PR

IFY

SG

OL

BA

NG

OR

/ B

AN

GO

R U

NIV

ER

SIT

Y

Functional Specificity and Sex Differences in the Neural CircuitsSupporting the Inhibition of Automatic ImitationDarda, Kohinoor Monish; Butler, Emily; Ramsey, Richard

Journal of Cognitive Neuroscience

DOI:10.1162/jocn_a_01261

Published: 01/06/2018

Peer reviewed version

Cyswllt i'r cyhoeddiad / Link to publication

Dyfyniad o'r fersiwn a gyhoeddwyd / Citation for published version (APA):Darda, K. M., Butler, E., & Ramsey, R. (2018). Functional Specificity and Sex Differences in theNeural Circuits Supporting the Inhibition of Automatic Imitation. Journal of CognitiveNeuroscience, 30(6), 914-933. https://doi.org/10.1162/jocn_a_01261

Hawliau Cyffredinol / General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/orother copyright owners and it is a condition of accessing publications that users recognise and abide by the legalrequirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of privatestudy or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access tothe work immediately and investigate your claim.

09. Oct. 2020

Journal:JoCN

Title:FunctionalSpecificityandSexDifferencesintheNeuralCircuitsSupportingthe

InhibitionofAutomaticImitation

Runningtitle:Functionalspecificityandautomaticimitation

KohinoorM.Darda*,EmilyE.Butler*andRichardRamsey

WalesInstituteforCognitiveNeuroscience,SchoolofPsychology,BangorUniversity,

Bangor,Gwynedd,Wales,LL572AS,UnitedKingdom

Correspondingauthor:r.ramsey@bangor.ac.uk

*=Jointfirstauthors

Keywords:fMRI;domainspecificity;imitation-inhibition.

2

Abstract

Althoughhumansshowaninvoluntarytendencytocopyotherpeople’sactions,which

buildsrapportbetweenindividuals,wedonotcopyactionsindiscriminately.Instead,

copyingbehavioursareguidedbyaselectionmechanism,whichinhibitssomeactionsand

prioritisesothers.Todate,theneuralunderpinningsoftheinhibitionofautomaticimitation

anddifferencesbetweenthesexesinimitationcontrolarenotwellunderstood.Previous

studiesinvolvedsmallsamplesizesandlowstatisticalpower,whichproducedmixed

findingsregardingtheinvolvementofdomain-generalanddomain-specificneural

architectures.Here,weuseddatafromExperiment1(N=28)toperformapoweranalysisto

determinethesamplesizerequiredforExperiment2(N=50;80%power).Using

independentfunctionallocalisersandananalysispipelinethatbolsterssensitivity,during

imitationcontrolweshowclearengagementofthemultiple-demandnetwork(domain-

general),butnosensitivityinthetheory-of-mindnetwork(domain-specific).Weaker

effectswereobservedwithregardtosexdifferences,suggestingthattherearemore

similaritiesthandifferencesbetweenthesexesintermsoftheneuralsystemsengaged

duringimitationcontrol.Insum,neurocognitivemodelsofimitationrequirerevisionto

reflectthattheinhibitionofimitationreliestoagreaterextentonadomain-general

selectionsystemratherthanadomain-specificsystemsupportingsocialcognition.

3

Introduction

Humansocialinteractionsareguidedbynon-verbalcues,suchascopyingbehaviours.In

thelasttwodecades,muchresearchhasinvestigatedtheinvoluntarytendencytocopy

other'sactions-aphenomenonknownasautomaticimitation(Heyes2011).Automatic

imitationisthoughttobebeneficialinsocialsituationsbecauseitdevelopsaffiliative

attitudes,betterco-operation,andfeelingsofclosenessbetweeninteractingpartners

(ChartrandandLakin,2013).Priorneuroscienceresearchhasshownthatimitationis

supportedbythemirrorneuronsystem(MNS),aneuralnetworkengagedinperceivingand

performingactions(Iacobonietal.1999;Iacoboni2009;RizzolatiandCraighero2004).

Imitation,however,isunlikelytorelyonasinglecognitiveorbrainsystem(Southgateand

Hamilton2008).Forexample,inmanycircumstances,imitationismaladaptive,and

requiresinhibition(Newman-Norlundetal.2007;vanSchieetal.2008;Crossetal.,2013;

Cross&Iacoboni,2014).Insuchsituations,aselectionmechanismisrequiredtosuppress

thetendencytoimitateandprioritisealternativeactions(Brassetal.2009).Todate,

studiesinvestigatingtheneuralmechanismsofimitationcontrolhavebeenlimitedbysmall

samplesizesandlowstatisticalpower,whichhasproducedmixedfindings(Table1).

Further,noneuroscienceresearchhasinvestigatedhowindividualdifferencessuchassex

modulateimitationcontrol,eventhoughbehaviouralresearchhasshownthatimitative

tendenciesvaryasafunctionofsex(Butler,Ward,andRamsey2015;Dimberg1990;

Sonnby-Borgströmetal.2008).Acrosstwofunctionalmagneticresonanceimaging(fMRI)

experiments,whichhadhigherstatisticalpowerandfunctionalsensitivitythanprior

studies,weinvestigatedtheextenttowhichimitationinhibitionreliesonadomain-specific

ordomain-generalneuralnetwork,whichvariesitsresponseasafunctionofsex.

4

Muchlikecognitivescienceingeneral(HirschfeldandGelman1994;Kanwisher

2010),inhibitorycontrolresearchhasfocusedonaneatdivisionbetweendomain-general

anddomain-specificmentaloperations.Domain-generalinhibitorysystems,whichoperate

acrossmultipletasks,havebeenidentifiedindorsalfrontoparietalcortices(Aronetal.

2014;Bungeetal.2002;Hazeltineetal.2007;Neeetal.2007;Wageretal.2005).This

braincircuithasbeenlabelledthemultipledemandnetwork(MD)network,duetoits

engagementinadiversityofmentaloperations(Duncan,2010).Bycontrast,evidencefrom

fMRI,neurostimulationandneuropsychologicalpatientstudieshassuggestedthata

domain-specificcircuitinananteriorportionofmedialprefrontalcortex(mPFC)andright

temporoparietaljunction(rTPJ)operatesduringtheinhibitionofimitation(Brassetal.,

2001;2003;Klapperetal.,2014;Spengleretal.,2009;2010;Santiestebanetal.,2012;

Wangetal.,2011;Hogeveenetal.,2014;Santiestebanetal.,2015;Sowden&Catmur,2015;

Bardietal.,2017).Beyondthecontrolofimitation,mPFCandrTPJhavebeenconsistently

implicatedinavarietyofsocialcognitionfunctions,whichrequiredistinguishingbetween

selfandother,aswellasreasoningaboutotherpeople’smentalstates(TheoryofMind,

ToM;Amodio&Frith,2006;Frith&Frith,1999;vanOverwalle,2009;Saxe&Kanwisher,

2003).Theseresultsledtotheorisingthatakeyneuralcircuitforsocialcognitionalso

regulatesimitativetendencies(Brassetal.,2009).

Althoughtheoriesofimitationcontrolhavedevelopedthatarebasedonfunctioning

ofthetheoryofmindnetwork,evidencefromfMRIstudiesthatusedareactiontime

measureofimitationinhibitionhavenotprovidedconsistentsupportfortheinvolvement

ofadomain-specificneuralnetwork(Table1).Thereactiontimemeasureofimitation

involvesmakingfingermovementswhilesimultaneouslywatchingcompatibleor

5

incompatiblefingermovements(Brassetal.,2000).Thedifferencebetweenreactiontimes

inthesetwoconditions(i.e.thegeneralcompatibilityeffect)hasbeenarguedtoindex

imitativecontrolasgreatercognitiveresourcesarerequiredtoinhibitmovementsthatare

incompatibletoone’sownresponses(Brass&Heyes,2005;Heyes,2011).Approximately

halfofthefMRIstudiesusingthisparadigmfailedtofindengagementofrTPJandanterior

mPFC.Inaddition,anumberofstudiesshowedengagementofregionsassociatedwiththe

MDnetworkincludingdorsalfrontoparietalcortex,supplementarymotorarea(SMA)and

anteriorinsula(Bien,Roebroeck,Goebel,&Sack,2009;Crescentini,Mengotti,Grecucci,&

Rumiati,2011;Cross&Iacoboni,2013;Marsh,Bird,&Catmur,2016;Mengottietal.,2012).

Moreover,themostcommonmeasureofimitationinterferenceisconfoundedbyspatial

compatibilityorthetendencytorespondfastertoastimuluswhenitisonthesamesideof

spaceastheresponse(e.g.Simon1969).Inordertomeasureimitationinterference

independentofspatialcompatibilityeffects,spatialandimitativeprocessesneedtobe

dissociated(Berthentaletal.2006;Boyeretal.2012;CatmurandHeyes2011;Cooperetal.

2012;Gowenetal.2016;Wiggettetal.2011;Marshetal.2016).Therefore,theextentto

whichimitationinhibitionreliesondomain-specificanddomain-generalarchitectures

remainsunclear.Indeed,noresearchtodatehasdissociatedspatialfromimitative

processesandusedafunctionalregionofinterestapproach(fROI;Kanwisher2010

Fedorenkoetal.2013).UsingafROIapproachenablesinvestigationofhowfunctionally-

definedbraincircuits,suchastheMDandToMnetworks,operateduringthecontrolof

imitation.

Afurtherareaofimitationresearchthathasreceivedlittleattentionistheextentto

whichimitativecontrolvariesacrossindividuals,especiallybetweenthesexes.Ithasbeen

6

arguedthatimitationismodulatedbystableindividualdifferencessuchasempathy

(ChartrandandLakin2013)andsex(Butleretal.2015;Sonnby-Borgströmetal.2008).

Althoughithasbeensuggestedthatwomenexcelacrossarangeofsocialprocesses

comparedtomen(Baron-Cohen2002),onlyalimitednumberofstudieshaveinvestigated

sexdifferencesinsocialcognitionandtheresultsareoftenmixed,donotreplicate,orare

specifictoveryselectcontextsorsamples(Hyde2014;MillerandHalpern2014).Further,

studiesofsexdifferencesinsocialcognitionhavemainlyfocusedonemotionalexpression

perceptionandmentalstatereasoningwithlittleemphasisplacedonimitation(Campbell

etal.,2002;Thayer&Johnsen,2000;Rahmanetal.,2004;Krachetal.,2009;Russelletal.,

2007).

Arecentstudythatusedareactiontimemeasureofimitationinhibition(Brassetal,

2000),showedthatfemalesshowedagreaterlevelofinterferencethanmales(Butler,

Ward,&Ramsey,2015).Itispossiblethatthissexdifferenceinimitationcontrolmaybe

mediatedbyempathy-femaleshavebeenshowntobemoreempatheticcomparedtomales

(Baron-Cohen&Wheelwright,2004;Christov-Mooreetal.,2014).However,eventhough

empathyhasbeenassociatedwithdifferenttypesofimitationparadigms(Chartrand&

Bargh,1999;Mülleretal.,2013;Sonnby-Borgström,Jönsson,&Svensson,2003;Soonby-

Borgström,2002),theevidencetodatesuggeststhatthereisnolinkbetweenimitation,as

measuredbyreactiontimes,andempathy(Butleretal.,2015;Genschowetal.,2017).In

addition,inthestudybyButlerandcolleagues(2015),itisunclearwhethersexmodulates

thetendencytoautomaticallyimitateorthetendencytoautomaticallyrespondinthesame

spatiallocationtotheobservedaction.Theformerindicatesasexdifferencethatis

specificallytiedtoimitationcontrol,whilethelattermightindicateasexdifferencein

7

processesassociatedwithresolvingspatialconflict.Morerecentworkalsoshoweda

greaterinterferenceeffectforfemalescomparedtomales(Genschowetal.,2017),aswell

asgreatererrorratesforpredominantlyfemalesamplesthanmales(Craccoetal.,inpress).

TheimitationtaskusedbyGenschowandcolleagues(2017)wascontrolledforleft-right

spatialcompatibility,bypresentingthestimulushandorthogonaltotheresponse.Even

thoughthisshowsthatthesexdifferenceremainswhenspatialcompatibilityisreduced,it

doesnotruleoutthepossibilityoforthogonalspatialcompatibility(Weeks&Proctor,

1990).Moregenerally,sexdifferenceshavebeenfoundonawiderangeofinhibitory

controltasksincludingFlanker,Gaze-Cueing,Arrow-Cueing,Oddball,andSimontasks,

whereinfemaleshavebeenshowntorequiremorecognitiveresourcesthanmalesto

inhibitautomaticresponsetendencies(Figure1;Baylissetal.,2005;Claysonetal.,2011;

Rubiaetal.,2010;Stoet,2010;2017).Itispossible,therefore,thatadomain-generalsystem

mayunderpinthesexdifferencesobservedacrossthesetasks,includingduringimitation

control,butnoresearchtodatehasdirectlyinvestigatedthisproposal.

AcrosstwofMRIexperiments,thecurrentstudyinvestigatedfunctionalspecificity

andsexdifferencesinimitationcontrol.Severalaspectsoftheexperimentaldesignprovide

groundstoextendcurrentunderstandinginmeaningfulandconcreteways.First,thisisthe

firststudytouseindependentfunctionallocaliserstoidentifyMDandToMnetworksin

singlesubjectsanddirectlytesttheinvolvementofthesenetworksinimitationcontrol.By

doingso,wecandirectlytesthypothesesregardingtheroleoffunctionally-definedneural

circuits(i.e.,MDandToMnetworks)andthereforeminimisetherelianceonreverse-

inferencetoinfercognitivefunctionbasedonanatomicallocalisation(Poldrack,2006).

Second,weuseddatafromExperiment1toperformapoweranalysistodeterminethe

8

samplesizerequiredtoachieveadesiredlevelofpowerinExperiment2.Giventhe

inconsistentfindingsinpriorstudies,whichhadrelativelysmallsamplesizes,thismulti-

experimentapproachmadesurethatourkeyexperimenthadover80%powertodetect

expectedeffectsizes.Third,toavoidspatialcompatibilityconfounds,inExperiment2,we

usedamodifiedversionoftheimitationinhibitionparadigmthatallowedforan

independentmeasureofspatialandimitativecompatibility(Catmur&Heyes,2011).Ifthe

inhibitionofautomaticimitationreliesonadomain-specificneuralarchitecturethatis

associatedwithsocialcognition,asproposedbyBrassandcolleagues(Brassetal.,2009),

mPFCandrTPJwouldbeengagedinimitativecontrol.Incontrast,engagementoftheMD

networkwouldsuggestthatdomaingeneralprocessessubserveimitationcontrol.Further,

thesexdifferencefoundpreviously(Butleretal.2015)maybesupportedbydifferencesin

ToMorMDnetworks.

MaterialsandMethods

Overviewoftheexperimentalapproach

Experiment1usedagroup-levelwhole-brainanalysis,whichprovidedthebasisfor

poweranalysesthatsetupExperiment2asthecriticalexperimentwithhighstatistical

power(80%).InExperiment2,inordertoincreasesensitivityandfunctionalresolution,we

usedindependentlocaliserstoidentifykeyfunctionalcircuits(i.e.MDandToMnetworks)

andanalyseswereperformedinsinglesubjectstopreciselyquantifytheconsistencyof

networkengagementacrossindividuals(Kanwisher,2010;Nieto-Castanon&Fedorenko,

2012).Group-levelanalysesrequireresponsesacrossindividualstooverlapinindividual

voxels.Incontrast,thefROIapproachallowsidentificationofcorrespondingfunctional

9

regionswithouttherequirementofexactvoxeloverlapacrossindividuals.Therefore,the

exactsamevoxelsneednotbeactiveacrossindividuals,aslongasvoxelswithina

functionally-definedROIareconsistentlyactiveacrossindividuals.Consequently,group

levelanalysesmayunderestimatefunctionalspecificity,whereasfROIanalysescanshow

increasedsensitivity(Nieto-Castanon&Fedorenko,2012).Inaddition,duetoa

constrainedsearchvolume,fROIanalysestypicallyhavehigherstatisticalpowerthan

whole-brainanalyses(Fedorenkoetal.,2010;Saxeetal.,2006).

Experiment1

Participants

28participants(Mage=23.96,SDage=5.52;14females)participatedformonetary

compensationof£15.Participantsgaveinformedconsentinlinewiththeguidelinessetby

theResearchEthicsandGovernanceCommitteeoftheSchoolofPsychologyatBangor

University,wereright-handed,hadnormalorcorrected-to-normalvision,andreportedno

historyofneurologicaldamage.

Designandprocedure

Allparticipantsperformedtheimitationtaskinsidethescanner.Theparticipants

alsodidfouradditionaltasksinthesamescanningsessionaspartofanotherexperiment.

Thescanningsessionstartedwiththeimitationtask,followedbyarunofafaceperception

task,aflankertask(Erikson&Erikson,1974),anotherrunofthefaceperceptiontask,a

dynamicfacelocaliser(Pitcheretal.,2011),andatheoryofmindlocaliser(Dodell-Federet

al.,2011).Theorderofthetaskswascounterbalancedacrossparticipantssuchthatoutof

10

the28participants,14participantsdidtheimitationtaskfirst,and14participantsdidthe

flankertaskfirst,withtheorderoftheothertasksremainingthesame.

TheImitationInhibitionTask

Theimitationtaskwasbasedonastimulus-responsecompatibilityparadigm

developedbyBrassetal.(2000)consistingofobservationandexecutionoffingerlifting

movementsduringfMRIscanning.Beforethetask,participantswereinstructedtohold

downthe‘blue’and‘yellow’buttonsontheresponseboxwiththeirindexandmiddle

fingersrespectivelyoftherighthand.Anumbercue(either‘1’or‘2’)waspresentedto

participants,andtheywereaskedtolifttheirindexfingeronpresentationofthenumber‘1’

andthemiddlefingerforthenumber‘2.’Simultaneously,theyalsoviewedanimageofan

indexormiddlefingerliftofalefthandviewedfromthethirdpersonperspectivesuchthat

thefingersextendedtowardstheparticipants.Thus,therewerefourtrialtypesinanevent-

relateddesignthatledtotwoconditions–participantsperformingthesame(congruent)or

different(incongruent)fingermovementtotheobservedhandimage.

Eachtrialstartedwithafixationcross(500ms)followedbyaneutralhand(fora

randominter-stimulusintervalof500,700,or1000ms),andahandimagewithanindex/

middlefingerlift,whichstayedonscreenfor2000msirrespectiveofwhentheparticipant

madetheresponse.After2000ms,thenexttrialstartedimmediatelywithafixationcross

(500ms).Inordertoseparatelymodeltheinfluenceofindividualeventssetclosertogether

intimeseparatelyinanevent-relateddesign,thefourtrialtypeswerepseudorandomized

suchthateachtrialtypewasprecededbyeachothertrialtypeandbyitselfanequal

numberoftimes(Josephs&Henson,1999;Wager&Nichols,2003).Therewere17trialsin

eachblock.Thefirsttrialwasusedtosetuptherandomizationsequencebutexcludedfrom

11

theanalysisasitwasnotprecededbyanyothertrial.Theremaining16trialswithinablock

wereanalysedandconsistedof8trialspercondition.Eachrunconsistedof5blocks

separatedbya3,4,or5secondfixationcross.Allparticipantscompletedonerunofthe

imitationtask.Thus,therewere80trialsofinterest(40congruentand40incongruent).

BehaviouralDataAnalysis

Reactiontimeontheimitationinhibitiontaskwasmeasuredasthetimefromnumbercue

onsettowhenparticipantsmadearesponse.Toensureparticipantswereengaging

correctlywiththetask,participantswhohadlessthan80%accuracywereremoved.In

addition,RTsmorethan3SDawayfromthemeanwereexcludedfromtheanalyses.

Further,trialsonwhichparticipantsmadean‘error’wereexcludedfromtheanalyses.

Errorsincludedanincorrectresponse,noresponse,aresponseafter2000msandpressing

aninvalidkey.ThegeneralcompatibilityeffectwascalculatedastheRTdifferencebetween

incompatibleandcompatibletrials.Aone-samplet-testwasperformedtoverifythe

presenceofageneralcompatibilityeffect.Aone-tailedindependentsamplet-testswas

performedtodetermineifthecompatibilityeffectwasgreaterforfemalesthanmales.Mean

differences,95%confidenceintervals(CI),andCohen’sd(Cohen,1992)arereportedforall

effectsofinterest.Fortheonesamplet-test,Cohen’sdzwascalculatedasmeandifference

dividedbythestandarddeviationofthesample(Lakens,2013).Fortheindependent

samplest-test,Cohen’sdwascalculatedasmeandifferencebetweenthetwogroups

dividedbythepooledstandarddeviation(Cohen,1992).

fMRIDataAnalysis

Dataacquisition

12

Participantswereplacedsupineina3-TeslaPhilipsMRIscannerusingaSENSE32

channelphasedarraycoil.Theywererequestedtoavoidheadmotionduringthescanning

sessionandwerepresentedstimulionacomputerscreenplacedbehindthescannermade

visiblebyamirrorattachedtotheheadcoil.Responsesonthetaskwererecordedwiththe

helpofabuttonboxthatrecordedreactiontimes.35axialsliceswereacquiredinan

ascendingorderusingaT2*-weightedEPIsequence.Thereferencesliceforslicetime

correctionwasthesliceacquiredinthemiddleofthesequence(slice17).Parametersareas

follows:voxelsize=3x3x4mm;repetitiontime(TR)=2000ms;echotime(TE)=30ms;

flipangle=90degrees;slicethickness=4mm;slicegap=0.8mm;FOV:230x230x167

mm3.174volumeswerecollectedfortheimitationtask.

Fourdummyscanscollectedatthebeginningofeachrunofthetaskwerenot

includedinanyanalyses.Ahigh-resolutionT1-weightedanatomicalimagewasalso

collectedwiththefollowingparameters:TR=12ms;TE=3.5ms,flipangle=8degrees;

numberofaxialslices=170;voxelsize=1mm3;FOV:250x250x170mm3.

DataPreprocessingandGeneralLinearModel

FunctionalimageswerepreprocessedinSPM-8.Datawererealigned,unwarped,and

correctedforslicetiming.DatawerenormalisedtotheMNItemplatewitharesolutionof

3mm3andimageswerespatiallysmoothed(8mm).

Fortheimitationtask,adesignmatrixwasfitforeachparticipantwiththree

regressors:oneeachforthecoorecttrialsofthetwoconditions,andoneforthe‘new’trials

(i.e.thefirsttrialofeachblock).Thenewtrialswerenotusedinanyfurtheranalyses.

Stimulusonsetsweretime-lockedtothepresentationofthenumbercuewithadurationof

zerosecondsandconvolvedwiththestandardhaemodynamicresponsefunction.

13

Whole-brainanalyses

Contrastimages(incompatible>compatible)werecalculatedatthesinglesubject

levelfortheimitationinhibitiontaskinordertoidentifyregionsofthebrainshowinga

compatibilityeffect.Grouplevelcontrastimageswerecreatedfromthesesingle-subject

contrastimagesinordertoidentifyregionsthatwereconsistentlyengagedforthe

compatibilityeffectacrossthesampleusingone-samplet-tests.Inordertoidentifyaneural

signatureofthesexdifferenceinimitationinhibition,asex*compatibilityANOVAwas

computed[Female(incompatible>compatible)>Male(incompatible>compatible)]as

femaleshavebeenshowntohaveahighercompatibilityeffectthanmalesintheimitation

task(Butleretal.,2015).Forallanalyses,contrastimagesweretakentothegroupleveland

thresholdedusingavoxel-levelthresholdofp<.001andavoxel-extentof10voxels.

Correctionformultiplecomparisonswasperformedattheclusterlevel(Fristonetal.,

1994)withclustersthatsurvivecorrectionformultiplecorrectionsusingafamily-wise

error(FWE)correction(p<.05)(showninboldfontinTable2(A)and(B);seeResults).

ClustersofactivitywereidentifiedwiththeSPMAnatomytoolbox(Eickhoffetal.,2005).

Experiment2

Participants

55participants(Mage=22.04,SDage=3.70;27females)wererecruitedfromthe

Bangorcommunityandwereeitherreimbursedwith£15or3coursecreditsfortheir

participation.InformedconsentwasobtainedinlinetheguidelinessetbytheResearch

EthicsandGovernanceCommitteeoftheSchoolofPsychologyatBangorUniversity.All

participantswereright-handed,didnothavedyslexiaordyspraxia,werenotonany

medication,didnotreportneurologicaldamage,andhadnormalorcorrected-to-normal

14

vision.ThesamplesizewasdeterminedbyapoweranalysisbasedonExperiment1data

(seeResults).

Designandprocedure

Eachparticipantperformedthreetasksinsidethescanner–theAutomaticImitation

task,aTheoryofMind(ToM)networklocalisertask,andaMultipleDemand(MD)network

localisertask.Theorderofthetaskswasasfollows:tworunsoftheMDnetworklocaliser

taskwereinterspersedbetweenthreerunsoftheimitationtaskinordertooffsetboredom,

followedbytworunsoftheToMtask.TheToMtaskwasalwayspresentedattheendto

reducethelikelihoodthatbeliefreasoningduringtheToMtaskwouldinfluence

performanceintheimitationtask.Theorderwasthesameforallparticipants.Participants

alsocompleteda50itemInternationalPersonalityItemPool(IPIP)questionnaire

(Donnellanetal.,2006;Goldberg,1992)(unrelatedtothecurrentstudy),andastimulus

ratingformwheretheywereaskedtoratethehandstimulusfromtheimitationtaskas

eithermale,female,orneutral.Theentiresessionlastedapproximately1.5hours,with60

minutesinsidethescanner.AllstimuluspresentationwascodedinMatlab2015b,and

presentedwithPsychToolBox3.0.6.

TheImitationInhibitionTask

TheautomaticimitationtaskwassimilartotheoneusedinExperiment1butwithtwo

changes.First,weusedadifferenthandstimulus,whichwasratedassex-neutralby

observers.Thesexofthehandwasanimportantconsiderationinordertominimisethe

possibilityofanown-sexbiaswhileexploringsexdifferencesinimitationinhibition.As

such,weconductedpilotworkthataskedobserverstoevaluatearangeofhandstimuliin

termsofmasculinityandfemininityandweselectedthemostsex-neutralstimulus(see

15

SupplementaryInformation,DevelopmentofStimuli).Weonlyusedonehandstimulusin

ordertosimplifythedesignspace.Althoughusingonesex-neutralhandstimulusprovided

greaterexperimentalcontrol,itmayhaveharmedourabilitytostudyorelicitsex

differences.Futureworkcouldprobethisfurtherbyvaryingthesexofthestimulusand/or

byusingmoresex-typicalstimuli.

Thesecondchangethatwemadewastocalculateanimitativecompatibilityeffect

independentofspatialcompatibility(Catmur&Heyes,2011).Todoso,participantsviewed

animageofanindexormiddlefingerliftofeitherarightorlefthand,butalwaysresponded

withtheirrighthand.Usingrightandleft-handimagesproducedeighttrialtypesandfour

mainconditionsofinterest(seeFigure2(A)).Forexample,whencuedtolifttheirindex

fingerwhileobservingaleft-handindexfingerlift,theobservedmovementisboth

imitativelycompatible(samefinger),aswellasspatiallycompatible(samesideofspaceto

theexecutedmovement).Incontrast,whenobservingaright-handindexfingerlift,the

participant’sresponseisimitativelycompatible(samefinger)butitisnotonthesameside

ofspace(theyarespatiallyincompatible).Thus,participantsperformedthesame

(imitativelycompatible)ordifferent(imitativeincompatible)fingermovementonthesame

(spatiallycompatible)ordifferent(spatiallyincompatible)sideofspacetotheobserved

fingermovement,givingrisetothefollowingfourconditions:

1. Imitativelyandspatiallycompatible

2. Imitativelyandspatiallyincompatible

3. Imitativelycompatibleandspatiallyincompatible

4. Imitativelyincompatibleandspatiallycompatible

16

Sequencinginformationandpseudo-randomisationwasthesameasExperiment1.There

were65trialsineachblock.Thefirsttrialwasusedtosetuptherandomizationsequence

butexcludedfromtheanalysisasitwasnotprecededbyanyothertrial.Theremaining64

trialswereanalysed,consistingof16trialspercondition.Eachrunconsistedof2blocks

separatedbya3secondfixationcross.Allparticipantscompletedthreerunsofthe

imitationtask.Intotal,therewere384trialsofinterest,96percondition.Experiment2,

therefore,hadmorethantwicethenumberoftrialsperconditionthanExperiment1.

Localisertasks

TheMultipleDemandNetworkLocaliser

InordertoidentifyregionsoftheMDnetwork,averbalWorkingMemory(WM)task

wasused(Fedorenko,etal.,2011).Participantswereaskedtorememberthesequencein

whicheitherfour(easycondition)oreight(hardcondition)digitsequenceswerepresented

onscreen(seeFigure2(B)).Aftereachtrial,participantshadtochoosebetweentwodigit

sequencespresentednumerically,oneofwhichmatchedthesequenceinwhichthedigits

werepresentedaswords.Feedbackwasprovidedastowhethertheyansweredcorrectlyor

incorrectly.Thehard>easycontrasthasbeenfoundtorobustlyactivateregionsoftheMD

network(Fedorenkoetal.,2011;2013).Eachrunconsistedof10experimentalblocks(each

34secondslong)and6fixationblocks(each16secondslong).Thetotalexperimentlasted

for436seconds.EachparticipantcompletedtworunsoftheWMtask.

TheTheoryofMindLocaliser

Inordertolocalisebrainregionsinvolvedinmentalstatereasoning,weuseda

paradigmdevelopedbyDoddell-Federandcolleagues(2011;

http://saxelab.mit.edu/superloc.php).Thislocalisertask(seeFigure2(C))includes20

17

stories,eachdescribingafalserepresentation.10storiesincludedout-of-datebeliefs(the

falsebeliefcondition)andtheother10includedout-of-datephysicalrepresentations

(photographs/maps;thefalsephotographcondition).TheFalseBelief>FalsePhotograph

contrasthasbeenshowninpriorworktorobustlyactivateregionsinvolvedinmentalising

(Dufouretal.,2013).Alltrialsconsistedofastory(10s),followedbyatrueorfalse

question(4s).Eachstorywasseparatedbya12-srestperiod.Theorderofthestoriesand

conditionswasthesameforallparticipants.Eachparticipantcompletedtworunsofthis

task:5trialsofeachconditionwerepresentedineachrun.

BehaviouralDataAnalysis

RTandaccuracywererecordedinthesamewayasExperiment1.Compatibility

effectswerecalculatedasfollows:spatialcompatibility=spatiallyincongruenttrials–

spatiallycongruenttrials;imitativecompatibility=imitativelyincongruenttrials–

imitativelycongruenttrials.Behaviouraldatawereanalysedinthesamefashionas

Experiment1onlyseparatelyforimitativeandspatialcompatibilityeffects.Themainaim

oftheexperimentwastotestforthepresenceofimitativeandspatialcompatibilityeffects,

aswellasfordifferencesbetweenthesexes(females>males).Hence,weusedaone-

samplet-testinordertoverifythepresenceofspatialandimitativecompatibilityeffects,

andaone-tailedindependentsamplest-testinordertotestwhetherfemalesshoweda

higherspatial/imitativecompatibilityeffectthanmales.

fMRIDataAnalysis

DataAcquisition

18

DataacquisitionprocedureswerethesameasExperiment1.Therewere249

volumescollectedfortheimitationtask,219fortheMDnetworklocaliser,and136forthe

ToMlocaliserforeachrun.

DataPreprocessingandGeneralLinearModel

AllMRIdatawerepreprocessedinSPM-8.Datawererealigned,unwarped,and

correctedforslicetiming.DatawerenormalisedtotheMNItemplatewitharesolutionof

3mm3.Normalisingtoacommonspaceinsteadoftheindividual’snativeanatomicalspace,

allowsforcomparisonswithpreviousstudies(relyingonthecommonspace)andis

preferredwhendefinitionoffROIsisbasedongroup-constrainedfunctionaldata(Nieto-

Castanon&Fedorenko,2012).Imageswerespatiallysmoothed(8mm).

Fortheimitationtask,adesignmatrixwasfitforeachparticipantwithfive

regressors:oneeachforthecorrecttrialsofthefourconditions,andonefor‘new’trials(i.e.

thefirsttrialofeachblock).Stimulusonsetsweretime-lockedtothepresentationofthe

numbercuewithadurationofzerosecondsandconvolvedwiththestandard

haemodynamicresponsefunction.Contrastimageswerecalculatedforeachindividual

subjectinordertoidentifyregionsofthebrainshowingaspatial(spatiallyincompatible>

spatiallycompatible)orimitative(imitativelyincompatible>imitativelycompatible)

compatibilityeffect.

Forthelocalisertasks,thedesignmatrixconsistedofregressorsforeach

experimentalcondition(‘Belief’and‘Photo’fortheToMlocaliser,and‘Hard’and‘Easy’for

theMDlocaliser).Theonsetanddurationofeachconditionwasspecifiedandconvolved

withthestandardhaemodynamicresponsefunction.Contrastimageswerethencalculated

19

foreachindividualsubjectinordertoidentifyregionsthatrespondedtocognitivedemand

(Hard>Easy)andmentalising(Belief>Photo).

Definitionofgroup-constrainedsubject-specific(GSS)analyses

FortheGSSanalyses,thespm_sstoolboxwasused,whichrunsinSPMusingMatlab

(http://web.mit.edu/evelina9/www/funcloc.html).TheGroup-constrainedSubject-Specific

(GSS)approachdevelopedbyFedorenkoetal.(2010)andJulianetal.(2012)wasusedto

definefunctionalregionsofinterestforeachparticipant.ThesefROIsweredefinedusing1)

eachindividual’sactivationmapforthelocalisertasks,and2)group-constraintsormasks.

Thesemasksrefertoasetof“parcels”,whichdemarcateareasinthebrainwhereprior

workhasbeenshowntoexhibitactivityforthelocalisercontrasts.

TwosetsoffROIsweredefined(Figure3):MDnetworkfROIsthathavebeenknown

toexhibitactivityforavarietyofcognitivecontroltasks(Duncanetal.,2010;Fedorenkoet

al.,2013)andToMnetworkfROIsthatsupportmentalisingandhavebeenspecifically

implicatedforimitationinhibition(Saxe&Kanwisher,2003;Brassetal.,2009).Forthe

ToMnetwork,fourparcelswerederivedfromagroup-levelmapfrom462participantsfor

theFalseBelief>FalsePhotographcontrast(Dufouretal.,2013).Theseregionsincluded

thedorsal,medial,andventralprefrontalcortex(DMPFC,MMPFC,VMPFC),andtheright

temporo-parietaljunction(rTPJ).FortheMDnetwork,weused16parcelsderivedfroma

setoffunctionalparcelscreatedbyIdanBlankbasedonaprobabilisticoverlapmapfrom

197participants(availableat:https://evlab.mit.edu/funcloc/download-parcels).These

includedareasinbilateralsuperiorandinferiorparietallobules(SPL,IPL),inferiorparietal

sulcus(IPS),inferiorandmiddlefrontalgyrus(IFG,MFG),precentralgyrus(PrecG),insula

(Ins),andthesupplementarymotorarea(SMA).Theseareaswerechosenfortworeasons:

20

1)theywerepartoftheMDnetwork(Fedorenkoetal.,2013)and2)theyhavebeenshown

torespondinpriorworktothespecifictypeofinterferencecontrolofrelevancetothe

currentstudy(Marshetal.,2016;alsoseeExperiment1).

Foreachindividual,thesemaskswereusedtoconstraintheselectionofsubject-

specificfROIs.Foreachindividual,fortheToMnetworkmask,theBelief>Photocontrast

wasusedandthetop10%ofvoxels(basedont-values)withineachparcelweredefinedas

thatindividual’sfROI.Similarly,fortheMDnetworkmask,eachindividual’stop10%of

voxels(basedont-values)intheHard>Easycontrastweredefinedasthatindividual’sfROI.

Usingthetop10%ofvoxels,ratherthanafixedthreshold(e.g.,allvoxelswithp<0.001),

ensuresaconstantsizeofeachfROIacrossindividuals(Blank,Kanwisher,&Fedorenko,

2014).Wealsorantheanalysesusingafixedthreshold(p<0.001,uncorrected)andfound

thesamepatternofresults(seeSupplementaryTablesS1.1andS1.2).Allanalysesreported

belowarebasedonthetop10%ofvoxelsbasedonthelocaliserdata.Percentsignalchange

(PSC)valueswereextractedfromallfROIs.Forthemainanalysis,allrunsofthelocaliser

taskswereusedtodefinefROIsineachindividual.ResponsesinthesefROIswereestimated

forspatialandimitativecompatibilityeffects.

Inasupplementaryanalysis,responsestothelocalisercontrastswerealso

estimatedinordertoensurethatallthefROIsshowedtheexpectedresponsewithrespect

tothelocalisercontrastsi.e.theToMnetworkshowedarobustBelief>PhotoandtheMD

networkshowedarobustHard>Easyeffect.Fortheselocaliseranalyses,anacross-runs

crossvalidationapproachwasused(Nieto-Castanon&Fedorenko,2012)toensurethat

datausedfordefiningfROIswasindependentofdatausedforestimatingresponse

(Kriegeskorteetal.,2009).

21

AsimplementedinGSS,statisticaltestswereperformedonthePSCvaluesusing

standardStudent’st-tests.One-samplet-testswereperformedtoinvestigatetheresponse

oftheMDandToMnetworkfROIstospatialandimitativecompatibilityeffects.Basedon

priorbehaviouralfindings,whichshowedgreaterRTinterferenceforfemalesthanmales

duringimitationinhibition(Butleretal.,2015),weexpectedtoobservesexdifferencesin

thoseregionsthatalsoshowsimplecompatibilityeffects.Thatis,weexpectedbrainregions

thatweregenerallyinvolvedinspatialand/orimitativecontroltoshowsexdifferences.As

such,weonlyinvestigatedsexdifferenceinthosefROIsthatshowedspatialorimitative

compatibilityeffects.Todoso,one-tailedindependentsamplest-testswereperformedthat

testedforgreaterengagementforfemalesthanmales.FalseDiscoveryRate(FDR)multiple

comparisoncorrection(p<.05)wasusedtocorrectforthenumberoffROIsineach

functionalnetwork.

Results

Experiment1

Behaviouralresults

Aone-samplet-testconfirmedageneralcompatibilityeffect(Mean=80.02,SE=

8.19),t(27)=9.77,p=<.001,95%CI(63.22,96.82),Cohen’sdz=1.85.Aone-tailed

independentsamplest-testshowednodifferencesbetweenmales(Mean=70.94,SE=

13.30)andfemales(Mean=89.10,SE=9.43),t(26)=1.114,p=.138,95%CI(-45.96),

Cohen’sd=0.42.Allparticipantshad>80%accuracyandhenceallwereincludedinthe

analysis.Trialsonwhichparticipantsmadeanincorrectresponse(0.95%),didnotmakea

22

responseorrespondedafter2000ms(0.52%)orpressedaninvalidkeyorrespondedtoo

fast(0.09%)wereexcludedfromtheanalyses.

fMRIresults

Inawhole-brainanalysis,compatibilityeffects(generalincompatible>general

compatible)wereobservedindorsomedialfrontalcortexandbilaterallyindorsolateral

frontalandparietalcortices(Figure4(A);Table3(A)).Asmallvolumecorrection(SVC)

usingMDandToMnetworkparcelswasperformedinordertorestrictthesearchareato

ToMandMDnetworks.UsingtheMDnetworkSVC,resultsshowedwidespreadactivation

offrontalandparietalregions,whichsurvivedcorrectionformultiplecomparisons(Figure

(4A,Ci)).Incontrast,usingtheToMnetworkSVC,noclusterssurvivedcorrectionfor

multiplecomparisonandonlyrTPJshowedacompatibilityeffectatmorelenientthreshold

(p<.001,uncorrected)(seeSupplementaryTablesS2.1andS2.2).AnteriormPFCdidnot

showthegeneralcompatibilityeffectevenatthismorelenientthreshold.

Thesex*compatibilityinteractionrevealedclustersinleftsuperiorparietallobule

extendingintopostcentralgyrusandafurtherclusterinthecerebellum(Figure4(B);Table

3(B)).NoclustersemergedfollowingaSVCanalysisusingtheMDandToMnetworkmasks,

whichdemonstratesthattheclustersemergingfromthesex*compatibilityinteractiondo

notoverlapwiththeMDorToMnetworks(seeSupplementaryTablesS2.1andS2.2;Figure

4(Cii)and(Dii)).

PowerAnalysis

WesetupExperiment1asapilotinordertoestimatetheappropriatesamplesize

forourcriticalexperiment(Experiment2).Tothisend,apoweranalysiswasperformed

usingthefMRIpowersoftwarepackage(fMRIpower.org;Mumford&Nichols,2008).We

23

performedthepoweranalysisasfollows:first,awholebrainmapoftheimitationtask

generalcompatibilityeffect(Incompatible>Compatible)fromExperiment1wasentered

intofMRIpower.Next,twoROIswereidentified:theMDnetwork(Duncan,2010)andthe

ToMnetwork(Saxe&Kanwisher,2003).TheMDandToMnetworkmasksusedwerethe

sameasinExperiment2(seeMaterialsandMethods).Asrecommended,wecorrectedthe

alphavaluebythenumberofROIs(0.05/2=0.025)beforeperformingpoweranalyses

(Mumford,2012).

Resultsfromthesepoweranalysesshowedthattesting50participantsin

Experiment2wouldprovide80%powertodetecteffectsaslargeas(orlargerthan)the

averageeffectsizethatwasobservedacrossallnodesintheMDnetworkinExperiment1

(Cohen’sd=0.4,MeanSignalChange=0.23,SD=0.58).Wedidnothavethesamelevelof

powertodetectsmallereffectsthanthese,suchasthoseobservedintheToMnetworkin

Experiment1.Indeed,theeffectsintheToMnetworkinExperiment1weresosmallthat

wewouldhaveneededanimpracticallylargesamplesizetoachieve80%power.Assuch,in

Experiment2wedecidedtotestparticipantsuntilwehad50usabledatasets.

DesigndifferencesbetweenExperiments1and2areworthconsideringwhen

interpretingthesepowercalculationsbecausewemaybeunderestimatingthepowerofour

designinExperiment2.Thetoolboxusedtorunpowercalculations(fmripower.org)can

onlyestimatepowerforafutureexperimentwiththesamedesignasthecurrentdataset

(Mumford&Nichols,2008).However,thedesignsofExperiment1and2differedintwo

ways.First,Experiment1measuredageneralcompatibilityeffect,whereasinExperiment

2,webrokethiseffectdownintospatialandimitativecompatibilityeffects.Second,

Experiment2hadmorethandoubletheamountoftrialsperconditionasExperiment1.

24

Therefore,theprimarycontrastusedtodeterminepowerwasnotidenticaltothecontrast

usedinExperiment2,butduetoagreaternumberoftrialsperconditiontoestimatethe

effectsofinterest,wemayunderestimatepowerinExperiment2.Giventhelackofsex

differencesinExperiment1inourregionsofinterest,wedidnothavesufficientpowerto

convincinglyinvestigateneuraldifferencesbetweenmalesandfemalesinExperiment2.

However,givenouraprioripredictionsregardingsex,wecontinuetoreportsexdifference

analysesthroughoutthepaper.

Experiment2

BehaviouralResults

ThehandstimulususedinExperiment2fortheimitationinhibitiontaskwas

perceivedas‘neutral’bymostparticipants(Meanrating=5.20,SDrating=2.04;ratedonascale

of1to9,where1=mostmasculine,5=neutral,and9=mostfeminine).Toensure

participantswereengagingcorrectlywiththetask,runsonwhichparticipantshadlessthan

80%accuracy(2runsofoneparticipant)wereremoved.Inaddition,RTsmorethan3SD

awayfromthemean(2runsofoneparticipantandonerunofanotherparticipant)were

excludedfromtheanalyses.Further,trialsonwhichparticipantsmadeanincorrect

response(1.52%),didnotmakearesponseorrespondedafter2000ms(0.61%)orpressed

aninvalidkey(0.22%)werealsoexcludedfromtheanalyses.Figure5showstheimitative

andspatialcompatibilityeffectsforboththesexes.ForRTdataseeSupplementaryTable

S3.

Spatialcompatibility.Aone-samplet-testconfirmedaspatialcompatibilityeffect

(Mean=41.94,SE=2.87),t(54)=14.618,p=<.001,95%CI(36.19,47.69),Cohen’sdz=

1.97.Aone-tailedindependentsamplest-testevidencedagreaterspatialinterferenceeffect

25

forfemales(Mean=50.98,SE=3.67)ascomparedtomales(Mean=33.20,SE=3.75),t

(53)=-3.38,p=<.001,95%CI(24.01),Cohen’sd=0.91.

Imitativecompatibility.Aone-samplet-testshowedasignificantimitative

compatibilityeffect(Mean=15.37,SE=2.86),t(54)=5.37,p<.001,95%CI(9.63,21.11)

Cohen’sdz=0.72.Therewasnosignificantdifferencebetweenmales(Mean=15.62,SE=

4.39)andfemales(Mean=15.11,SE=3.73),t(53)=0.09,p=.930,95%CI(-15.78),

Cohen’sd=0.02.

fMRIResults

FiveparticipantswereexcludedfromthefMRIanalysesbecauseoflessthan80%

accuracyintworunsoftheimitationtaskandtheMDnetworklocalisertask(N=1),and

excessiveheadmotion(N=4;displacement>4mm)inallrunsoftheimitationtaskand/or

allrunsofeitherofthelocalisertasks.Thus,thefinalsampleconsistedof50participants

(Mage=22.26;SDage=3.71;24females).Fromthese50participants,twosessionsofthe

imitationtaskwerealsoexcludedforoneparticipantduetoexcessiveheadmotionandone

participant’sdataforonesessionoftheimitationtaskcouldnotbeusedbecausethedata

filewascorrupted.

Localisertasks

AllfROIsshowedthepredictedresponsestothelocalisercontrasts(asestimated

usingdatanotusedfordefiningROIs;seeMethods).AlltheMDNetworkfROIsshoweda

robustHard>Easyeffect(ts>9.13,ps<.0001)andToMNetworkfROIsshowedarobust

Belief>Photoeffect(ts>5.70,ps<.0001).ForresponsesforeachindividualfROI

separately,seeSupplementaryTablesS4.1(MD)andS4.2(ToM).

TheAutomaticImitationTask

26

GSSAnalyses

Figure6showsthemeanpercentsignalchangeforeachfROIoverbaselineinthe

MDandToMnetworksforspatialandimitativecompatibility.

MultipleDemandNetworkfROIs

SpatialCompatibility.All16fROIsoftheMDNetworkshowedaspatialcompatibility

effect(ts>1.8,ps<.04;Figure6(A),Table3),whichsurvivedcorrectionformultiple

comparisons(p<.05,FDRcorrected).ThemeanpercentsignalchangeacrosstheMD

networkforspatialcompatibilitywas0.70,SD=1.66,Cohen’sd=0.42.Nosignificant

differenceswerefoundbetweenmalesandfemalesinpercentsignalchangevaluesinany

ofthefROIs(ts<1.6,ps>.1),exceptrightSPLwhichapproachedsignificance(p=.062)

(Figure7(A)).

ImitativeCompatibility.Noneofthe16MDNetworkfROIsshowedanimitative

compatibilityeffect,whichsurvivedcorrectionformultiplecomparisons(allps>.05,FDR-

corrected).5MDnetworkfROISshowedanimitativecompatibilityeffectatanuncorrected

threshold(ts>1.95,ps<.05).ThesefROISincludebilateralIPL,bilateralIPS,andtheright

IFG(Figure6(A),Table3).FourfurtherfROISshowedanimitativecompatibilityeffectthat

approachedsignificance,whichincludedleftIFG(p=.07),rightSPL,rightMFGandright

PrecG(p=.06).ThemeanpercentsignalchangeacrosstheMDnetworkforimitative

compatibilitywas0.54,SD=2.06,Cohen’sd=0.26.Therewasnosignificantdifference

betweenmalesandfemalesinanyofthesefROIs(ts<1.5,ps>.08)(seeFigure7(B)).

TheoryofMindNetworkfROIs

NoneoftheToMnetworkfROIsshowedimitative(ts<1.3,ps>.50)orspatial(ts<

1.6,ps>.06)compatibilityeffects,evenatanuncorrectedsignificancethreshold(Figure6

27

(B),Table4).rTPJshowedaspatialcompatibilityeffectthatapproachedsignificance

(p=0.065).ThemeanpercentsignalchangeacrosstheToMnetworkforspatial

compatibilitywas-0.16,SD=1.88,Cohen’sd=-0.08,andforimitativecompatibilitywas-

0.32,SD=2.02,Cohen’sd=-0.16.

Whole-brainanalyses

Forcompleteness,andforuseinfuturemeta-analyses,wealsocomputedgroup-

levelwhole-brainanalysesseparatelyforgeneral,spatial,andimitativecompatibility

effects,aswellasforsex*compatibilityinteractions(seeSupplementaryTableS4).

Openscience

DataforExperiments1and2arefreelyavailableonlineincludingbehaviouralandfROI

data(osf.io/45x6z),aswellaswhole-braint-maps

(https://neurovault.org/collections/EPSIHNQQ/).

Discussion

Thecurrentstudyprovidesthemostrobustneuroimagingevidencetodateforalackof

functionalspecificityintheneuralcircuitssupportingtheinhibitionofautomaticimitation.

Withhigherstatisticalpowerandfunctionalsensitivitythanpriorstudies,acrosstwo

experimentstheresultsdemonstratethatimitationinhibitionengagesadomain-general

neuralnetworkasopposedtoabrainnetworkthatsupportssocialcognition.Assuch,

modelsofimitationcontrolneedupdatingtoincludeanincreasedrolefordomain-general

processesandareducedoralteredrolefordomain-specificprocesses.Further,intermsof

behaviour,femalesshowedahigherspatialbutnotimitativecompatibilityeffectthan

males.However,therewasnosexdifferenceintheneuralmechanismsunderlyingspatial

28

orimitationcontrol,whichsuggeststhatfurtherexplorationofsexdifferencesininhibitory

controlisrequired.

Functionalspecificityinimitationinhibition

Ourfindingsshowthatbrainregionsthatareengagedinaverbalworkingmemory

task,whichareassociatedwiththeoperationoftheMDnetwork(Duncan,2010;Fedorenko

etal.,2013),arealsoengagedduringspatialandimitativeconflictresolution.Theseresults

supporttheinvolvementofadomain-generalcognitiveandneuralsystemduringthe

controlofimitation.Bycontrast,brainregionsthatareengagedinabeliefreasoningtask,

whichareassociatedwiththeoperationofthetheoryofmindnetwork(Frith&Frith,1999;

Saxe&Kanwisher,2003;vanOverwalle,2009),shownoengagementduringtheinhibition

ofimitation.Assuch,weprovidenoevidencefordomain-specificityincognitiveandneural

systemsthatcontrolimitation.

Brassandcolleagues(2009)proposedthatinthecontextofimitationcontrol,rTPJis

involvedinself/otherdistinctionandmPFCenforcestheself-generatedactionoverthe

observedaction.Ourfindingsareinconsistentwiththehypothesisthataspecificneural

systemrelatedtosocialcognitionisengagedintheinhibitionofautomaticimitative

tendencies.mPFCandrTPJhavebothbeenimplicatedinimitationinhibitionbysome

studies(Brassetal.,2005;2009;Spengleretal.,2009;Wangetal.,2011).Incontrast,other

studiesfoundengagementofmPFConly(Brassetal.,2001;Crossetal.,2013)orofdomain

generalregionsratherthanmPFCandrTPJ(Bien,Roebroeck,Goebel,&Sack,2009;

Crescentini,Mengotti,Grecucci,&Rumiati,2011;Cross&Iacoboni,2013;Marsh,Bird,&

Catmur,2016).Inbothexperimentsinthecurrentstudy,wehadlargersamplesizesthan

priorexperimentsand,inExperiment2,wehadsufficientstatisticalpowertobeconfident

29

indetectingeffectsaslargeaspreviouslyobservedinmPFCandrTPJ,shouldtheyexist.

Takentogetherwithpriorfindings(Table1),wesuggestthatduringtheinhibitionof

imitation,theconsistencyofmPFCandrTPJengagementacrossindividualsislow,whereas

theconsistencyofMDnetworkengagementacrossindividualsisrelativelyhigh.

Theseresultshavepotentialimplicationsforself-othercontroltheoriesofsocial

cognitionmoregenerally.Mostlybasedonimitationresearch,whichpreviouslysuggested

thatmPFCandrTPJareengagedinimitationinhibition,self-othercontrolisthoughttobea

candidatemechanismforadiversesetofsocialfunctions(Brassetal.,2009;deGuzmanet

al.,2016;Sowden&Shah,2014).Forexample,self-othercontrolprocesseshavebeen

linkedtoautism,empathy,andtheory-of-mind(Spengleretal.,2009;deGuzmanetal.,

2016;Sowden&Shah,2014).However,recentbehaviouralfindings,whichusedlarger

samplesizesthanpriorworkandmeta-analyticalapproaches,donotsupporttheviewthat

thecontrolofimitationvariesasafunctionofsocialdispositionasindexedbyautistic-like

traitsandempathy(Butleretal.,2015;Craccoetal.,inpress;Genschowetal.,2017).In

lightoftheserecentbehaviouralresults,thelackofengagementofmPFCandrTPJinthe

currentstudyraisesanimportantquestionabouttherelianceofimitationinhibitionona

self-otherdistinction.Onepossibilityisthatinsteadofadistinctlysocialmechanism(Boyer

etal.,2012;Bertenthal&Scheutz,2013),inhibitingimitativetendenciesmayinvolvethe

samecognitiveprocessesthatareusedwheninhibitingothernon-socialexternalinfluences

(Heyes,2011;Cooperetal.,2013).

Alternatively,theengagementofmPFCandrTPJduringself-othercontrolprocesses

maybemorecomplicatedthanwhatcurrentmodelsofsocialcognitionsuggest.Indeed,a

smallnumberofneurostimulationstudieshaveshownthatmodulationtorTPJcan

30

influenceperformanceonRTmeasuresofimitation(Hogeveenetal.,2014;Sowden&

Catmur,2015).Inaddition,mPFCandrTPJhavebeenfoundtobeinvolvedinthe

modulationofautomaticimitation.Forexample,Klapperetal.(2014)foundahigher

responseinrTPJwhenaninteractionpartnerlookedhumanandwasbelievedtobehuman

comparedtowhenneitheroftheseanimacycueswerepresent.Wangandcolleagues

(2011)demonstratedthatmPFChadatop-downinfluenceonotherbraincircuitsduring

socialmodulationofimitationviadirectgaze.ThesestudiessuggestthatmPFCand/orrTPJ

mayhavearegulatoryrole,besensitivetosocialcontext,andbefunctionallyconnectedto

otherregionsduringtheinhibitionofautomaticimitation.Indeed,regionsthatdonotshow

directengagementinacognitiveprocessofinteresthavebeenknowntohavearegulatory

influenceonotherregionsthataredirectlyengaged(Burnett&Blakemore,2009).Inline

withthisproposal,Crossandcolleagues(2013)suggestedthatimitationcontrolinvolves

top-downregulationbetweenadomain-generalcognitivecontrolnetwork,andadomain-

specificnetworkrelevantforimitation.Moregenerally,researchfromotherdomainsof

socialcognitionshowsgrowingevidenceforhighercomplexityandfunctionalinterplay

withinandbetweenso-calleddomain-specificanddomain-generalnetworks(Baetensetal.,

2014;Spunt&Adolphs,2015;Quadfliegetal.,2011;Zakietal.,2010).Thesestudies

suggestthatmodelsincludingneatdivisionsbetweenthesenetworksmaybeanoverly

simplisticcharacterisationofmentalfunction(Barrett,2012;Michael&D’Ausilio,2014).

Muchlikesocialcognitioningeneral,therefore,imitationcontrolmaybebestexplainedby

interactionsbetweencomponentfunctionalcircuits,whichthemselvesneednotbedomain-

specific(Spunt&Adolphs,2017).Acrucialdirectionforfutureresearchistestingformore

31

complexmodelsofimitation,whichmayinvolveconnectivityinandbetweenregionsofthe

MDandToMnetworks.

Animportantpointtonote,however,isthatanyconclusionsmaderegarding

possibledomain-specificityofmPFCandrTPJarebasedontheassumptionthatmPFCand

rTPJareatleastpartlyspecialisedforsocialcognition(Brassetal.,2009).Recentevidence

suggeststhatmPFCandrTPJmaybefunctionallyversatileinthesensethattheyshow

generalcognitiveproperties,whichmaynotbespecifictosocialcognition(Alexander&

Brown,2011;Carter&Huettel,2013;delaVegaetal.,2016;Dugueetal.,2017;Schurzet

al.,2017;Schuwerketal.,2017;Yarkonietal.,2011).Thus,theargumentthatthe

engagementofmPFCandrTPJinimitation-inhibitionmaybespecifictosocialcognition

mightneedfurthervalidation.Additionally,socialcognitionitselfhasbeenbrokendownin

‘bottom-up’and‘top-down’domains(Zaki&Oschner,2012).Thebottom-updomainrefers

topre-reflectiveprocessesthatarefastandstimulusdrivenwhereasthetop-downdomain

mapsonreflective,cognitivelylaboriousandflexibleprocesses(Bohl&vanderBoss,

2012).Whenextendedtoimitationcontrol,priorresearchhasconsistentlyimplicated

regionsinvolvedintop-downcontrolforautomaticimitation(Brassetal.,2009).However,

recentstudiessuggestthatthatimitationcontrol(andsocialcognitionmorebroadly)relies

oninteractionsbetweenbottom-upandtop-downprocesses(Cross&Iacoboni,2014;

Christov-Moore,Conway,&Iacoboni,2017;Bohl&vanderBoss,2012).Thus,another

importantavenueforfutureresearchwouldbetoinvestigateimitationcontrolbasedon

bottom-upandtop-downprocessesandtheirinteractions,ratherthanconsideringthese

processesasmutuallyexclusive.

32

Nonetheless,resultsfromthecurrentstudyremainclear:thebasicimitation

inhibitionmechanismengagesthemultipledemandnetwork,whichhasbeenconsistently

associatedwithdomain-generalprocesses(Duncan,2010).Giventhemixedfindingsin

priorimitationstudies(Table1),aswellaspsychologyandneurosciencemoregenerally

(OpenScienceFramework,2015;Buttonetal.,2013),futurefMRIresearchmayalso

considerreliabilityandreproducibilityaskeyconcernsinimitationresearchandconsider

thepossibleuseoffROIasameanstoquantifyconsistencyacrossindividuals.

Sexdifferencesinimitationinhibition

Thisstudyisthefirsttoinvestigatesexdifferencesintheneuralmechanismsthat

inhibitimitation.Thebehaviouraldatademonstratedthatfemalesshowagreaterspatial

butnotimitativecompatibilityeffectthanmales.Thisresultextendspriorbehavioural

researchonsexdifferences,whichdidnotseparatespatial(ororthogonalspatial)from

imitativeresponsesinimitationcontrol(Butleretal.,2015;Genschowetal.,2017).The

resultisalsoconsistentwithreportsinawiderangeofnon-socialinhibitorycontroltasks,

whichshowsimilarsexdifferences(Figure2;Baylissetal.,2005;Claysonetal.,2011;Rubia

etal.,2010;Stoet,2010;2017).Allthesetasksshareacommonfeature–theyrequirethe

inhibitionofaresponsetoatask-irrelevantspatialfeatureinordertoenforceatask-

relevantresponse.Takentogether,thispatternofresultssuggeststhatresponseinhibition

relatingtospatialconflictdiffersbetweenthesexes,ratherthanaprocessthatistiedtothe

controlofimitation.Analternativepossibilityisthatthedifferencebetweenthesexesfor

spatialcompatibilityislargerthanforimitativecompatibility,andwewereunabletodetect

theimitativeeffectbehaviourally.Futureresearchwillhavetoprobethesepossibilities

further.

33

GiventheproposedroleofMDandToMnetworksinimitationcontrol,we

anticipatedsexdifferencesinoneorbothofthesenetworks.Theneuroimagingdata,

however,demonstratednosexdifferencesintheToMorMDnetworksineither

experiment.Further,eventhoughregionsoutsideofourregionsofinterestmediatedthe

sexdifferenceinExperiment1,theseregionswerenotconsistentlyengageddifferentlyfor

malesandfemalesinExperiment2.Thus,basedondataacrossbothexperiments,ourbest

estimateisthatunivariateanalyses,whichassessthemagnitudeofBOLDresponse,donot

showlargeeffectsofsexinMDorToMneuralnetworks.Thisbeingsaid,theredoesseemto

beatrendforgreaterengagementintheMDnetworkforfemalescomparedtomalesfor

bothspatialandimitativeeffects,butthisdoesnotsurviveourstatisticalthresholding

(Figure7).Asaconsequence,wearecautioustointerpretthisnullresultaswedidnothave

thesamelevelofstatisticalpowertodetectsexdifferencesaswedidtodetectsimple

compatibilityeffects.Indeed,itremainsapossibilitythatsmallunivariateeffectsexistor

thatthesexdifferenceisunderpinnedbymorecomplexneuralorganisation.Futurestudies

thatuseconnectivitymeasures(Sporns,2005)ormultivoxelpatternanalysis(Normalet

al.,2006;Kriegeskorte,2008)mayshowincreasedsensitivityandbebetterabletocapture

thecomplexityofneuralorganizationthatweareaimingtomeasure.

Limitations

Theprimarylimitationofthecurrentworkisthatwestudiedarelativelysimple

modelofbrainorganisationbasedonunivariatemeasures.Giventhemixedevidencefrom

priorstudiesregardingimitationcontrol(Table1),wefeltitwasanimportantsteptofirst

establishtheextenttowhichgeneralandspecificsystemswereengagedinaunivariate

34

manner.Bydoingso,weaimedtobuildanappropriatefoundationforfutureworktobuild

upon.Moreover,asweonlyidentifiedtheMDandToMnetworks,itispossiblethatneural

regionsoutsideourkeynetworksmayplayaroleinimitationinhibitionormediatethesex

differenceinspatialresponseinhibitionorimitationcontrol.Eventhoughourwhole-brain

analysesshowednoconsistenteffectsoutsideofourfROIs,thisonlyshowsthattherewas

nounivariateengagementofextendedbrainregions.Wethusacknowledgethatwehave

testedarelativelysimplemodelofbrainorganisationthatislikelytounderestimatethe

complexityofneuralprocessesassociatedwithsocialandcognitivemechanismssuchas

imitationcontrol.Futureworkmayconsiderinteractionsbetweengeneralandspecific

systemsandmorecomplex,multivariatemeasuresofbrainorganisationsuchasMVPA.

Asecondlimitationregardsthefunctionallocalisationapproachusedtoidentifythe

ToMandMDnetworksinExperiment2.ThevalidityofthefROIapproachisbasedon

assumptionsaboutthefunctionalprocessesthatareengagedbythelocalisersusedto

identifyfROIs.Forexample,differentToMlocalisersmayengagepartlynon-overlapping

aspectsoftheToMnetwork(Spunt&Adolphs,2014;Schaafsma,Pfaff,Spunt,&Adolphs,

2015).Therefore,ourconclusionsabouttheroleofToMandMDnetworksarelimitedto

thetypeoflocaliserparadigmsthatweusedinthecurrentstudy.Futureresearchthatuses

differentfunctionalpartitionsofthesenetworkswouldbeinstructive.

AfinalpotentiallimitationisthattheorderoftasksinExperiment2couldhave

influencedourresults.WeorderedthetaskssuchthatToMlocaliserwasalwaysperformed

attheend,butMDtaskwasinterspersedbetweenimitationsrunstooffsetboredom.We

arrangedblocksinthismannerbecausewewereprimarilyconcernedthataskingpeopleto

performabeliefreasoningtaskwouldintroduceasocialbiastotreattheperson(hand

35

image)inanartificiallymoresocial/beliefreasoningmannerduringtheimitation-

inhibitiontask.Wedidnotsharethesamelevelofconcernthatperformingamemorytask,

whichweusedtolocalisethedomain—generalsystem,wouldintroduceamemoryor

“cognitivecontrol”biastotheimitation-inhibitiontask.However,wecannotruleoutthe

possibilityinthecurrentexperimentthattheMDtaskinfluencedthewaytheimitationtask

wasperformed.Thisbeingsaid,wedidgetthesameresultsinExperiment1,whentheMD

taskwasnotperformedbeforetheimitationtask.Assuch,althoughpossible,wefindit

unlikelythattaskorderhadameaningfulimpactonourresultsinExperiment2.

36

Funding

ThisworkwasfundedbyagrantfromtheEconomicandSocialResearchCouncil(grant

number:ES/K001884/1toR.R.).WethankRuudHortensiusforcommentsonaprevious

versionofthemanuscript.

References

Alexander,WH,&Brown,JW.2011.Medialprefrontalcortexasanaction-outcome

predictor.NatureNeuroscience,14(10),1338–1344.

AmodioDM,FrithCD.2006.Meetingofminds:themedialfrontalcortexandsocial

cognition.NatRevNeurosci.7:268-277.

AronAR,RobbinsTW,PoldrackRA.2014.Inhibitionandtherightinferiorfrontalcortex:

onedecadeon.TrendsCognSci.18:177-185.

Bardi,L.,Gheza,D.,&Brass,M.2017.TPJ-M1interactioninthecontrolofshared

representations:newinsightsfromtDCSandTMScombined.NeuroImage,146:734-

740.

Baron-CohenS.2002.Theextrememalebraintheoryofautism.TrendsCognSci.6:248-

254.

Baron-Cohen,S.,&Wheelwright,S.(2004).Theempathyquotient:aninvestigationofadults

withAspergersyndromeorhighfunctioningautism,andnormalsex

differences.Journalofautismanddevelopmentaldisorders,34(2),163-175.

BarrettHC.2012Ahierarchicalmodeloftheevolutionofhumanbrainspecializations.Proc

NatlAcadSci.109:10733-10740.

BaylissAP,diPellegrinoG,TipperSP.2005.Sexdifferencesineyegazeandsymboliccueing

ofattention.TheQuarterlyJournalofExperimentalPsychology.58A:631-650.

BertenthalBI,LongoMR,KosobudA.2006.Imitativeresponsetendenciesfollowing

observationofintransitiveactions.JournalofExperimentalPsychology:Human

PerceptionandPerformance.32:210.

38

Bertenthal,B.I.,&Scheutz,M.(2013).Inpraiseofamodelbutnotitsconclusions:

CommentaryonCooper,Catmur,andHeyes(2012).Cognitivescience,37(4),631-

641.

BienN,RoebroeckA,GoebelR,SackAT.2009.Thebrain'sintentiontoimitate:the

neurobiologyofintentionalversusautomaticimitation.CerebCortex.19:2338-51.

BlankI,KanwisherN,FedorenkoE.2014.Afunctionaldissociationbetweenlanguageand

multiple-demandsystemsrevealedinpatternsofBOLDsignalfluctuations.J

Neurophysiol.112:1105-18.

Bohl,V.,&vandenBos,W.2012.Towardanintegrativeaccountofsocialcognition:

marryingtheoryofmindandinteractionismtostudytheinterplayofType1and

Type2processes.FrontiersinHumanNeuroscience,6:274.

BoyerTW,LongoMR,BertenthalBI.2012.Isautomaticimitationaspecializedformof

stimulus–responsecompatibility?Dissociatingimitativeandspatialcompatibilities.

ActaPsychologica.139:440-8.

BrassM,BekkeringH,WohlschlagerA,PrinzW.2000.Compatibilitybetweenobservedand

executedfingermovements:comparingsymbolic,spatial,andimitativecues.Brain

andCognition,44:124-143.

BrassM,ZyssetS,vonCramonDY.2001.Theinhibitionofimitativeresponsetendencies.

NeuroImage,14:1416-1423.

BrassM,DerrfussJ,Matthes-vonCramonG,vonCramonDY.2003.Imitativeresponse

tendenciesinpatientswithfrontalbrainlesions.Neuropsychology,17:265-271.

BrassM,DerrfussJ,vonCramonDY.2005.Theinhibitionofimitativeandoverlearned

responses:afunctionaldoubledissociation.Neuropsychologia,43:89-98.

39

Brass,M.,&Heyes,C.2005.Imitation:iscognitiveneurosciencesolvingthecorrespondence

problem?TrendsCognSci,9:489-495.

BrassM,RubyP,SpenglerS.2009.Inhibitionofimitativebehaviourandsocialcognition.

PhilosTransRSocLondBBiolSci.364:2359-67.

BungeSA,HazeltineE.ScanlonMD,RosenAC,GabrieliJDE.2002.Dissociablecontributions

ofprefrontalandparietalcorticestoresponseselection.NeuroImage.17:1562-1571.

ButlerEE,WardR,RamseyR.Investigatingtherelationshipbetweenstablepersonality

characteristicsandautomaticimitation.PloSone.10:e0129651.

BurnettS,&BlakemoreSJ.2009.Functionalconnectivityduringasocialemotiontaskin

adolescentsandinadults.EuropeanJournalofNeuroscience.29:1294–301.

ButtonKS,IoannidisJP,MokryszC,NosekBA,FlintJ,RobinsonES,MunafòMR.2013.Power

failure:whysmallsamplesizeunderminesthereliabilityofneuroscience.NatRev

Neurosci.14:365-376.

CampbellR,ElgarK,KuntsiJ,AkersR,TersteggeJ,ColemanM,SkuseD.2002.The

classificationof‘‘fear’’fromfacesisassociatedwithfacerecognitionskillinfemales.

Neuropsychologia,40:575–584.

CarterR.M.&S.A.Huettel.2013.Anexusmodelofthetemporal–parietaljunction.Trends

Cogn.Sci.(Regul.Ed.)17:328–336.

CatmurC,HeyesC.2011.Timecourseanalysesconfirmindependenceofimitativeand

spatialcompatibility.Journalofexperimentalpsychology.Humanperceptionand

performance.37:409-21.

ChartrandTL,LakinJL.2013.Theantecedentsandconsequencesofhumanbehavioural

mimicry.AnnRevPsych.64:285-308.

40

Christov-Moore,L.,Simpson,E.A.,Coudé,G.,Grigaityte,K.,Iacoboni,M.,&Ferrari,P.F.

2014.Empathy:Gendereffectsinbrainandbehavior.Neuroscience&Biobehavioral

Reviews,46:604-627.

Christov-Moore,L.,Conway,P.,&Iacoboni,M.2017.DeontologicalDilemmaResponse

TendenciesandSensorimotorRepresentationsofHarmtoOthers.Frontiersin

integrativeneuroscience,11:34.

ClaysonPE,ClawsonA,LarsonMJ.2011.Sexdifferencesinelectrophysiologicalindicesof

conflictmonitoring.BiologicalPsychology.87:282-289.

Cohen,J.1992.Apowerprimer.QuantitativeMethodsinPsychology.112:155-159.

Cooper,R.P.,Catmur,C.,&Heyes,C.2012.Areautomaticimitationandspatialcompatibility

mediatedbydifferentprocesses?CognSci.37:605–30.

CraccoE,BardiL,DesmetC,RigoniD,RadkovaI,DeschrijverE,GenschowO,DeCosterL,

BrassM.AutomaticImitation:AMeta-Analysis.Inpress.PsychologicalBulletin.

CrescentiniC,MengottiP,GrecucciA,RumiatiRI.2011.Theeffectofobservedbiological

andnon-biologicalmovementsonactionimitation:anfMRIstudy.Brainresearch.

1420:80-92.

CrossKA,IacoboniM.2013.Optimisedneuralcoding?Controlmechanismsinlargecortical

networksimplementedbyconnectivitychanges.HumanBrainMapping.34:213-25.

Cross,K.A.,Torrisi,S.,Losin,E.A.R.,&Iacoboni,M.2013.Controllingautomaticimitative

tendencies:interactionsbetweenmirrorneuronandcognitivecontrol

systems.Neuroimage,83:493-504.

Cross,K.A.,&Iacoboni,M.2014.Toimitateornot:Avoidingimitationinvolvespreparatory

inhibitionofmotorresonance.Neuroimage,91:228-236.

41

deGuzmanM,BirdG,BanissyMJ,CatmurC.2016.Self–othercontrolprocessesinsocial

cognition:fromimitationtoempathy.Phil.Trans.R.Soc.B.371(1686):20150079.

delaVegaA,ChangLJ,BanichMT,WagerTD,YarkoniT.2016.Large-scalemeta-analysisof

humanmedialfrontalcortexrevealstripartitefunctionalorganization.Journalof

Neuroscience.36(24):6553-62.

DimbergU,LundquistLO.1990.Genderdifferencesinfacialreactionstofacialexpressions.

BiolPsychol.30:151-159.

Dodell-FederD,Koster-HaleJ,BednyM,SaxeR.2011.fMRIitemanalysisinatheoryofmind

task.Neuroimage.55:705-712.

DonnellanMB,OswaldFL,BairdBM,LucasRE.2006.Themini-IPIPscales:tiny-yet-effective

measuresoftheBigFivefactorsofpersonality.Psychologicalassessment.18:192-

203.

DowningPE,JiangY,ShumanM,KanwisherN.2001.Acorticalareaselectiveforvisual

processingofthehumanbody.Science.293:2470-3.

DufourN,RedcayE,YoungL,MavrosPL,MoranJM,TriantafyllouC,GabrieliJD,SaxeR.

2013.Similarbrainactivationduringfalsebelieftasksinalargesampleofadults

withandwithoutautism.PloSOne.8:e75468.

DuguéL,MerriamEP,HeegerDJ,CarrascoM.2017.SpecificVisualSubregionsofTPJ

MediateReorientingofSpatialAttention.CerebralCortex1–16.

DuncanJ.2010.Themultiple-demand(MD)systemoftheprimatebrain:mentalprograms

forintelligentbehaviour.TrendsCognSci.14:172-9.

42

EickhoffSB,StephanKE,MohlbergH,GrefkesC,FinkGR,AmuntsK,ZillesK.2005.Anew

SPMtoolboxforcombiningprobabilisticcytoarchitectonicmapsandfunctional

imagingdata.Neuroimage.25:1325–1335.

EklundA,NicholsTE,KnutssonH.2016.Clusterfailure:WhyfMRIinferencesforspatial

extenthaveinflatedfalse-positiverates.ProceedingsoftheNationalAcademyof

SciencesoftheUnitedStatesofAmerica113:7900–7905.

EriksonBA,EriksonCW.1974.Effectsofnoiselettersupontheidentificationofatarget

letterinanonsearchtask.PerceptionandPsychophysics.16:143-149.

FedorenkoE,HsiehPJ,Nieto-CastañónA,Whitfield-GabrieliS,KanwisherN.2010.New

methodforfMRIinvestigationsoflanguage:definingROIsfunctionallyinindividual

subjects.Journalofneurophysiology.104:1177-94.

FedorenkoE,DuncanJ,KanwisherN.2013.Broaddomaingeneralityinfocalregionsof

frontalandparietalcortex.ProcNatlAcadSci.110:16616-21.

FristonKJ,WorsleyKJ,FrackowiakRSJ,MazziottaJC,EvansAC.1994.Assessingthe

significanceoffocalactivationsusingtheirspatialextent.HumanBrainMapping.

1:210-220.

FrithCD,FrithU.1999.Interactingminds--abiologicalbasis.Science.286:1692-1695.

Genschow,O.,vanDenBossche,S.,Cracco,E.,Bardi,L.,Rigoni,D.,&Brass,M.2017.Mimicry

andautomaticimitationarenotcorrelated.PloSone,12(9),e0183784.

Goldberg,LR.1992.ThedevelopmentofmarkersfortheBig-Fivefactor

structure.PsychologicalAssessment,4,26-42.

43

GowenE,BoltonE,PoliakoffE.2016.Believeitornot:Movingnon-biologicalstimuli

believedtohavehumanorigincanberepresentedashumanmovement.Cognition.

146:431-8.

HazeltineE,PoldrackR,GabrieliJDE.2007.Neuralactivationduringresponsecompetition.

JCognNeurosci.12:118-129.

HeyesC.2011.Automaticimitation.Psychologicalbulletin.137:463.

HirschfeldLA,GelmanSA.1994.Mappingthemind:Domainspecificityincognitionand

culture.CambridgeUniversityPress.

HogeveenJ,ObhiSS,BanissyMJ,SantiestebanI,PressC,CatmurC,BirdG.2014.Task-

dependentanddistinctrolesofthetemporoparietaljunctionandinferiorfrontal

cortexinthecontrolofimitation.SocialCognitiveandAffectiveNeuroscience,

10:1003-1009.

HydeJS.2014.Gendersimilaritiesanddifferences.Annualreviewofpsychology.65:373-98.

IacoboniM,WoodsRP,BrassM,BekkeringH,MazziottaJC,RizzolattiG.1999.Cortical

mechanismsofhumanimitation.Science.286:2526-2528.

IacoboniM.2009.Neurobiologyofimitation.CurrOpinNeurobiol.19:661-665.

JosephsO,HensonRN.1999.Event-relatedfunctionalmagneticresonanceimaging:

modelling,inferenceandoptimization.PhilosophicalTransactionsoftheRoyal

SocietyofLondonB:BiologicalSciences,354:1215–1228.

JulianJB,FedorenkoE,WebsterJ,KanwisherN.2012.Analgorithmicmethodfor

functionallydefiningregionsofinterestintheventralvisualpathway.Neuroimage.

60:2357-64.

44

KanwisherN.2010.Functionalspecificityinthehumanbrain:awindowintothefunctional

architectureofthemind.ProcNatlAcadSci.107:11163-70.

KlapperA,RamseyR,WigboldusD,CrossES.2014.TheControlofAutomaticImitation

BasedonBottom-UpandTop-DownCuestoAnimacy:InsightsfromBrainand

Behavior.JCognNeurosci.,26:2503-2513.

KrachS,BlumelI,MarjoramD,LatasterT,KrabbendamL,WeberJ,vanOsJ,KircherT.

2009.Arewomenbettermindreaders?Sexdifferencesinneuralcorrelatesof

mentalizingdetectedwithfunctionalMRI.BMCNeuroscience.10.

KriegeskorteN,SimmonsWK,BellgowanPS,BakerCI,2009.Circularanalysisinsystems

neuroscience:thedangersofdoubledipping.Nat.Neurosci.12:535–540.

KriegeskorteN,MurM,BandettiniP.2008.Representationalsimilarityanalysis-connecting

thebranchesofsystemsneuroscience.Frontiersinsystemsneuroscience.2:4.

Lakens,D.(2013).Calculatingandreportingeffectsizestofacilitatecumulativescience:a

practicalprimerfort-testsandANOVAs.Frontiersinpsychology,4.

MarshLE,BirdG,CatmurC.2016.Theimitationgame:Effectsofsocialcueson‘imitation’

aredomain-generalinnature.NeuroImage.139:368-75.

MichaelJ,D’AusilioA.2015.Domain-specificanddomain-generalprocessesinsocial

perception–Acomplementaryapproach.Consciousnessandcognition.36:434-437.

MengottiP,Corradi-Dell'AcquaC,RumiatiRI.2012.Imitationcomponentsinthehuman

brain:anfMRIstudy.Neuroimage.59:1622-30.

MillerDI,HalpernDF.2014.Thenewscienceofcognitivesexdifferences.TrendsCognSci.

18:37-45.

45

Müller,B.C.,Leeuwen,M.L.,Baaren,R.B.,Bekkering,H.,&Dijksterhuis,A.2013.Empathyis

abeautifulthing:Empathypredictsimitationonlyforattractiveothers.Scandinavian

journalofpsychology,54(5),401-406.

MumfordJA,NicholsTE.2008.PowercalculationforgroupfMRIstudiesaccountingfor

arbitrarydesignandtemporalautocorrelation.Neuroimage.39:261-8.

MumfordJA.2012.ApowercalculationguideforfMRIstudies.Socialcognitiveandaffective

neuroscience.7:738-42.

NeeDE,WagerTD,JonidesJ.2007.Interferenceresolution:Insightsfromameta-analysisof

neuroimagingtasks.CognAffectBehavNeurosci.7:1-17.

Newman-NorlundRD,vanSchieHT,vanZuijlenAM,BekkeringH.2007.Themirrorneuron

systemismoreactiveduringcomplementarycomparedwithimitativeaction.Nat

Neurosci.10:817-819.

Nieto-CastañónA,FedorenkoE.2012.Subject-specificfunctionallocalisersincrease

sensitivityandfunctionalresolutionofmulti-subjectanalyses.Neuroimage.

63:1646-69.

NormanKA,PolynSM,DetreGJ,HaxbyJV.2006.Beyondmind-reading:multi-voxelpattern

analysisoffMRIdata.Trendsincognitivesciences.10:424-30.

PeelenMV,DowningPE.2005.Selectivityforthehumanbodyinthefusiformgyrus.Journal

ofneurophysiology93:603-608.

PitcherD,DilksDD,SaxeRR,TriantafyllouC,KanwisherN.2011.Differentialselectivityfor

dynamicversusstaticinformationinface-selectivecorticalregions.Neuroimage,

56:2356-2363.

46

PoldrackRA.2006.Cancognitiveprocessesbeinferredfromneuroimagingdata?Trendsin

cognitivesciences,10(2):59-63.

QuadfliegS,FlanniganN,WaiterGD,RossionB,WigGS,TurkDJ,MacraeCN.2011.

Stereotype-basedmodulationofpersonperception.Neuroimage.57:549–557.

RizzolattiG,CraigheroL.2004.Themirror-neuronsystem.AnnuRevNeurosci.27:169-192.

RubiaK,HydeZ,HalariR,GiampietroV,SmithA.2010.Effectsofageandsexon

developmentalneuralnetworksofvisual-spatialattentionallocation.NeuroImage.

51:817-827.

RussellTA,TchanturiaK,RahmanQ,SchmidtU.2007.Sexdifferencesintheoryofmind:A

maleadvantageonHappé's“cartoon”task.CognitionandEmotion.21:1554–1564.

SantiestebanI,BanissyMJ,CatmurC,BirdG.2012.EnhancingSocialAbilitybyStimulating

RightTemporoparietalJunction.CurrBiol.22:2274-2277.

Santiesteban,I.,Banissy,M.J.,Catmur,C.,&Bird,G.2015.Functionallateralizationof

temporoparietaljunction–imitationinhibition,visualperspective-takingandtheory

ofmind.EuropeanJournalofNeuroscience,42(8),2527-2533.

SaxeR,KanwisherN.2003.Peoplethinkingaboutthinkingpeople:theroleofthetemporo-

parietaljunctionin“theoryofmind”.Neuroimage.19:1835-42.

SaxeR,BrettM,KanwisherN.2006.Divideandconquer:adefenseoffunctionallocalisers.

Neuroimage.30:1088-96.

SchurzM,TholenMG,PernerJ,MarsRB,SalletJ.2017.Specifyingthebrainanatomy

underlyingtemporo-parietaljunctionactivationsfortheoryofmind:Areviewusing

probabilisticatlasesfromdifferentimagingmodalities.Humanbrainmapping.

47

SchuwerkT,SchurzM,MüllerF,RupprechtR,SommerM.2017.TherTPJ’soverarching

cognitivefunctioninnetworksforattentionandtheoryofmind.Socialcognitiveand

affectiveneuroscience.12(1):157-168.

SimonJR.1969.Reactionstowardthesourceofstimulation.Journalofexperimental

psychology.81:174.

SchaafsmaSM,PfaffDW,SpuntRP,AdolphsR.2015.Deconstructingandreconstructing

theoryofmind.Trendsincognitivesciences,19(2),65-72.

Sonnby–Borgström,M.2002.Automaticmimicryreactionsasrelatedtodifferencesin

emotionalempathy.Scandinavianjournalofpsychology,43(5),433-443.

Sonnby-Borgström,M.,Jönsson,P.,&Svensson,O.(2003).Emotionalempathyasrelatedto

mimicryreactionsatdifferentlevelsofinformationprocessing.JournalofNonverbal

behavior,27(1),3-23.

Sonnby-BorgströmMA,JönssonP,SvenssonO.2008.Genderdifferencesinfacialimitation

andverballyreportedemotionalcontagionfromspontaneoustoemotionally

regulatedprocessinglevels.ScandJPsychol.49:111-22.

SouthgateV,HamiltonAF.2008.Unbrokenmirrors:Challengingatheoryofautism.Trends

CognSci.12:225-229.

SowdenS,&ShahP.2014.Self-othercontrol:acandidatemechanismforsocialcognitive

function.Frontiersinhumanneuroscience.8:789.

SowdenS,CatmurC.2015.Theroleoftherighttemporoparietaljunctioninthecontrolof

imitation.CerebralCortex,25:1107-1113.

SpenglerS,vonCramonDY,BrassM.2009.Controlofsharedrepresentationsreliesonkey

processesinmentalstatereasoning.HumanBrainMapping.30:3704-3718.

48

SpenglerS,vonCramonDY,BrassM.2010.Resistingmotormimicry:Controlofimitation

involvesprocessescentraltosocialcognitioninpatientswithfrontalandtemporo-

parietallesions.SocialNeuroscience,5:401-416.

SpornsO,TononiG,KötterR.2005.Thehumanconnectome:astructuraldescriptionofthe

humanbrain.PLoSCompBiol.1:e42.

Spunt,RP,Adolphs,R.2014.Validatingthewhy/howcontrastforfunctionalMRIstudiesof

theoryofmind.Neuroimage,99:301-311.

SpuntRP,AdolphsR.2015.FolkExplanationsofBehavior:ASpecializedUseofaDomain-

GeneralMechanism.PsychologicalScience.26:724–736.

Spunt,RP,AdolphsR.2017.Anewlookatdomainspecificity:insightsfromsocial

neuroscience.NatureReviewsNeuroscience,18(9):559.

StoetG.2010.Sexdifferencesintheprocessingofflankers.TheQuarterlyJournalof

ExperimentalPsychology.63:633-638

StoetG.2017.SexdifferencesintheSimontaskhelptointerpretsexdifferencesinselective

attention.Psychologicalresearch.81:571-81.

ThayerJF,JohnsenBH.2000.Sexdifferencesinjudgementoffacialaffect:amultivariate

analysisofrecognitionerrors.ScandinavianJournalofPsychology,41:243–246.

vanSchieHT,vanWaterschootBM,BekkeringH.2008.Understandingactionbeyond

imitation:reversedcompatibilityeffectsofactionobservationinimitationandjoint

action.JExpPsycholHumPerceptPerform.34:1493.

VanOverwalleF.2009.Socialcognitionandthebrain:ameta-analysis.Humanbrain

mapping.30:829-858.

49

WagerTD,NicholsTE.2003.OptimizationofexperimentaldesigninfMRI:ageneral

frameworkusingageneticalgorithm.NeuroImage.18:293-309.

WagerTD,SylvesterCYC,LaceySC,NeeDE,Franklin,M,JonidesJ.2005.Commonand

uniquecomponentsofresponseinhibitionrevealedbyfMRI.NeuroImage.27:323-

340.

WangY,RamseyR,HamiltonAFdeC.2011.Thecontrolofmimicrybyeyecontactis

mediatedbymedialprefrontalcortex.JNeurosci.31:12001e12010.

WeeksDJ,ProctorRW.1990.Salient-featurescodinginthetranslationbetweenorthogonal

stimulusandresponsedimensions.JournalofExperimentalPsychology:General

119:355–366.

Wiggett,AJ,HudsonM,TipperSP,DowningPE.2011.Learningassociationsbetweenaction

andperception:Effectsofincompatibletrainingonbodypartandspatialpriming.

BrainandCognition,76:87–96.

YarkoniT,PoldrackRA,NicholsTE,VanEssenDC,WagerTD.2011.Large-scaleautomated

synthesisofhumanfunctionalneuroimagingdata.Naturemethods.8(8):665.

ZakiJ,HenniganK,WeberJ,OchsnerKN.2010.Socialcognitiveconflictresolution:

Contributionsofdomaingeneralanddomainspecificneuralsystems.Journalof

Neuroscience.30:8481–8488.

ZakiJ,OchsnerKN.2012.Theneuroscienceofempathy:progress,pitfallsandpromise.

Natureneuroscience.15(5):675.

50

Tables

Table1.fMRIstudiesinvestigatingimitationcontrolusingmodifiedversionsofthe

imitationinhibitiontask.EvidenceforimplicationofmPFCandrTPJisinconsistent

acrossstudies.

Sample

(Males:

Females)

Dissociatio

nof

Imitative

andSpatial

Processes

Analysis BrainNetworks

ROI Whole

-Brain

ToM MD

mPFC rTPJ

Brass,

Zysset,&

von

Cramon,

2001

10(4:6) ü ü ü

Brass,

Derfuss,

&van

Cramon,

2005

20(8:12) ü ü ü ü

Brass,

Ruby,&

Spengler,

20* ü ü ü

52

2009

Spengler,

von

Cramon,

&Brass,

2009

18(9:9) ü ü ü ü ü

Bien,

Roebroec

k,Goebel,

&Sack,

2009

15(5:10) ü ü ü

Crescenti

ni,

Mengotti,

Grecucci,

&

Rumiati,

2011

19(9:10) ü ü ü

Cross&

Iacoboni,

2013

24(12:12) ü ü ü

Mengotti,

Corradi

22(10:12) ü ü ü

53

Dell’Acqu

a,&

Rumiati,

2012

Cross,

Torrisi,

Losin,&

Iacoboni,

2013

25(5:15) ü ü ü ü

Klapper,

Ramsey,

Wigboud

s,&

Cross,

2014

19(2:17) ü ü**

Marsh,

Bird,&

Catmur,

2016

24(7:17) ü ü ü

Wang,

Ramsey,

&

Hamilton,

20(5:15) ü ü ü ü

54

2011

Table1.fMRIstudiesinchronologicalorderthatinvestigatedimitationcontrolusing

modifiedversionsoftheimitationinhibitiontask.Forallstudies,engagementofmPFCor

rTPJisreportedonlyforcontraststhattestforinhibitingtheurgetoautomaticallyimitate.

Engagement(ornoengagement)oftheMDnetworkisreportedonlyforwhole-brain

analyses.ExceptforinWang,Ramsey,&Hamilton(2011),whichusedhandmovements,all

othertasksusedmodifiedversionsoftheimitationinhibitiontasksinvolvingfinger

movements(Brassetal.,2000).Foramoredetailedversionofthistable,see

SupplementaryTableS6.*Numberofmalesandfemalesnotmentioned.**mPFCshowed

engagementonlyatp<.005,uncorrected.

Table2.Generalcompatibilityeffectandsex*compatibilityinteractionforthe

imitationinhibitiontask(Experiment1).

MNIcoordinates

Region Cluster

Size

PFWE

Corr

t-

valu

e

x y z

(A) GENERALCOMPATIBILITYEFFECT(Incompatible>Compatible)

Linferiorparietallobuleextending

intosuperiorparietallobuleand

superiorfrontalgyrus

986 <0.001 8.40 -39 -40 43

6.50 -36 -37 70

6.38 -27 -7 70

Lcerebellum 150 0.001 5.79 -21 -55 -41

4.95 -30 -55 -35

4.72 -9 -70 -44

Rcerebellum 198 <0.001 5.71 21 -58 -44

5.12 45 -46 -32

4.32 39 -55 -23

Rprecentralgyrusextendingacross

superiorandmiddlefrontalgyri

183 <0.001 5.12 27 -1 70

5.04 42 2 58

4.39 39 -10 61

Rpostcentralgyrusextendinginto

superiorandinferiorparietal

481 <0.001 5.18 33 -40 73

4.55 42 -40 67

56

lobules

4.50 48 -34 37

Rposteriormiddletemporalgyrus 41 0.179 4.61 66 -46 1

4.13 57 -40 -8

3.63 60 -43 13

Linsula 24 0.458 4.67 -36 17 -2

Rposteriormedialfrontalcortex 20 0.564 4.79 3 -4 73

Lposteriormedialfrontalcortex 55 0.083 4.40 -3 -1 52

3.82 -6 11 52

Rpallidumextendingintothalamus 11 0.834 4.14 21 -7 -2

3.80 15 -6 7

Lparacentrallobule 11 0.834 3.89 -12 -19 79

Rmiddlecingulatecortex 20 0.564 3.85 9 14 43

3.78 6 8 49

3.67 6 17 52

(B) SEX*COMPATIBILITY

[Female(incompatible>compatible)>Male(incompatible>compatible)]

Lsuperiorparietallobuleextending

intopostcentralgyrus

93 0.011 4.98 -21 -37 70

4.80 -30 -19 46

4.60 -24 -31 52

57

Lcerebellum 16 0.679 4.34 -24 -55 -38

4.04 -21 -55 -29

Table2.Regionssurvivingavoxel-levelthresholdofp<.001and10voxelsarereportedfor

the(A)generalcompatibilityeffectand(B)sex*compatibilityinteractionfortheimitation

inhibitiontask.Subclustersatleast8mmfromthemainpeakarelisted.Boldfontindicates

clustersthatsurvivecorrectionformultiplecorrectionsusingafamily-wiseerror(FWE)

correction(p<.05).MNI=MontrealNeurologicalInstitute.

Table3.ResponsesineachMDnetworkfROIforspatialandimitativecompatibility.

ROI ROI

size

Inter-

subject

overlap

Average

ROI

mask

size

(voxels)

SpatialCompatibility ImitativeCompatibility

T p-value p-FDR t p-

value

p-FDR

L_SPL

L_IPS

L_IPL

L_MFG

L_PrecG

L_IFG

L_Insula

L_SMA

R_SPL

R_IPS

R_IPL

R_MFG

R_PrecG

R_IFG

R_Insula

1173

287

641

536

338

181

197

294

1181

227

599

535

269

265

184

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

117

28

64

53

33

18

19

29

118

22

59

53

26

26

18

2.00

2.00

2.72

2.16

2.17

1.83

2.78

2.52

2.30

2.03

2.65

3.57

2.43

2.61

2.09

0.026

0.026

0.005

0.018

0.018

0.040

0.004

0.008

0.013

0.024

0.005

<0.001

0.009

0.006

0.021

0.028

0.028

0.019

0.028

0.028

0.037

0.019

0.020

0.026

0.028

0.019

0.006

0.021

0.019

0.028

1.13

1.96

2.05

0.53

0.91

1.53

0.52

0.39

1.56

2.30

2.50

1.55

1.59

2.53

1.24

0.131

0.028

0.023

0.301

0.184

0.066

0.304

0.349

0.062

0.013

0.008

0.064

0.060

0.007

0.120

0.191

0.089

0.089

0.324

0.227

0.118

0.324

0.349

0.118

0.069

0.063

0.118

0.118

0.063

0.175

59

R_SMA

328

1 32

2.30

0.022

0.028

1.06

0.148

0.198

Table3.ResponsesineachMDnetworkfROIforspatialandimitativecompatibility.For

eachindividual,fortheMDnetworkmask,theHard>Easycontrastwasusedandthetop

10%ofvoxels(basedont-values)withineachparcelweredefinedasthatindividual’sfROI.

Uncorrectedp-valuesaswellasFDRcorrectedpvaluesarereported.CellsinboldarefROIs

whichsurvivecorrectionformultiplecomparisons(p<.05,FDR-corrected).

Table4.ResponsesineachToMnetworkfROIforspatialandimitative

compatibility.

ROI ROI

size

Inter-

subject

overlap

Average

ROI

mask

size

(voxels)

SpatialCompatibility ImitativeCompatibility

T p-value p-FDR t p-value p-FDR

DMPFC 576 1 57 -1.38 0.913 0.951 -1.167 0.876 0.898

MMPFC 494 1 49 -0.043 0.517 0.951 -1.081 0.857 0.898

VMPFC 382 1 38 -1.690 0.951 0.951 -1.286 0.898 0.898

RTPJ 1018 1 101 1.543 0.065 0.258 -0.106 0.542 0.898

Table 4. Responses in eachToMnetwork fROI for spatial and imitative compatibility.

Foreachindividual,fortheToMnetworkmask,theBelief>Photocontrastwasusedand

the top 10% of voxels (based on t-values) within each parcel were defined as that

individual’sfROI.Uncorrectedp-valuesaswellasFDRcorrectedpvaluesarereported.

CellsinboldarefROIswhichsurvivecorrectionformultiplecomparisons(p<.05,FDR-

corrected).

61

FigureCaptions

Figure1.Sexdifferencesininhibitorycontroltasks.Femalesexperiencegreater

interferencethanmalesinmultipleinhibitorycontroltasks.

Figure1.Sexdifferencesininhibitorycontroltasks.Femalesexperiencegreaterinterferencethanmalesinmultipleinhibitorycontroltasks.Imagesareproducedbasedonfiguresanddescriptionineachexperimentapartfrom

Butleretal.(2015)whicharetheactualimagesused.Also,inRubiaetal.(2010),thesex

differenceshowedincreasedinterferencebytheoddballtrialsratherthanthe

incongruenttrialsandthisiswhatisrepresentedbytheimages.Finally,inButleretal.

(2015),participantscompleted60trialsthatwere30±2trialspercompatibleand

incompatiblecondition.

62

Figure2.Stimulifortheimitationinhibitionandfunctionallocalisertasks.

Figure2.Stimulifortheimitationinhibitionandfunctionallocalisertasks.

StimuliandtrialdesignfortheImitationInhibitiontask(A),theMultipleDemand

Networklocalisertask(B),andtheTheoryofMindlocalisertask(C).

63

Figure3.TheparcelsusedtodefinetheMDandToMnetworks.

Figure3.TheparcelsusedtodefinetheMDandToMnetworks.

GraphicalrepresentationoftheparcelsusedtodefinetheMDandToMnetworkfROIs.

TheMDnetworkconsistedin16parcelsandtheToMnetworkincluded4parcels.

64

Figure4.GeneralCompatibilityEffectandSex*CompatibilityInteractionintheimitationinhibitiontask(Experiment1).

65

Figure4.GeneralCompatibilityEffectandSex*CompatibilityInteractioninthe

imitationinhibitiontask(Experiment1).

(A)ResultsfortheGeneralCompatibilityEffect(Incompatible>Compatible).Clusters

emergedinthedorsalfrontoparietalcortices.(B)ResultsfortheSex*Compatibility

Interaction(definedas[Female(CompatibilityEffect)>Male(CompatibilityEffect)].

Clustersemergedintheleftsuperiorparietalcortexextendingintothepostcentral

gyrus.

BoththeMDnetworkparcels(C)andToMnetworkparcels(D)andgeneral

compatibility(i)andsex*compatibility(ii)contrastsaredisplayedonacommonbrain

template.TheoverlapshowsoverlapbetweenareasoftheMDnetworkandbrain

regionsengagedbythegeneralcompatibilityeffect.Voxel-wisethresholdusedforall

imageswasp<.001,k=10.Foracompletesetofresults,seeTable2,andSupplementary

TablesS2.1andS2.2.

66

Figure 5. Behavioural Sex Differences in Imitative and Spatial Compatibility

Effects

Figure 5. Behavioural Sex Differences in Imitative and Spatial Compatibility

Effects

The spatial and imitative compatibility effects (RTs) in males and females in

milliseconds.Errorbarsdenotestandarderrorofmeanbyparticipants.

67

Figure6.ResponsesintheMD(A)andToM(B)networkfROIsforimitativeand

spatialcompatibilityeffects.

Figure6.ResponsesintheMD(A)andToM(B)networkfROIsforimitativeand

spatialcompatibilityeffects.

Theparcelsusedtodefine individual fROIsandtheresponsestospatialand imitative

compatibility effects in the MD (A) and ToM (B) network fROIs. Error bars denote

standarderrorofmeanbyparticipants.AllMDnetworkfROIsweresensitivetospatial

compatibility effects (FDR corrected, p<.05). Bilateral IPL, bilateral IPS, and the right

IFG showed a significant response for imitative compatibility effects, but at an

uncorrectedthresholdofp<.001.

68

Figure7.SexdifferencesintheresponsesinMDnetworkfROIsforimitativeand

spatialcompatibilityeffects.

Figure7.SexdifferencesintheresponsesinMDnetworkfROIsforimitativeand

spatialcompatibilityeffects.

Responsestospatial(A)andimitativecompatibility(B)effectsseparatelyformalesand

femalesintheMDnetwork.Errorbarsdenotestandarderrorofmeanbyparticipants.

NoneofthefROIsshowedasexdifferenceeitherinimitativeorspatialcompatibility.

69

SupplementaryInformation

DevelopmentofStimuli

TheimitationinhibitiontaskusedinExperiment2inthepresentstudyisamodified

versionofapreviouslyexistingparadigm(Brassetal.,2000;Catmur&Heyes,2010).In

ordertoavoidanyown-sexbias,wedecidedtouseahandstimuluswhichwasratedas

neutral.51participants(19males,Meanage=23.49,SDage=3.12)otherthantheones

whoparticipatedinthecurrentstudywereaskedtorate18whiteCaucasianhand

stimuli(9malehands,9femalehands)onascaleof1to9,with1being“extremely

masculine,”9being“extremelyfeminine”and5being“neutral.”Anaverageoftherating

scorewasobtainedforeachhandstimulus.Thehandwhichhadanaveragerating

closestto5(Mean=5.08)wastakenasthefinalstimulusasitwasconsideredtobe

ratedmost“neutral”amongstallotherstimuli.Inthepresentexperiment,weagain

askedallparticipantstoratethehandtheysawintheimitationinhibitiontaskonthe

samescale:1to9,with1being“extremelymasculine,”9being“extremelyfeminine”

and5being“neutral.”Themeanofthehandratingswas5.27;thus,onaverage,

participantsperceivedthehandas‘neutral.’

70

TableS1.1.ResponsesineachMDnetworkfROIforspatialandimitativecompatibilitywhenindividualcontrastswerethresholdedatp<.001,uncorrected.

TableS1.1.ResponsesineachMDnetworkfROIforspatialandimitativecompatibility

whenindividualcontrastswerethresholdedatp<.001,uncorrected.Cellsinboldshow

fROIswhichsurvivedcorrectionformultiplecomparisons(p<.05,FDRcorr.).

ROI ROIsize

Inter-subjectoverlap

AverageROImasksize

SpatialCompatibility ImitativeCompatibility

t p-value p-FDR t p-value p-FDR

L_SPL

L_IPS

L_IPL

L_MFG

L_PrecG

L_IFG

L_Insula

L_SMA

R_SPL

R_IPS

R_IPL

R_MFG

R_PrecG

R_IFG

R_Insula

R_SMA

1173

287

641

536

338

181

197

294

1181

227

599

535

269

265

184

328

0.92

0.86

0.82

0.84

0.86

0.74

0.7

0.88

0.9

0.84

0.76

0.88

0.74

0.7

0.78

0.84

413

85

89

140

103

58

48

86

415

71

111

144

56

56

44

101

1.71

1.36

1.91

2.37

1.76

0.11

2.45

3.00

1.66

1.36

1.89

3.61

2.06

2.34

2.22

2.32

0.047

0.090

0.032

0.012

0.043

0.457

0.010

0.002

0.052

0.091

0.034

<0.001

0.024

0.013

0.017

0.013

0.063

0.097

0.055

0.035

0.062

0.457

0.035

0.019

0.064

0.097

0.055

0.007

0.048

0.035

0.037

0.034

1.49

2.91

2.70

0.99

0.33

0.77

0.55

0.03

2.32

2.21

2.40

1.80

1.57

1.81

0.67

1.05

0.072

0.003

0.005

0.165

0.371

0.223

0.294

0.489

0.012

0.017

0.011

0.039

0.064

0.041

0.255

0.150

0.137

0.043

0.043

0.240

0.396

0.297

0.336

0.489

0.050

0.053

0.050

0.093

0.128

0.093

0.314

0.240

71

TableS1.2.ResponsesineachToMnetworkfROIforspatialandimitativecompatibility

whenindividualcontrastswerethresholdedatp<.001,uncorrected.

ROI ROIsize

Inter-subjectoverlap

AverageROImasksize

(voxels)

SpatialCompatibility ImitativeCompatibility

t p-value p-FDR t p-value p-FDRDMPFC 576 1 0.58 -1.23 0.88 0.88 -0.09 0.54 0.82MMPFC 494 1 0.56 -0.02 0.49 0.88 -0.56 0.71 0.82VMPFC 382 1 0.44 -0.96 0.82 0.88 -0.93 0.82 0.82RTPJ 1018 1 0.92 1.20 0.12 0.47 0.78 0.22 0.82

TableS2.1Smallvolumecorrection(SVC)withMDnetworkmaskforthegeneral

compatibilityeffectandthesex*compatibilityinteraction.

Region ClusterSize

PFWECorr

t-value

MNIcoordinatesx y z

GENERALCOMPATIBILITY(Incompatible>Compatible)Leftinferiorparietallobule(extendingintotheleftpostcentralgyrus)

382 <0.001 8.40 -39 -40 43 5.57 -48 -40 61 5.31 -36 -43 67

Leftintraparietalsulcus 79 0.005 6.50 -36 -52 46 4.27 -30 -61 49

Leftprecentralgyrus 180 <0.001 6.38 -27 -7 70 5.11 -33 -4 52 4.78 -42 -1 55 4.32 -27 -10 55 4.22 -39 -4 52

Leftsuperiorparietallobule 14 0.249 5.31 -27 -52 70 3.95 -15 -55 73 3.80 -12 -58 70

Rightmiddlefrontalgyrus

86 0.004 5.12 27 -1 70 4.04 42 2 58

Linsula 21 0.149 4.67 -36 17 -2Rightinferiorparietallobule(extendingintotherightintraparietalsulcusandpostcentralgyrus)

270 <.001 4.50 48 -34 37 4.47 39 -46 67 4.46 48 -34 43 4.41 45 -43 64 4.41 42 -34 40 4.39 33 -37 37 4.38 54 -40 40 4.35 45 -40 52 4.29 36 -46 46 4.28 39 -46 52 4.27 45 -46 55

Rightsuperiorparietallobule 29 0.086 4.40 36 -52 67 3.97 42 -52 52

SEX*COMPATIBILITY[Female(Incompatible>Compatible)>Male(Incompatible>Compatible)]

NosuprathresholdclustersTableS2.1.Regionssurvivingavoxel-levelthresholdofp<.001and10voxelsarereportedforthegeneralcompatibilityeffectandsex*compatibilityinteraction,smallvolumecorrectedusingtheMDmask.Subclustersatleast8mmfromthemainpeakarelisted.Boldfontindicatesclustersthatsurvivecorrectionformultiplecorrectionsusingafamily-wiseerror(FWE)correction(p<.05).MNI=MontrealNeurologicalInstitute.

TableS2.2.Smallvolumecorrection(SVC)withToMnetworkmaskforthegeneral

compatibilityeffectandthesex*compatibilityinteraction.

Region Cluste

rSize

P

FWE

Corr

t-

value

MNIcoordinates

x y z

GENERALCOMPATIBILITY(Incompatible>Compatible)

Righttemporo-parietaljunction

(supramarginalgyrus)

10 0.124 3.60 57 -43 37

3.54 51 -40 34

SEX*COMPATIBILITY

[Female(Incompatible>Compatible)>Male(Incompatible>Compatible)]

Nosuprathresholdclusters

TableS2.2.Regionssurvivingavoxel-levelthresholdofp<.001and10voxelsarereportedforthegeneralcompatibilityeffectandsex*compatibilityinteraction,smallvolumecorrectedusingtheToMmask.Subclustersatleast8mmfromthemainpeakarelisted.Boldfontindicatesclustersthatsurvivecorrectionformultiplecorrectionsusingafamily-wiseerror(FWE)correction(p<.05).MNI=MontrealNeurologicalInstitute.

Table S3. ShowingMeanRT and SD for each condition of the imitation task for both

malesandfemales.

Spatially

Compatible

Spatially

Incompatible

Imitatively

Compatible

Imitatively

Incompatible

Mean SD Mean SD Mean SD Mean SD

Males 666.20 157.50 699.42 159.76 675.00 158.56 690.62 158.93

Females 736.94 130.34 787.83 134.19 754.88 133.91 769.99 130.65

TableS4.1.ResponsesineachMDnetworkfROIfortheMDnetworklocalizercontrast.

ROI ROIsize Inter-

subject

overlap

Average

ROImask

size

(voxels)

t p-value p-FDR

L_SPL

L_IPS

L_IPL

L_MFG

L_PrecG

L_IFG

L_Insula

L_SMA

R_SPL

R_IPS

R_IPL

R_MFG

R_PrecG

R_IFG

R_Insula

R_SMA

1173

287

641

536

338

181

197

294

1181

227

599

535

269

265

184

328

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

117

28

64

53

33

18

19

29

118

22

59

53

26

26

18

32

11.69

11.15

10.30

11.08

10.40

9.38

12.26

14.56

11.36

10.18

9.25

11.61

9.43

9.13

12.43

11.38

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

<.0001

TableS4.1.AllMDnetworkfROIsweresignificantlyresponsivetotheHard>Easycontrastandsurvivedcorrectionforfalsediscoveryrate(p<.05).

TableS4.2.ResponsesineachToMnetworkfROIfortheToMnetworklocalizer

contrast.

ROI ROIsize Inter-

subject

overlap

Average

ROI

mask

size

t p-value p-FDR

DMPFC 576 1 57 7.097 <.001 <.001

MMPFC 494 1 49 7.065 <.001 <.001

VMPFC 382 1 38 5.704 <.001 <.001

RTPJ 1018 1 101 15.025 <.001 <.001

TableS4.2.AllToMnetworkfROIsweresignificantlyresponsivetotheBelief>Photocontrastandsurvivedcorrectionforfalsediscoveryrate(p<.05).

TableS5.Whole-brainanalysis(Experiment2).

Region Cluster

Size

PFWE

Corr

t-

value

MNIcoordinates

x y z

(A) GENERALCOMPATIBILITY

Leftinferiorparietallobule 333 <0.001 5.28 -42 -28 43

5.10 -48 -28 34

3.91 -30 -49 43

Leftmiddlefrontalgyrus 130 0.018 5.05 -27 -10 49

Rightinferiorparietallobuleextending

intotherightpostcentralgyrus

437 <0.001 4.85 45 -34 46

4.81 48 -25 43

4.28 30 -46 46

Rightmiddleandsuperiorgyri

extendingintotherightposterior-

medialfrontalcortex

259 0.001 4.85 30 -4 52

4.28 21 -7 64

4.23 15 5 52

Leftposterior-medialfrontal 29 0.476 4.19 -9 -1 55

Rightinferiorfrontalgyrusandrightinsula

lobe

96 0.050 3.87 42 2 19

48 8 22

42 17 1

Rightprecuneus 36 0.374 3.77 12 -64 52

3.56 18 -70 49

Rightsupramarginalgyrus 26 0.527 3.75 63 -40 31

3.68 63 -49 28

Rightcerebellum 12 0.805 3.68 36 -49 -35

(B) SPATIALCOMPATIBILITY

Intraparietalsulcusextendingintothe

rightpostcentralgyrus

136 0.019 4.65 33 -43 46

3.87 48 -25 43

Bilateralposteriormedialfrontal

117 0.033 4.55 -6 -1 55

4.22 9 5 52

3.48 -9 5 46

Leftprecentralgyrus 26 0.538 4.49 -57 5 31

Leftprecentralgyrus 68 0.139 4.34 -24 -16 58

4.28 -27 -10 52

Rightprecentralgyrus 27 0.521 4.30 63 8 28

Rightinsulalobe 46 0.282 4.23 39 14 1

Rightsuperiorfrontalgyrus 122 0.028 4.16 27 -7 58

4.15 27 -7 49

Leftinferiorparietallobule 75 0.025 3.95 -36 -37 40

3.66 -48 -31 40

Leftinsulalobe 17 0.702 3.92 -33 14 7

Leftpostcentralgyrus 13 0.780 3.69 -63 -16 34

Rightprecuneus 22 0.607 3.67 12 -70 46

3.60 12 -61 46

10 0.838 3.39 57 -16 22

(C) IMITATIVECOMPATIBILITY

Rightsupramarginalgyrusandright

postcentralgyrus

40 0.349 3.76 45 -34 43

3.64 42 -34 55

Leftinferiorparietallobule 15 0.666 3.50 -42 -28 40

(D) SEX*COMPATIBILITY(GENERAL)

Rightsuperioroccipitalgyrus 10 0.845 3.48 30 -67 28

(E) SEX*COMPATIBILITY(SPATIAL)

Rightsuperioroccipitalgyrus 10 0.838 3.65 33 -64 31

(F) SEX*COMPATIBILITY(IMITATIVE)Nosuprathresholdclusters

Table5.Regionssurvivingavoxel-levelthresholdofp<.001and10voxelsarereportedfor(A)generalcompatibility(B)spatialcompatibility(C)imitativecompatibilityeffectsandsex*compatibilityinteractionsseparatelyfor(D)general(E)spatialand(F)imitativecompatibilityeffects.Subclustersatleast8mmfromthemainpeakarelisted.Boldfontindicatesclustersthatsurvivecorrectionformultiplecorrectionsusingafamily-wiseerror(FWE)correction(p<.05).MNI=MontrealNeurologicalInstitute.

TableS6.DetailedtableoffMRIstudiesusingtheimitationinhibitiontaskandthecontributionsofToMandMDnetworksinimitation

inhibition.

Sample No.oftrialsper

condition

Design TaskInstructions ROI/Who

lebrain

Regions Thresholding

M:F Age Block/

Event-

related

ROI W

B

mPFC rTPJ MD

1. Brass,

Zysset,&

von

Cramon,

2001

10

(4:6)

23.

5

80congruent,80

incongruent

Mixed Block1:tapindex

finger

Block2:liftindex

finger

Y Y

(Frontopolar

cortex,BA10)

N Y(MFG,Cuneus,

Anteriorparietal

cortex)

p<.001,

uncorrected

2. Brass,

Derfuss,&

von

Cramon.

2005

20

(8:12)

–10for

Imi,10

for

Stroop

26 40congruent,40

incongruent,40

baseline,40null

Event-

related

Indexfingerfor‘1’

Middlefingerfor

‘2’

Y Y

aFMC

Y

(BA40)

Y p<.001,

uncorrected

3. Brass,

Ruby,&

Spengler,

2009

20 35simultaneous

congruent,35

simultaneous

incongruent,

35delayedcongruent,

35delayed

incongruent,

Mixed Sameasabove

Simultaneous:

numbercue

appearedwith

irrelevanthand

Delayed:response

ledtoappearance

ofirrelevanthand

Y Y

(for

simultaneous

incongruent

>congruent

only)*

Y

(for

simultaneous

incongruent

>congruent

only)*

P<.001,

uncorrected

4. Spengler,

von

18

(9:9)

25 72incongruent,72

congruent,36null

Event-

related

Sameas2. Y** Y

(overlapwith

Y

(overlapwith

Y

(seesupple-

P<.05,

correctedfor

Cramon,&

Brass,

2009

self-

referential

andToM

tasks;-ve

correlation

withRT

interference)

agencyand

ToMtasks)

mentarymaterial:

SII,MFG)

multiple

comparisons

5. Bien,

Roebroeck

,Goebel,&

Sack,2009

15

(5:10)

23 64imitativecongruent,

64imiincongruent,64

spatcongruent,64spat

incongruent

Mixed Block1:imitate

fingermovement

(imitativetrials)

Block2:follow

spatialcuefor

movement

(spatialtrials)

Y N N Y***

(premotorcortex,

bilateralposterior

parietaland

frontal/parietal

opercularcortex,

rightSTS)

P<.045,

clustersize

threshold=

50

6. Crescentin

i,

Mengotti,

Grecucci,

&Rumiati,

2011

19

(9:10)

24.

6

60biological

congruent,60bio

incongruent,60non-

biocongruent,60non-

bioincongruent

Mixed Modifiedversion

of1.With

biological(human

hand)andnon-

biological(white

dot)stimuli;but

pptsresponded

aftermovement

offsetinsteadof

onset

Y N N Y(onlyright

Insula)

P<.05,

correctedfor

multiple

comparisons

7. Cross&

Iacoboni,

2013

24

(12:12)

Mixed Block1:imitate

fingermovements

Block2:imitate

spatialdot

movement

Y N N N P<.05,FWE

corrected

8. Mengotti,

Corradi-

Dell’Acqua

,&

Rumiati,

2012

22

(10:12)

24.

4

80percondition Mixed Task1:tap

anatomicalfinger

Task2:tapfinger

onsamesideof

space

Y N N/Y(foronly

AN_NSover

allothersi.e.

forthe

condition‘imi

comp+spat

incomp’)

Y**** P<.05FWE

9. Cross,

Torrisi,

25

(5:15);

19-

39

80imicongruent,80

imiincongruent,80

Block Block1:liftindex

fingeronfinger

Y Y N Y(ACC,bilateral

insulaextending

P<.05,FWE

corrected

Losin,&

Iacoboni,

2013

20

include

din

analysi

s

spatcong,80spat

incong,80nulls

movement

Block2:lift

middlefingeron

fingermovement

Block¾:lift

index/middle

fingerwhendot

moves

intofrontalpole

andorbitofrontal

cortex,IFG,PrecG,

SPL)

10. Klapper,Rasey,

Wigboldus

,&Cross,

2014

19

(2:17)

21.

95

160congruent,160

incongruent

Event

related

Indexfor‘1’

Middlefor‘2’

Y Y Y(atp<.005,

uncorrected:

incong>cong,

human>non-

human)

N/Y(at

p<.005,

uncorrfor

human>non-

human)and

3-waycongx

formxbelief

atp<.05FWE

corr

11. Marsh,Bird,&

Catmur,

2016

24

(7:17)

23.

71

80imicong,80imi

incong,80spatcong,

80spatincong

Event

related

Liftindexfor‘1’

Liftmiddlefor‘2’

Y N N Y(IFG,IPL,ACC);

alsoleftTPJ

P<.05,FWE

corrected

12. Wang,Ramsey,&

Hamilton,

2011

20

(5:15)

23 96congruent(averted

+directgaze),96

incongruent,54catch

trials

Mixed Block1:Hand

open

Block2:Hand

closed

Y Y(foraverted

incong>

cong)

Y(foraverted

incong>

cong)

Y(maineffect:IPL,

Cuneus);

(averted:MOG,

MTG,STS,

temporalpole,

IFG,precuneus,

MFG,SPL,PMC,

IPL,cuneus)

P<.05,FWE

corrected