEURALCORRELATESOFSEXUALAROUSAL 241
formedadditionalANOVAsforSLpreferredandSDnonpreferred sexual men averaged 0.6 positive evaluations. He also tended to
stimuli. give less negative evaluations for nonpreferred stimuli compared
withotherheterosexualmen(thepercentagesrated“stronglydis-
Results like” were 48% and 66%, respectively). Furthermore, he rated
preferredstimulilesspositivelycomparedwithotherheterosexual
Subjective Responses men (the percentages rated “strongly like” were 46% and 82%,
respectively).
Figure1depictstheresultsofthesubjectivemeasuresofliking.
Participants reported liking their preferred stimulus types more
Whole Brain Analysis
thantheirnonpreferredstimulustypes,F(1,20)(cid:1)405,p(cid:6).001.
This pattern was found in all participants. The magnitude of the The preferred versus nonpreferred comparison (N (cid:1) 22) iden-
differencesbetweenpreferredandnonpreferredstimulustypesdid tified greater activity for preferred stimuli in a large number of
notdiffersignificantlybetweenhomosexualandheterosexualmen, brain regions (see Figure 2 and Table 2), including widespread
F(1,20)(cid:1)1.59,p(cid:1).222.Whenabsoluteratingsofnonpreferred visual regions (BA 17, BA 18, BA 19), posterior cingulate, pre-
stimuli were compared, however, heterosexual participants gave cuneus,leftsuperiorparietallobule,leftglobuspallidus,thalamus,
lowerratingstomalestimulithanhomosexualparticipantsdidto left putamen, left insula, bilateral caudate (head and body), left
femalestimuli,F(1,20)(cid:1)4.90,p(cid:6).05.Nodifferencewasfound claustrum,leftdorsalamygdala/sublenticularextendedamygdala,
forratingsofpreferredstimuli,F(1,20)(cid:1).006,p(cid:1).949. anterior cingulate, hypothalamus, nucleus accumbens/subcallosal
Although heterosexual participants gave approximately equal cortex,leftsuperiorfrontalgyrus,andleftcerebellum.
ratings to male and female sports stimuli, F(1, 20) (cid:6) .001, p (cid:1) The only areas with greater activity to nonpreferred stimuli
.990,homosexualparticipantsgavemorepositiveratingstomale likelyreflectedtheuseofthelefthandtopressthe“dislike”and
thanfemalesportsstimuli,F(1,20)(cid:1)7.19,p(cid:6).05.Becauseof “strongly dislike” buttons (sensory–motor areas of right pre-/
thisdifferenceinreportedvalance,subsequentanalysesusedonly postcentral gyri). Corresponding greater activation for preferred
femalesportsstimuliascomparisonsfortheheterosexualpartici- stimuliwasseeninthematchingipsilateralregions.
pantsandmalesportsstimuliascomparisonsforthehomosexual For the preferred versus sports comparison (see Figure 2b and
participants.Ratingsofmalesportsstimulibyhomosexualpartic- Table 3), preferred stimuli elicited greater activation than sports
ipantsandratingsoffemalesportsstimulibyheterosexualpartic- stimuli in a widespread parieto–occpito–temporal network: BA
ipantsweresimilar,F(1,20)(cid:1)1.60,p(cid:1).220. 17, BA 18, BA 19, left sensory–motor cortex, bilateral anterior
When the responses given to different stimulus types were cingulate,bilateralcaudate/putamen/thalamus,leftinsula,leftpos-
examinedinindividualparticipants,Participant16showedunusu- teriorcingulate,andrightinferiorparietalcortex.Asintheprevi-
allypositiveevaluationsfornonpreferredstimuli.Thisparticipant, ous comparison, the only area with greater activation for sports
a self-reported heterosexual, had 10 instances in which he gave stimuli was the right sensory–motor cortex, reflecting the greater
positiveevaluationstononpreferredstimuli,whereasotherhetero- tendencytousethelefthandtorespondtothosestimuli.
Figure1. A:Averagepreferenceratings((cid:2)SE)ofhomosexualandheterosexualparticipantstothepreferred
andnonpreferredsexualstimuli;(cid:3)2correspondstostronglydislikingastimulusand(cid:7)2correspondstostrongly
likingastimulus.B:Individualparticipants’averagedifferencesinratingsofpreferredandnonpreferredstimuli.
Foreachgroup,themiddlelinerepresentsthemean,andtheouterlinesmarkthe95%confidenceintervals.
242 SAFRONETAL.
Table5),16outof22participantsexhibitedsignificantdifferential
activation between stimulus types; of those 16 participants, 15
showedgreateractivationforstimulifeaturingtheirpreferredsex.
The one individual whose predicted brain activity did not corre-
spondwithreportedsexualorientationwasParticipant16.
WhendifferentialactivationwasassessedforSLpreferredand
SDnonpreferredstimuli(seeTable5),participantstendedtoshow
larger magnitudes of activation than when subjective preference
was not used to select trials, F(1, 21) (cid:1) 3.77, p (cid:1) .067. Eight
participants, however, did not show reliable differences between
SL preferred and SD nonpreferred stimuli. The larger magnitude
contrasts and fewer significant results may be explained by the
smallernumberoftrialsthatwentintotheseanalyses.
Table6depictstheresultsoftheROIanalysis.Allregionsinthe
hypothesized arousal network exhibited greater activity for pre-
ferred stimuli than for nonpreferred stimuli. Group differences
were found in the amygdala for the preferred–nonpreferred com-
parison with homosexual men showing category-specific activa-
tion to male sexual stimuli and heterosexual men showing tonic
activation across stimulus conditions (see Figure 4). When sub-
jectivedatawereusedtoselectSLpreferredandSDnonpreferred
stimuli,theinteractionbetweenstimulustypeandsexualorienta-
tionfellbelowsignificanceandbecameatrend,F(1,20)(cid:1)3.18,
p(cid:1).091.AnadditionaltestwasperformedwithoutParticipant16
toensurethatananomalousdatapointwasnotskewingtheresults.
Theseresultswerestillsignificantforthepreferred–nonpreferred
Figure2. Areasofdifferentialactivitytopreferredversusnonpreferred comparisons including all trials, but not the subjectively filtered
stimuli(A;increasedactivitytopreferredstimulishowninred),preferred preferred–nonpreferred contrasts, F(1, 20) (cid:1) 4.41, p (cid:6) .05, and,
versussportsstimuli(B;increasedactivitytopreferredstimulishownin F(1,20)(cid:1)3.53,p(cid:1).077,respectively.
red),andnonpreferredversussportsstimuli(C;increasedactivitytosports
stimulishowninblue).Sportsimagescontainedactorsoftheparticipants’
preferredsex.Axialslicesareatz(cid:1)0,(cid:7)15,and(cid:7)30. Discussion
Homosexual and heterosexual men exhibited category-specific
networksofactivitywhenviewingmaleandfemalesexualstimuli.
Forthenonpreferredversussportscomparison(seeFigure2and
The finding of large networks of activity spanning multiple ana-
Table 4), greater activity for sports stimuli was observed in the
tomical domains suggests that the neurophysiological state of
posterior cingulate bilaterally, left sensory–motor cortex, left in-
sexual arousal involves coordination of cognitive/affective/
sula,rightinferiorparietallobule,leftmiddleandsuperiortempo-
sensory–motor systems such that increased vigilance, positive
ral gyrus, left parahippocampal gyrus, left middle frontal gyrus,
affect,andmotivationalresponsesaretriggeredbysalientstimuli.
left middle occipital gyrus, left precuneus, cerebellum, and right
When sexual stimuli matched the participant’s stated preference,
middletemporalgyrus.Greateractivitywasobservedfornonpre-
dramaticallyincreasedactivitywasobservedacrossmultiplecor-
ferred stimuli in the left inferotemporal cortex, right precuneus,
ticalandsubcorticalregions.
and right sensory–motor cortex. Differential activity in pre-/
Most previous studies mapped arousal by comparing sexual
postcentralgyrilikelyreflectsmotoractivity(asabove).
stimuliwithneutralstimuli.Thisstudy,however,usedthreecon-
Inthewhole-brainanalysis,thepatternofgreaterevokedactiv-
ditions (preferred, nonpreferred, and neutral sports) and three
ity for preferred stimuli did not differ reliably across the partici-
comparisons to examine the neural correlates of sexual arousal.
pant groups (i.e., the comparison of activity to female and male
The preferred–nonpreferred subtraction yielded the most wide-
stimuliforheterosexualparticipantswassimilartothecomparison
spread network of activity, with greater activation to preferred
of activity to male and female stimuli for homosexual partici-
stimuli.Althoughmensometimesreportaversiontononpreferred
pants).Nosignificantclusterswerefoundatthewholebrainlevel
sexual stimuli (Freund, Langevin, Cibiri, & Zajac, 1973; Freund,
for the preferred–nonpreferred, preferred–sports, and
Langevin, & Zajac, 1974), no regions of consistently greater
nonpreferred–sports, homosexual–heterosexual double subtrac-
activity to the nonpreferred stimuli were identified (outside of
tions.
motor areas, because of the method of reporting arousal level).
Brain regions regulating aversion may have been insufficiently
ROI Analysis
activated in the context of this study. Another possibility is that
Differences between the participant groups were assessed with aversion activates some of the same regions as the appetitive
greater sensitivity in the network of areas hypothesized to be response,buttoasmallerdegree.
involvedinarousalusinganROIanalysis.Whencategory-specific When nonpreferred sexual stimuli were compared with sports
activity was assessed over the entire network (see Figure 3 and stimuli, however, greater activity to the nonpreferred stimuli was
NEURALCORRELATESOFSEXUALAROUSAL 243
Table2
AreasExhibitingDifferentialActivitytoPreferredandNonpreferredStimuli
Region BA X Y Z Size(mm3)
Preferred(cid:5)Nonpreferred
Verylargeclusterincluding: 239,203
Lmiddleoccipitalgyrus 18/19 (cid:3)28 (cid:3)85 (cid:3)5
(cid:3)32 (cid:3)84 14
Rmiddleoccipitalgyrus 18/19 34 (cid:3)82 (cid:3)4
Lprimaryvisualcortex 17 (cid:3)18 (cid:3)89 1
Rprimaryvisualcortex 17 (cid:3)21 (cid:3)89 4
Rsuperioroccipitalgyrus/precuneus 19 29 (cid:3)85 24
Lprecuneus 19 (cid:3)22 (cid:3)83 35
Lposteriorcingulate (cid:3)12 (cid:3)53 22
Rposteriorcingulate 12 (cid:3)56 22
Lprecuneus/inferiorparietalcortex 7 (cid:3)9 (cid:3)61 64
Rhippocampus 31 (cid:3)41 (cid:3)2
Lbasalgangalia/thalamus (cid:3)27 (cid:3)24 (cid:3)1
(cid:3)19 (cid:3)15 (cid:3)1
Lthalamus (cid:3)19 (cid:3)23 7
Lclaustrum/pulvinar/insula (cid:3)24 (cid:3)25 17
Lcaudatehead (cid:3)14 7 14
Lcaudatebody (cid:3)19 (cid:3)18 27
Rcaudatehead 9 9 16
Rcaudatebody 18 (cid:3)17 27
Linsula 13 (cid:3)33 2 18
(cid:3)29 27 17
Lanteriorcingulate 24/32 (cid:3)1 26 21
Ranteriorcingulate 24/32 10 20 21
Lsensory–motorcortex (cid:3)36 (cid:3)32 52
4 (cid:3)38 (cid:3)17 58
(cid:3)39 (cid:3)43 65
(cid:3)16 (cid:3)36 66
(cid:3)27 (cid:3)38 65
Rcerebellum 26 (cid:3)40 (cid:3)34
26 (cid:3)53 (cid:3)24
Lsuperiorfrontalgyrus 8 (cid:3)19 19 48 3,625
Rcingulate 24/32 17 1 36 672
Rsuperiorfrontalgyrus 8 16 23 45 672
Rmiddlefrontalgyrus 46 45 21 26 562
Nonpreferred(cid:5)Preferred
Rsensory–motorcortex 4 34 (cid:3)30 63 8,266
Note. L(cid:1)left;R(cid:1)right;BA(cid:1)Brodmannarea.
observedinareasassociatedwithvisualprocessingandattention, studyingtheinitialstagesofarousal,whichseemsespeciallylikely
includingtheinferotemporalcortexandprecuneus.Greateractiv- to reflect automatic processes. Finally, Ponseti et al. (2006) gen-
ityintheinferotemporalcortexfornonpreferredstimulicompared eratedcategory-specificarousalusinganevenshorterimagepre-
withsportsstimulimayreflectresponsestofacesandbodiesinthe sentationwitha300-mstrialonsetand2,700-msintertrialinterval.
sexual images (Downing, Jiang, Shuman, & Kanwisher, 2001). BecauseMEGstudieshaveshowncategory-specificneuronalac-
Greater activity in the cerebellum and basal ganglia during the tivity to visual sexual stimuli at as early as 126 ms (Costa et al.,
viewingofsportsstimulimightreflectatendencyforparticipants 2003),itseemslikelythatcategory-specificarousaldoesnotrely
to engage in mental rehearsal of motor routines relating to the onconsciouscontrolforitsexpression.
particularsport(Decetyetal.,1994). Examination of previous reports on arousal-related neural ac-
Ourresultssuggestthattheneuralprocessesunderlyingarousal tivitysuggestedanetworkofspecificregionsparticipatinginthe
quickly, specifically, and robustly activate in response to sexual arousal response (see Table 1). Each of these regions exhibited
stimuli of a preferred nature. Beauregard et al. (2001) found that category-specificactivitythatwashigherforthepreferredstimuli
thesuperiorfrontalcortexandtheanteriorcingulatewereinvolved thanforthenonpreferredstimuli(seeTable6).Groupdifferences
in the conscious self-regulation of sexual arousal. In the current were found in the amygdala, a region that has shown differential
study,thoseareasdidnotshowelevatedactivationtononpreferred activation between men and women in previous neuroimaging
stimulicomparedwithpreferredorsportsstimuli.Thus,wethink studies (Cahill, Uncapher, Kilpatrick, Alkire, & Turner, 2004;
itisunlikelythatconsciousregulationofsexualarousalaccounts Hamann et al., 2004; Kilpatrick, Zald, Pardo, & Cahill, 2006). If
for the category-specific pattern we obtained. Furthermore, the neural structures are organized differently in homosexual and
shortstimuluspresentationtimes(3,500ms)andrapidsuccession heterosexual men in accordance with the neurohormonal hypoth-
of images (5,000 ms between onsets) should ensure that we are esis of sexual orientation, then it is possible that these variations
244 SAFRONETAL.
Table3
AreasExhibitingDifferentialActivitytoPreferredandSportsStimuli
Region BA X Y Z Size(mm3)
Preferred(cid:5)Sports
Largeclusterincluding: 67,797
Learlyvisualcortex 17 (cid:3)4 (cid:3)90 (cid:3)3
Rearlyvisualcortex 17 11 (cid:3)90 (cid:3)3
Lmiddleoccipitalgyrus 18 (cid:3)31 (cid:3)88 (cid:3)4
Lmiddleoccipitalgyrus 18/19 (cid:3)34 (cid:3)88 9
Rmiddleoccipitalgyrus 19 27 (cid:3)90 4
Rcuneus/middleoccipitalgyrus 17/18/19 22 (cid:3)89 9
Rprecuneus/inferiorparietalcortex 7/19 27 (cid:3)84 24
Lsensory/motorcortex (cid:3)38 (cid:3)39 56 27,453
Bilateralanteriorcingulate/caudate (cid:3)1 21 15 18,281
Lbasalganglia/thalamus (cid:3)22 (cid:3)26 5 6,781
Lbasalganglia (cid:3)14 0 (cid:3)6 1,062
Linsula 13 (cid:3)37 (cid:3)3 4 969
Lposteriorcingulate (cid:3)20 (cid:3)62 7 625
Rcaudatetail 16 (cid:3)34 28 484
Rinferiorparietalcortex 40 56 (cid:3)26 38 359
Sports(cid:5)Preferred
Rsensory–motorcortex 29 (cid:3)29 64 1,359
Note. L(cid:1)left;R(cid:1)right;BA(cid:1)Brodmannarea.
could manifest as differential activation in sexually dimorphic finding that heterosexual men rated the nonpreferred stimuli lower
regions.Althoughtheamygdaladifferenceswefoundareconsis- thandidhomosexualmen.Groupdifferencesinamygdalafunctioning
tentwiththehypothesisthathomosexualmen’sbrainsshowatyp- maybeduetonegativeattitudestowardhomosexualityonthepartof
ical patterns of activation in sexually dimorphic regions, these heterosexuals. However, a significant trend persisted when we re-
results do not paint a simple picture of homosexual men having strictedtheanalysistothosetrialsonwhichheterosexualandhomo-
“femalebrains.” sexualmengavesimilarratings.Ourfindingofagroupdifferencein
Interpretation of group differences remains challenging because theamygdalaneedstobereplicatedbeforemuchenergyisexertedto
such variations may represent either dissimilar levels of arousal or explainit.Assumingthedifferenceisreplicable,evidencefromfunc-
dissimilarprocessingduringsimilarlevelsofarousal.Asocialalter- tionalimagingisunabletoresolvewhetherthedifferenceisdueto
native to the sexual dimorphism explanation is suggested by the innateorsocialfactorsortoboth.
Table4
AreasExhibitingDifferentialActivitytoSportsandNonpreferredStimuli
Region BA X Y Z Size(mm3)
Sports(cid:5)Nonpreferred
Rcerebellum 20 (cid:3)47 (cid:3)30 6,391
Bilateralposteriorcingulate 30 9 (cid:3)52 11 4,375
Lsensory–motorcortex (cid:3)31 (cid:3)32 65 4,031
Linsula 13 (cid:3)40 (cid:3)30 20 2,125
Rinferiorparietallobule 7/39 40 (cid:3)68 41 1,344
Lmiddletemporalgyrus 22/38 (cid:3)54 (cid:3)2 (cid:3)16 1,156
Lpostcentralgyrus (cid:3)51 (cid:3)21 45 875
Lsuperiortemporalgyrus 22 (cid:3)55 (cid:3)47 11 875
Lparahippocampalgyrus 36 (cid:3)24 (cid:3)34 (cid:3)14 812
Lmiddlefrontalgyrus 6 (cid:3)12 (cid:3)17 74 656
Lmiddleoccipitalgyrus 18 (cid:3)10 (cid:3)102 11 578
Lcerebellum (cid:3)41 (cid:3)59 (cid:3)40 484
Lprecuneus (cid:3)1 (cid:3)69 39 469
Lmiddlefrontalgyrus 9 (cid:3)52 17 32 422
Rmiddletemporalgyrus 22 58 (cid:3)43 0 406
Lmiddlefrontalgyrus 10/46 (cid:3)38 48 11 391
Nonpreferred(cid:5)Sports
Rsensory–motorcortex 43 (cid:3)32 54 3,859
Linferioroccipital/temporalcortex 37 (cid:3)46 (cid:3)70 (cid:3)5 2,359
Rcuneus/precuneus 19 26 (cid:3)85 26 766
Note. L(cid:1)left;R(cid:1)right;BA(cid:1)Brodmannarea.
NEURALCORRELATESOFSEXUALAROUSAL 245
Figure3. Differentialactivityaveragedacrossallregionsofinterest.A:Averagerawactivity((cid:2)SE)foreach
stimulus type for homosexual and heterosexual participants. B: Differences in activation to the male sexual
stimuliminusactivationtothefemalesexualstimuliforindividualparticipants.Foreachgroup,themiddleline
representsthemean,andtheouterlinesmarkthe95%confidenceintervals.
ActivityassessedacrosstheseROIswasareliablepredictorof 6 additional participants failing to exhibit significant differences
self-reported sexual orientation, with greater responses to female betweenconditions).Althoughpatternsofstatedlikingseemedto
stimuliforheterosexualmenandgreaterresponsestomalestimuli perfectly track self-reported sexual orientation, this association
forhomosexualmenin15outof16participantsinthisstudy(with maybeinflatedbyfactorsunrelatedtoarousalthatinfluenceboth
Table5
DiscriminativeValidityTestforDifferentRegionsHypothesizedtobeInvolvedintheArousalResponse
Sexualmale(cid:3)Sexual Stronglylikedpreferred(cid:3)
female Stronglydislikednon-preferred
Sexual Sexual
# Orientation femalerating malerating LC[0] F(1,1,152) LC[0] F(1,1,152)
2 Hetero 1.98 (cid:3)1.99 (cid:3)1.770 5.47* 1.790 2.33*
3 Hetero 0.65 (cid:3)1.39 (cid:3)3.290 13.40*** 6.220 10.30***
4 Hetero 1.14 (cid:3)1.83 (cid:3)0.615 0.52 3.530 3.18***
5 Hetero 1.98 (cid:3)1.92 0.332 0.51 (cid:3)0.310 0.43
7 Hetero 1.32 (cid:3)2.00 (cid:3)1.670 10.20** 2.560 16.10***
8 Hetero 1.18 (cid:3)2.00 (cid:3)3.750 26.60*** 4.130 21.50***
9 Hetero 1.68 (cid:3)1.33 (cid:3)0.655 0.15 1.050 1.76
11 Hetero 1.84 (cid:3)1.93 (cid:3)2.790 28.80*** 3.270 33.90***
14 Hetero 1.35 (cid:3)1.70 (cid:3)3.930 8.54** 1.350 0.62
16 Hetero 1.26 (cid:3)1.14 1.610 4.09* 0.080 0.01
20 Hetero 0.77 (cid:3)1.48 0.435 0.65 0.620 0.44
10 Homo (cid:3)1.93 1.57 1.320 5.80* 1.430 5.14*
12 Homo (cid:3)0.88 1.53 3.070 24.10*** 0.760 0.41
13 Homo (cid:3)1.53 1.60 2.060 4.72* 2.870 6.04*
15 Homo (cid:3)1.56 1.66 1.620 7.65** 2.720 14.40***
17 Homo (cid:3)1.40 1.06 0.820 2.38 1.490 3.45*
18 Homo (cid:3)1.40 0.82 3.020 8.35** 4.210 4.30*
19 Homo (cid:3)1.20 1.54 0.440 0.40 0.220 0.06
21 Homo (cid:3)0.90 0.73 2.100 7.54** 2.020 1.35
22 Homo (cid:3)1.39 1.39 4.430 68.20*** 5.140 51.20***
23 Homo (cid:3)0.67 1.18 5.020 63.70*** 6.990 31.20***
24 Homo (cid:3)2.00 1.93 2.030 9.75*** 1.940 8.54**
Note. Generallineartestingwasperformedforneuralresponsivenesstospecificstimulustypesforindividualparticipants.TheLC[0]termindicatestheestimated
differentialevokedneuralactivity,andtheFstatisticindicatesthesignificanceofthismeasure(Fstatisticswithoutasterisksarenotsignificant).NegativeLC[0]
termscorrespondtogreaterevokedactivityforthesecondstimulusset,andpositivescorescorrespondtogreateractivityforthefirststimulusset.
*p(cid:6).05. **p(cid:6).01. ***p(cid:6).001.
246 SAFRONETAL.
Table6
AnalysesofVarianceTestingforCategory-SpecificActivityinResponsetoPreferredandNonpreferredStimulusTypesandWhether
ThisVariedasaFunctionofSexualOrientation
Stronglylikedpreferred(cid:3)Stronglydisliked
Preferred(cid:3)Nonpreferred nonpreferred
Stimulus(cid:4)Orientation, Stimulus(cid:4)
Regions Stimulus,F(20,1) F(20,1) Stimulus,F(20,1) Orientation,F(20,1)
AC(BA24/32) 15.20*** 2.870 16.50** 1.210
Amy 7.81** 5.670* 4.61* 3.180
Hip 20.60*** 1.540 36.50*** 0.662
Hyp 6.12* 0.446 10.30*** 0.064
MDTN 13.20** 0.232 15.30*** 0.149
Midbrain 10.50** 0.861 5.18* 1.180
mOFC 5.86* 0.860 7.22* 1.490
NAc 13.00** 0.169 22.70*** 0.203
SLEA 13.00** 0.748 22.90*** 0.459
Visual(BA17/18) 41.20*** 0.184 46.30*** 0.283
Note. Allregionsexhibitedgreateractivityforpreferredthanfornonpreferredstimuli.Fstatisticswithoutasterisksarenotsignificant.AC(cid:1)rostral
anteriorcingulate;BA(cid:1)Brodmannarea;Amy(cid:1)basolateral/medialamygdala;Hip(cid:1)hippocampalcomplex;Hyp(cid:1)hypothalamus;MDTN(cid:1)medial
dorsalthalamicnucleus;mOFC(cid:1)medialorbitofrontalcortex;NAc(cid:1)nucleusaccumbens/subcallosalcortex;SLEA(cid:1)sublenticularextendedamygdala.
*p(cid:6).05. **p(cid:6).01. ***p(cid:6).001.
types of self report. It is interesting that the single participant Becausepsychophysiologicalresponsestosexualstimulibegin
whoseself-reportedorientationdidnotmatchhispatternofbrain inthebrain,neurologicalactivitymaybeamoreeffectivewayto
activation provided unusually positive ratings to nonpreferred assessthebasisofthearousalstatethanperipheralmeasures.The
stimuli. A recent study showed that a majority of bisexually “gold standard” measure of sexual arousal in men has been the
identified men have a pattern of genital sexual arousal similar to penile plethysmograph, but this measure has the substantial limi-
that of homosexual men (Rieger, Chivers, & Bailey, 2005). For tationthatapproximatelyonethirdofmendonothavesufficient
this participant, too, self-identified sexual orientation may reflect erections for valid measurement (e.g., Chivers et al., 2004). Our
something other than relative degrees of arousal to male and results suggest that fMRI can be a useful psychophysiological
femalestimuli. measure of sexual arousal. A relatively imprecise method of a
priori ROIs drawn on an average brain achieved a degree of
sensitivityandspecificitycomparablewiththatobtainedbyother
physiological measures of arousal (Blanchard, Klassen, Dickey,
Kuban, & Blak, 2001). These results indicate that neurological
measuresmayprovideausefulmeansofassessingpopulationsin
whicherectilemeasuresarenotanoptionandmayhavepotential
clinical applications in the diagnosis of sexual dysfunctions and
theevaluationoftreatmentefficacy.
PredictivepowermaybeimprovedinfuturefMRIstudies.For
example, a block design using either sexual pictures with longer
stimulusonsetsorvideoeroticamaybemoreeffectiveatproduc-
ing robust arousal responses. Furthermore, functional ROIs from
studiesusingsimilarpreferredversusnonpreferredsexualstimuli
could be used to choose regions of greatest sensitivity and thus
providemorepreciseestimatesofarousalresponses.
References
Addis, D. R., Moscovitch, M., Crawley, A. P., & McAndrews, M. P.
(2004). Recollective qualities modulate hippocampal activation during
autobiographicalmemoryretrieval.Hippocampus,14,752–762.
Aharon,I.,Etcoff,N.,Ariely,D.,Chabris,C.F.,O’Connor,E.,&Breiter,
H. C. (2001). Beautiful faces have variable reward value: FMRI and
behavioralevidence.Neuron,32,537–551.
Figure 4. Average amygdala activation ((cid:2) SE) to male versus female Arnow,B.A.,Desmond,J.E.,Banner,L.L.,Glover,G.H.,Solomon,A.,
stimuli, separately by stimulus type, for homosexual and heterosexual Polan, M. L., et al. (2002). Brain activation and sexual arousal in
participants. healthy,heterosexualmales.Brain,125,1014–1023.