Dorsalandventralstreamactivationandobjectrecognitionperformanceinschoo - PDF document

Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildrenTessaDekker,DenisMareschal,MartinI.Sereno,MarkH.JohnsonBirkbeckCollege,UniversityofLondon,CentreforBrainandCognitiveDevelopment,MaletStreet,London,UKUniversityCollegeLondon,DepartmentofPsychology,26BedfordWay,London,UKarticleinfoArticlehistory:Received3June2010Revised26October2010Accepted1November2010AvailableonlinexxxxWeexploredhowdevelopingneuralartifactandanimalrepresentationsinthedorsalandventralstreamplay NeuroImagexxx(2010)xxx Correspondingauthor.Fax:+442076316587.E-mailaddresses:(T.Dekker),d.mareschal@bbk.ac.uk(D.Mareschal),(M.I.Sereno),mark.johnson@bbk.ac.uk(M.H.Johnson). YNIMG-07794;No.ofpages:12;4C:3,6,7,8,9 seefrontmatter©2010ElsevierInc.Allrightsreserved.10.1016/j.neuroimage.2010.11.005 ContentslistsavailableatScienceDirectjournalhomepage:www.elsevier.com/locate/ynimg Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 processingkeeponimprovingintoadolescence(Mondlochetal.,,althoughseeCrookesandMcKone,2009)andtworecentstudiesalsoindicatethattoolprocessingkeepsondevelopinguntilatleastwellintotheschool-ageyears.Forexample,Bovaetal.(2007)haveshownthattheabilitytorecognizetoolsfromanon-canonicalviewpoint,improvesrapidlybetweenthe6thand12thyear.InMounoudetal.(2007)reportedthatseeingtoolactionprimesspeedsupdecisionsabouttoolsin59-year-olds,butnotatolderages,suggestingthatthelinkbetweentoolactionrepresenta-tionsandtoolperceptionischangingduringchildhood.Theventralvisualpathwayplaysanimportantrolewhenadultsprocessaspectsofobjectsforperceptionsuchasshape,color,texture,location-andsize-constancyandtoanextentorientationconstancyGrill-Spector,2003,2009).Anotherwell-knowncharacteristicoftheadultobjectsensitivecortexintheventralstreamisitsorganizationbycategory;complexobjectsarerepresentedinadistributedmanneracrosstheinferotemporalcortex(ITC),butthereareclusteredregionswitharelativelystrongerBOLDresponseforcertainobjectcategoriesHaxbyetal.,2001).Theseregionsshowahighlyconsistentspatialorganizationacrossindividualsthatevenemergesintheabsenceofvisualexperiencewithtool-andhouse-selectiveclustersarelocatedmoremediallyinthefusiformgyrus(FFG)comparedtoface-,animal-andbody-partselectiveclusters(Hassonetal.,2003;Mahonetal.,2009).Objectprocessinginthedorsalvisualpathwayislessextensivelyinvestigated.However,earlierreportsfromsingle-unitstudiesinmacaques(SerenoandMaunsell,1998)suggestedthatinthelateralintraparietalarea(LIP),thereisselectivityforobjectidentityindependentoflocation,whichhasbeenconrmedinstudiesinadulthumans(KonenandKastner,2008).Inaddition,dorsalcorticalregionsthoughttobehumanhomologuesofthemacaqueinferiorparietalcortex(IPc)andanteriorinferiorparietalcortex(AIP),arethoughttoplayaroleinadultvisuo-motortransformationsforplanningandexecutionofobjectmanipulation(Arbib,2005;RizzolattiandMatelli,2003;Valyearetal.,2007).Perhapsbecauseofthecloselinkbetweentoolsandspecicmanualactions,apreferencefortoolsoveranimateobjectsisdistributedwidelyacrossaventraldorsalstream-spanningnetwork.Toolpreferencewasmostconsistentlyshowninthemedialfrontalgyrus,AIP,IPc,thepremotorcortexandanareainthemedialtemporalgyrus(MTG)thatrespondstonon-biologicaltoolmovementandtheearliermentionedmedialFFG.Thisactivationpatternevenpersistsduringpassiveviewingtasksthatdonotrequireexplicitactionsorinvolvetoolmotion,whichhasledtosuggestionsthattheseregionsmaybeimportantforrepresentingactionandmotionrelatedaspectsoftoolsBeauchampetal.,2003;ChaoandMartin,2000;Johnson-Frey,2004However,towhatextentregionswithacategorypreferenceperformsomespecialcomputationsrelevanttotheirspecicpreferredcategory,andwhatdrivingprinciplesliebehindtheconsistentorganizationbycategoryacrossthedorsalandventralstream,iscurrentlyunclear(Grill-Spector,2009).Recently,somedevelopmen-talfMRIstudieshaveshednewlightonthisdebatebyinvestigatinghowthecategoryselectivityintheventralstreamemerges.Mostconsistently,thesestudieshaveshownthattheresponsetofacesintheventralstreamkeepsondevelopinguntillateinchildhood.AccordingtotheInteractiveSpecializationview(Johnson,2001,2010),weneedtodiscriminatebetweenstimulus-stimulus-)tissue.Forexample,face-sensitivetissueisrevealedbyacontrastbetweencorticalresponsestofacesandabaselineresponsetoanon-objectstimulusthathascomparablelow-levelvisualcharacteristics.Face-preferring(orselective)tissuereferstocorticalregionsthatrespondtofacesconsiderablymorethancloselyrelatedstimulisuchasothercategoriesofcomplexvisualAscorticalregionsbecomebettertunedtoparticularstimuli,theInteractiveSpecializationviewpredictsincreasinglyfocalpatternsofactivationofcorticaltissuesensitivetostimulisuchasfaceswithdevelopment,alongsidetheemergenceofface-preferringtissue.ThesepredictionshavebeenconrmedforfacesinseveralrecentdevelopmentalfMRIstudies.(Aylwardetal.,2005;Gathersetal.,2004;Golaraietal.,2007;Josephetal.,2010;Passarottietal.,2003;Scherfetal.,2007Johnsonetal.,2009forreview).ThesestudiesthussupporttheInteractiveSpecializationaccountandprovideevidenceagainstmodularviewsofthebrainthathypothesizethatregionsofthecortexthatperformevolutionaryimportantfunctionssuchasfaceprocessingarehard-wiredforthatpurposefrombirth.Thedevelopmentoftoolprocessingandobjectprocessinginthedorsalstreamingeneralhasnotbeensystematicallyinvestigated.Becauseregionsthatwilleventuallyshowapreferenceforgraspableobjectsduringpassiveviewingarelocatedacrossthewholebrain,trackingthedevelopmentofthecorticalresponsetotoolscanprovideimportantinsightsintoobjectprocessinginthedorsalstreamanditsdevelopmentalinteractionswiththeventralstream.Previousdevel-opmentalfMRIstudiesonobjectprocessinghavemainlyfocusedoncategoryspecicregionsofinterest(ROIs)intheventralstream.However,objectrepresentationsdonotsolelydependontheregionsthataremostspecicallytunedtothecategoryanobjectbelongsto.Infact,substantialinformationabouttoolsandothertypesofobjectsisrepresentedinadistributedmanneracrosstheinferotemporalcortexHaxbyetal.,2001)andindorsalregionsofthebrain(Barsalou,2008;KonenandKastner,2008).Thepossibilitythatdevelopmentalchangestakeplaceintheobjectsensitivecortexoutsideregionswithacategorypreference,andmaynotevenbespecictoacategorybutgeneraltocomplexobjects,hasnotyetbeenaddressed.Throughoutthisreport,weusethewordreferringtograspableutensils,thewordanimalswhenonlyreferringtoanimals,andthetermwhenreferringtocategoriesofcomplexobjects(bothanimalsandtools).InlinewiththeInteractiveSpecializationapproachdiscussedearlier,wedistin-guishbetweenthedevelopmentofacorticalcategorypreferencethedevelopmentofcorticalobjectsensitivity.Morespecically,inordertogetafullpictureofthedevelopmentalchangesinobjectprocessingacrossthebrainduringapassiveviewingtaskinthescanner,weidentifyobjectsensitiveregions(tools+animalsscrambledimages)toexplorethedevelopingresponsedistributiontoobjectsirrespectiveofthelocalcategorypreferenceandweidentifytool-preferringregions(toolsscrambledminusanimalsscrambled)andanimal-preferringregions(animalsscrambledminustoolsscrambledimages)toexplorethedevelopmentofacategorypreference.Additionallywerelatethesetwotypesofdevelopmentalchangestoanobjectrecognitionskillthatisstilldevelopingbetweenage6and10,theage-rangeofthechildreninthepresentstudy.Althoughitiscurrentlyunknownifmerepassiveviewingoftoolswillpreferentiallyactivatedorsalstreamregionsinchildren,wecanformulatetwocontrastinghypothesesaboutthedevelopmentofadorsalandventralcategorypreferencebasedonpreviousresearch.Firstly,behavioraldevelopmentalstudiessuggestthattoolrepresen-keepondevelopingduringchildhood(Bovaetal.,2007;Mounoudetal.,2007)andithasbeensuggestedthatdorsalstreamactionrepresentationsmayplayanimportantroleinthedevelopmentoftoolrepresentations(Mahonetal.,2007,2009).Wethereforemayexpectprotractedtuningofregionswithatoolpreference,similartowhathaspreviouslybeenreportedforventralregionswithafacepreference(Hypothesis1).Secondly,threerecentfMRIstudieshavereportedthatanadult-likepreferenceforabstractsculptures,bodypartsandlmmaterialinwhichbodyparts,furnitureandtoolswereintermixedispresentintheventralLOCfrom58yearsonwardsGolaraietal.,2007;Scherfetal.,2007;Pelphreyetal.,2009Althoughtheseauthorsdidnotspecicallycomparetheresponsetotoolsacrossageandrestrictedtheiranalysistotheventralstream,ndingsraisethepossibilitythatacorticalpreferencefornon-socialobjectsmaymaturebeforethepreferenceforfaces.Wemaythereforehypothesizethatapreferencefortoolsinthedorsalandventralstreamwillbeadult-likeinchildhood,contrarytowhathasT.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 previouslybeenreportedforfaces(Hypothesis2).Finally,wecanformulateahypothesiswithrespecttothedevelopmentofobjectsensitivecortex:totheextentthatobject-processingskillsthatkeepondevelopinguntillateintochildhooddependonprocessesthatarerelevanttomultipletypesofobjects(Nishimuraetal.,2009),wemayexpectthattheresponsetobothtoolsandanimalsintheobjectsensitivecortexwillshowaprotracteddevelopmentwithageandincreasingobjectexperience(Hypothesis3).Notethatthisthirdhypothesisisnotmutuallyexclusivewitheitherofthersttwohypotheses.Tosummarize,inthepresentstudyweaimtogetmoreinsightintowhichaspectsoftheBOLDresponsetoobjects(tooloranimalpreferenceand/orobjectsingeneral)indorsalandventralcorticalregionsshowaprotracteddevelopmentduringchildhood.Inaddition,weexplorewhichdevelopmentalchangesincorticaltoolandanimalprocessingrelatetoperformanceonanindependentlymeasuredobjectrecognitiontaskthatisstillimprovinginchildhood,inordertogainabetterunderstandingofthefunctionalrolethatdevelopingneuralartifact-andanimalrepresentationsinthedorsalandventralstreamplayinchildren'sdaytodayinteractionswithobjects.MaterialsandmethodsParticipantsThirty-three6-to10-year-oldchildrenandeleven21-to34-year-oldadultstookpartinapassiveviewingtaskintheMRIscannerandinanobjectperceptiontaskthattookplaceoutsidethescanner.Theparticipantsweresplitintoanagegroupoffteen67-year-oldsthatconsistedof6boys(averageage=6.9years,SD=0.7)and9girls(averageage=7.1years,SD=0.5),agroupofeighteen8oldchildrenconsistingof8boys(averageage=9.6years,SD=0.9)and10girls(averageage=9.4years,SD=0.8)and11adultsconsistingof5women(averageage=25.3,SD=3.0)and6men(averageage=30,SD=3.9).Allparticipantswereright-handedwithnormalorcorrectedvisionandwithoutanypastorcurrentneurologicalorpsychiatricconditions,orstructuralbrainabnormal-ities.Childrenwererecruitedthroughadvertisementsinnewspapers.Adultparticipantswererecruitedvialocalrecruitmentfacilities.Beforetesting,adultparticipantsandcaretakerssignedaninformedconsentformandchildrensignedanassentform.AmedicalclearancequestionnaireandametaldetectiontestwereadministeredtoeachparticipanttoensureMRsafetycriteriawereadheredto.TheresearchwasexecutedundertheapprovedUniversityprotocolsfortheuseofhumanadultandminorparticipantsinresearch.Unusualperspectivetask(outsidethescanner)AllparticipantsperformedanobjectrecognitiontaskoutsidetheMRIscanner.ThetaskwasbasedontheunusualperspectivetaskdevelopedbyBovaetal.(2007).Tocreatethestimuli,twenty-onecommon,graspableobjectswerephotographedfromaviewpointinwhichtheyaretypicallyviewedandfromwhichtheyareeasytorecognize.Thesameobjectswerealsophotographedfromanunusual,moredifculttorecognize,viewpoint.ObjectsintheoriginalcolorphotographswereseparatedfromtheiroriginalbackgroundusingadobePhotoshopandplacedonaneutral,graybackground(rgb:200×200×200,600×800pixelsize,seeFig.1Aforexamplesandsupplementarymaterialformoredetailsontheitemsinthetask).Theobjectswerepresentedona15.7in.monitorwithaviewingdistanceofca50cm(37°×24.2°visualangle)usingE-primestimuluspresentationsoftware.ProcedureTheunusualperspectivetasktookplaceinatestingroomseparatefromthescannersuite.Inthisself-pacedtask,participantsedgraspableobjectsthatappearedconsecutivelyonthescreenfromanunusualviewpoint.Next,theyidentiedthesameobjects,butnowpresentedfromtheirtypicalviewpoint.Thissecondpartofthetaskensuredthatparticipantscouldrecognizetheobjectsintypicalview.Allparticipantswereinstructedtolookcarefullyateachobjectpresentedonthescreenandtopressthespacebarassoonastheywerereadytoguesswhattheobjectwas.Aftereachpress,abuzzersoundedandtheobjectwasreplacedwithablankscreenwithaxationcross.Aftertheparticipanthadmadeaguessandrefocusedattentiontothecross,theexperimenterinitiatedthenexttrial.Thesameprocedurewasfollowedforobjectsthatwerepresentedfromanunusualandfromatypicalviewpoint.Asthiswasarecognitiontestandnotanamingorlanguagetest,descriptions Fig.1.A:Examplesofitemsviewedfromanunusualperspectivefromtheobjectrecognitiontaskthatparticipantsperformedoutsidethescanner.B:Examplesofthestimuliparticipantswerepresentedwithinthescannerduringthepassiveviewingtaskwhiletheyxatedonthecentralcross.C:Accuracyscoresontheunusualperspectivetask(basedonasimilartaskbyBovaetal.,2007)areplottedagainstage.Answerswereconsideredtobecorrectifparticipantscorrectlynamedordescribedtheobjectfromanunusualviewpointaswellasfromatypicalviewpointandincorrectifonlythetypicalviewpointwasrecognized.Itemsthatwerenotcorrectlyrecognizedintheirtypicalviewwereexcludedfromtheanalysis.Thegraphshowsthatperformanceonthistaskincreasessubstantiallywithagefromchildhoodtoadulthood(Pearsonr=0.7,p0.000,R=0.49;F(1,43)=40.33,0.000),however,performanceonlyincreaseswithagefrom6to10(R=0.251,F(1,32)=10.406,p=0.003)andnolongerimprovesafterthe22ndyearoflife,F=1.23,p=n.s.).T.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 oftheobjectorindicationsregardingtheiruse(e.g.,insteadofsayingnailclipper,saying:athingtocutyournails)wereacceptedascorrectanswers.Anitemwasconsideredasnotrecognizedwhentheanswergivenreferredtoadifferentobject(e.g.,fork:rake,:plug).SeeBovaetal.(2007)forsimilarscoringcriteria.Itemswereexcludedfromtheanalysisiftheusualperspectivewasnotrecognizedcorrectlyoriftheunusualperspectivewasrecognizedincorrectlybuttheanswerwasplausible(insteadofsayingphone,saying:remotecontrol).SeeSupplement1formoreinformationabouttheitems.Passiveviewingtask(insidethescanner)FifteentypesofhighlyfamiliaranimalsandtoolswereselectedforthepassiveviewingtaskintheMRIscanner.Asinglecolorphotographwasselectedforeachofthese30typesofobjectsandtheanimalortoolonthephotowasplacedonaneutralgraybackgroundusingAdobePhotoshop(background:200×200×200rgb,stimulussize:600×450pixels).AscrambledversionofeachpicturewascreatedinMatlab,byapplyinga6×9gridtotheobjectstimuliandshufingthegridcells(gridcellsize100×50pixels).Objectdetailswerestillvisibleinthesescrambledstimulibuttheoverallobjectshapewasgrosslydistorted.Aredxationcrosswithablackoutline(30pixelsheightandwidth)wasdisplayedinthecenter,ontopofallstimuliandduringinter-stimulusintervalswhennoobjectswerepresentedFig.1Bforexamplesofthestimuli).Imageswereprojectedontoabackprojectionscreen(23°×14°visualangle,screenresolution800×600)attachedtotheboreofthescanner.Participantssawthisscreenthroughamirrorthatwasmountedontheradiofrequencycoilthatsurroundedtheirhead.StimuliwerepresentedusingMatlab6.0(Mathworks)andCogent2000extensions(www.vislab.ucl.ac.uk/cogent.phpProcedureandtaskInthescanner,weaskedallsubjectstolookatthescreenwhilexatingacrossinthecenterthroughouttherun.Thistaskhastwoimportantbenets.Firstly,taskdemandsareverylow,soparticipantsofawiderangeofagescandothesametaskwithminimalconfoundinginuencesofdevelopmentaldifferencesinperformance.Secondly,becauseallparticipantslookedatthesamelocationonthescreen,developmentaldifferencesineye-movements,whichcanmasktheBOLDresponsetoobjectsintheretinotopicregionsthatmakeupabout50%ofthebrain,includingthefusiformgyrus,regionsintheparietalandfrontalcortex(SayginandSereno,2008)areminimized.Animals,toolsandscrambledobjectswerepresentedin16blocksof15s(4animalblocks,4toolblocksand8scrambledobjectbaselineblocks).Duringablock,15itemsfromoneofthethreestimulusconditionswereeachpresentedfor800ms,followedbya200msxationscreen.Theorderoftheblockswasrandomized,aswastheorderofthestimuli,withtheconstraintthatnostimulusoccurredmorethanonceduringablock.Thetotaldurationofarunwas16×15=240s.Eachstimuluswasrepeatedfourtimesduringarun.Tworunswereacquiredforeachparticipant.Therunswereseparatedbyastructuralscantolimitstimulusadaptationeffectsandtopreventyoungchildrenfromgettingboredwiththetask.Childrenweretrainedtoliestillinthescannerbeforetheexperimentalrunsbegan.EachtimeanMR-compatiblevideocamerarecordedexcessivemovementduringthetraining,thescanneroperatorstoppedacartoonthatthechildwaswatchingandexplainedthatthemovementhe/shewasmakingwouldharmtheimagequality.Thistrainingcontinueduntilthechildwaslyingcientlystillforafewminutes.Childrenweremonitoredwiththecameraandviaanintercomthroughoutthesessiontoensurethattheyremainedstill,thattheywerexatingthecentralcrossonthescreen,andthattheywerecomfortableduringscanning.Allparticipantsheldanalarmbuttonintheirlefthandsothattheycouldnotifythescanneroperatoratanytimetheychoseto,incasetheywantedtostoptheexperiment.dataacquisitionandpreprocessingMRdatawerecollectedwithaSiemensTIMAvanto1.5TMRIscannerusinga12-channelreceive-onlyheadcoil.Ahigh(1×1×1mm)resolutionT1-weighted3DMPRAGEanatomicalsequence(magnetizationpreparedlowanglespoiledgradientecho)wasperformedtoacquirehigh-resolutionimagesofthebrainstructureofeachparticipant(imagematrix=224×256,160parti-tions,TR:2730,TE:3.57,effectiveTI1000ms,ipangle:7°).Functionaldatawerecollectedusinganechoplanar2Dimagingsequencewithimage-basedprospectiveacquisitioncorrectionforheadmotion(Thesenetal.,2000).Perrun,wecollectedonehundredvolumesthatcoveredthewholebrainandconsistedofthirty-twoslices,acquiredintheaxialplaneininterleavedascendingorder(bandwidth=1906Hz/pix,TR:2.5,TE:39,ipangle:90,voxelsize:3.5×3.5×3.5cm,matrix64×64).AllfunctionaldatawereconvertedtoNIFTIformatandanalyzedusingFSL(http://www.fmrib.ox.ac.uk/fsl).First,theimageswereippedintothestandardFSLorientationandthenon-brainstructuresintheco-planarhigh-resolutionT2-weightedEPIvolumeandtheT1-weightedMPRAGEvolumewereremovedwithFMRIB'sBrainExtractionTool(Smith,2002).Beforestatisticalestimation,thefollowingpreprocessingstepswereundertaken:therst4volumesoffunctionaldatafromeachrunwerediscarded,brainvolumesweremotioncorrectedtothemiddlevolumeusingtheOxfordCentreforFunctionalMagneticResonanceImagingoftheBrain(FMRIB)LinearImageRegistrationTool(MCFLIRT)(Jenkinsonetal.,2002).SpatialsmoothingwasappliedusingaGaussian-weightedkernelof5mmatfull-widthhalf-maximum,anddatawerehigh-passlteredtoremovelineartrends.EstimatesofthedegreesoffreedominthestatisticalmodelwerecorrectedforautocorrelationinthedatabyusingtheFSLpre-whiteningtechnique(Woolrichetal.,2001).Onlyrunswithlessthan2mmmaximalabsolutemovementswereincluded(53childrunsand21adultruns).TodealwithremainingnoiseduetoexcessivesubjectmovementnotdealtwithbycorrectionforheadmotionDiedrichsenandShadmehr,2005),themeanabsolutedeviationfromthemedianwascalculatedforeachvolumeusingAfni(3DToutcount,http://afni.nimh.nih.gov/afni/).Ineachrun,thevolumeswiththefteenmostextremevalueswereidentiedandmodeledinthedesignmatrixasregressorsofnointerest.Delaysandundershootsinthehemodynamicresponsewereaccountedforbyconvolvingthemodelwithadouble-gammabasisfunction.Functionalimageswereregisteredtothehigh-resolutionT1weighted3DMPRAGEusingthelowresolution3DMPRAGEacquiredinthesameplaneasthefunctionalimages.Thehigh-resolutionstructuralT1-weightedEPIvolumewasregisteredtotheMontrealNeurologicalInstituteTalairachcompatibleMRatlasaveraging152normalsubjectsusingFMRIB'sLinearImageRegistrationTool(FLIRT).Childbrainnormal-izationisanacceptedmethodindevelopmentalfMRIstudiessincetotalcerebralvolumedoesnotchangesignicantlywithageafter5yearsofagetoadulthood(Klingbergetal.,2002;Passarottietal.,2007;Reissetal.,1996)andfrom6yearsofage,standardnormalizationproceduresdonotleadtoartifacts(Kangetal.,2003;Muziketal.,2000).Thereforedatafrom6-to10-year-oldsandadultscanbeeffectivelytransformedintothesamestereotacticspace.fMRIanalysesAnimalandtoolblocksweremodeledasregressorsofinterestinthedesignmatrixwithrespecttoascrambledimagebaseline.EachfunctionalrunforagivensubjectwasmodeledseparatelyattheT.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 level.Statisticsforthecontrastsofinterest,averagingacrosstherunsofeachsubject,wereestimatedusingxedeffectsmodeling.Atthegrouplevel,random-effectscomponentsofmixedeffectsvarianceweremodeledandestimatedforeachcontrastofinterestusingFLAME1(Beckmannetal.,2003).Toidentifysignicantclustersofactivation,allZ-statisticimages(GaussianizedT/F)werethresholdedatanuncorrectedvoxelthresholdofz=2.3,p=0.01.Correctionformultiplecomparisonswasperformedattheclusterlevelbyapplyingaclustersizeprobabilitythresholdofp0.05totheZ-statisticimages.Weonlyreportclustersthatsurvivedthisclustercorrectedthreshold.Weincludedadultsinourstudyasamaturebenchmarkfordevelopment,butanumberofchangesthatarenotofinterestwhenstudyingthedevelopmentofcorticalobjectprocessing,suchaseye-movementsandcognitivestrategiesthatareunrelatedtoobjectprocessing,motion,respirationdifferences,andcardiacactivity,canaffectthemagnitudeorextensionoftheBOLDresponse.Eachofthesefactorshasthepotentialtoleadtoartifactualdifferencesbetweenagegroups.WedealtwiththesecommonchallengesofdevelopmentalfMRIbyusingarobustblockeddesign,byminimizingeye-movements,andbyapplyingstringentcorrectionsformotionartifacts(seeprevioussection).Inaddition,weperformedmultiple,complementarytypesofanalysesinwhichwetrackedchangesinactivationfrom6yearsofagetoadulthoodandinvestigatedchangesthatoccurredduringchildhoodseparatelyfromtheadultgroup.Weidentiobjectsensitiveregions(tools+animalstool-preferringregionsversusanimalsscrambled)andanimal-preferringregions(animalsscrambledscrambled)ineachagegroup.Weexploredwheretheobjectsensitiveandcategory-selectiveBOLDresponsevariedwithageinwholebraincorrelationanalysesaswellasina-priori,structurallynedregionsinwhichweexpecteddevelopmentalchangesinobjectprocessingtotakeplacebasedonbothdevelopmentalandadultliteratureontoolandobjectprocessing(ChaoandMartin,2000;Golaraietal.,2007;Grill-Spector,2003;KonenandKastner,2008;RizzolattiandMatelli,2003;SerenoandMaunsell,1998;Valyearetal.,).ThesestructuralROIsencompassedthefusiformgyrus,theparietalcortex,theinferiorlateraloccipitalcortexandthesuperiorlateraloccipitalcortex/posteriorparietalcortex.Becausenormaliza-tionfromagesixonwardsdoesnotleadtoartifacts(Muziketal.,)wedidnotmanuallydrawthebordersofanatomicalROIsineachindividualanatomicalscanbutdenedtheminstandardspaceafternormalizationwasperformed,ashasbeendoneinseveralotherdevelopmentalfMRIstudies(Passarottietal.,2003,2007;Nelsonetal.,2003).ThefourstructurallydenedROIsconsistedof(1)thecombinedoccipitalandtemporalfusiformcortexatlasregionmasks,(2)thecombinedsuperiorparietalcortexandsupra-marginalgyrus,anteriordivisionatlasregionmasks(3)theinferiorLOCatlasregionmaskand(4)thesuperiorLOCatlasregionmaskderivedfromtheprobabilisticHarvardOxfordCorticalStructuralAtlas(probabilisticmaskswereconvertedtobinarymasks).Tofurtherinvestigatethedevelopmentofcategory-selectiveregionsinthedorsalandventralobjectsensitivecortex,wecomparedtheanimalandtoolpreferenceof6-to7-year-oldchildren,8-to10-year-oldchildren,andadultsinindependentlydenedfunctionalregionsofinterests.Finally,weexploredwhereinthebrainandinthedescribeda-prioridenedstructuralROIs,individualdiffer-encesin(1)toolandanimalpreferenceand(2)objectsensitivity,correlatedwithageandwithindividualscoresonanindependenttoolrecognitiontaskbetweenage6to10.Thebehavioralresultsacquiredfromthesamesubjectsthatweresubsequentlyscannedarepresentedrst,followedbyfMRIresults.TheunusualperspectivetaskThepercentageofgraspableobjectsthatwerecorrectlyrecognizedfromanunusualviewpointisplottedagainstageinFig.1C.Thegraphshowsthataccuracyincreasedoverallwithagefrom6yearsto34yearsofage(R=0.49,F(1,43)=40.33,p0.001).Aregressionanalysesshowedthatrecognitionofgraspableobjectsfromanunusualperspectiveimprovedmarkedlyduringchildhood,betweenage6and10(R=0.25,F(1,32)=10.406,p=0.003)butdidnotimprovefurtherinthegroupofadults(R=0.032,F(1,10)=0.009,0.92).Thehighadultscoresonthistask,whichwasspecidesignedforchildren,mayreectaceilingeffectratherthanstabilizationoftheabilitytorecognizegraspableobjectsfromunusualviewpointsinadulthood.ThedevelopmentofobjectsensitivecortexWeexploredhowthedistributionofobjectsensitivecortex,irrespectiveofitscategorypreference,changesbetweenage6andadulthood(contrastingtools+animalswithscrambledimages).Collapsingacrossallforty-fourparticipants,weobtainedamapofobjectsensitivecortexthatextendedfromtheoccipitalcortexintothefusiformgyrusandintothelateraloccipital-andmedialtemporalcortexanddorsallyintotheleftsuperiorparietalcortexextendingtotheanteriorintraparietalsulcus(rightpanelofFig.2A).Objectsensitiveregionsaredisplayedseparatelyfor6-to7-year-olds,8-to10-year-oldsandadultsintheleftthreepanelsofFig.2WeusedawholebraincorrelationanalysistodetermineifthecorticalBOLDresponsetoobjectsvariedlinearlywithageinanyregion.Clusterswhereobjectsensitivitychangedfromthe6thyearoflifetoadulthoodwerelocatedinthebilateralinsularcortex,posteriorcingulateandvisualcortex.Intheinsular,theresponsetoobjectsincreasedwithage;childrenshowedapreferenceforscrambledimagesvsobjectsandadultsapreferenceforobjectsvsthescrambledimagebaseline.Conversely,intheposteriorcingulatetheresponsetoobjectsreducedwithage,withapreferenceforobjectsoverscrambledimagesinchildhoodbutastrongerresponseduringthebaselineinadulthood(Fig.2B,left).Finally,atallagestherewasapreferenceforthebaselineintheoccipitalpoleandthelingualgyrus,butthispreferencereducedwithage(Fig.2B,right).Theposteriorcingulatecortexandtheinsulargyrusareoftenlinkedtotheregulationofcognitiveresourcesandthedefault-network(1996;FranssonandMarrelec,2008;Haydenetal.,2008)wethereforesuggestthatthepatternsofresponseintheseregionsreectage-relateddifferencesincognitivestrategiesduringthetask.Forexample,thereducingresponseintheposteriorcingulatemayretheallocationoffewerresourcestoscrambledimagesvsobjectsinadultsthaninchildren,presumablybecauseadultsweremorecientatthepassiveviewingtask,i.e.,theyengagedlesswiththestimuli.Likewise,theactivationdifferencesinornearprimaryvisualcortexmayhavereectedslightdifferencesinxationstrategies.Thus,wehaveidentiedseveralregionswheretheresponsetoobjectscorrelatedlinearlywithage.Inlaterdiscussionsofage-relateddifferencesintheresponsetoobjectsinchildhoodalone,weexploreagedifferencesinobjectprocessingwithamorecomplexdevelop-mentalpattern.ThedevelopmentofacorticalpreferencefortoolsandanimalsWenextperformedmultiplecomplementaryanalysestoinvesti-gatehowcorticalpatternsofthepreferencefortoolsandanimalsinthedorsalandventralstreamchangefromthe6thyearoflife(contrastingtoolsscrambledtoobtaintool-preferringregionsandanimalsscrambledtoolsscrambledtoobtainanimal-preferringregions).Collapsingacrossallparticipantsweobtainedanaveragemapofcortexwithatoolpreference.AscanT.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 beseenintherightpanelofFig.3,dorsaltool-selectiveregionswerelocatedinthebilateraldorsaloccipitalandparietalcortex,extendingintoLIP,VIP,AIP,thebilateralprecentralsulcusnearthefrontaleyeelds,therightinferiorfrontalgyrus,parsopercularisandtriangularisandthefrontalpole.Ventraltool-selectiveregionswerelocatedintheleftandrightLOCadjacenttothemedialtemporalgyrusandinthemoremedialaspectsofthebilateralFFG.AnimalselectivecortexextendedfromV1intothefusiformgyrus.Ontheright,thisextensionencompassedregionsintheLOCandFGGthatarereportedtobeselectivetofacesaswell(Grill-Spector,2003Adultdorsalandventraltool-selectiveregionswerealreadyunmistakablypresentin6-and7-year-oldchildren,andshowedasimilarspatialorganizationineachagegroup.Corticalregionsthatshowedtooloranimalselectivityduringpassiveviewingaredepictedseparatelyfor6-to7-year-oldchildren,8-to10-year-oldchildrenandadultsintheleftpanelsofFig.3.Althoughatrstglancethereappeartobesomedevelopmentalchangesintoolpreference,forexampleinmedialFFGandposteriorparietallobe,weidentiednoregionswheretooloranimalselectivityvariedwithageinawholebraincorrelationanalyses.Evenwhenlimitingtheanalysestoa-prioridenedstructuralROIsthatencompassedallvisualizedregionswithatoolpreference,namely,parietalcortex,thefusiformgyrusandtheinferiorandsuperiorLOCwefoundnoregionswheretheBOLDresponsecorrelatedwithage.Wethusfoundnoevidencefordevelopmentalchangesinthedistributionofcategory-selectivecortexinthedorsalorventralvisualstreamincorrelationanalysisofthewholebrainorwithinstructurallydenedROIsatthresholdsadjustedforsmallervolumes(seeMaterialsandmethods).Compar-ingfunctionallydenedROIsacrossgroups(Scherfetal.,2007;Mahonetal.,2009;Golaraietal.,2007)isconsiderablymorepowerfulthanawholebrainanalysis,orthananalysesthatarerestrictedtorelativelylargestructurallydenedROIsandthusmaybeabletopickuponsubtledevelopmentaldifferences,ormorecomplexdevelop-mentalpatternsthanlinearactivationdecreasesorincreaseswithage. Fig.2.Onlyclustersexceedingathresholdofz=2.3,p=0.01,withaclustersizeprobabilityofp0.05aredepicted.A:Objectsensitiveregionsthatrespondmoretotoolsandanimalsthanscrambledpicturesaredisplayedseparatelyinred/yellowontheFreesurferaveragesurfacefor6-and7-year-olds,8-to10-year-oldsandadultsintheleftthreepanels.Intherightpanelallobjectsensitiveregionsaveragedoverall44subjectsisdisplayed.Regionsthatshowedsensitivitytoobjectscomparedtoscrambledpictureswerelocatedinthebilateralfusiformgyrus,thelateraloccipitotemporalcortex,themedialtemporalcortexandintheleftandrightinferiorparietalcortexandtheleftinferiorfrontalgyrus.NotethatthethreeleftsurfacesindicateacomplexdevelopmentaltrendwithadecreasingresponsetoobjectsvsscrambledstimuliduringchildhoodandasubsequentincreasingresponseinadulthoodintheposteriorparietalcortexandtheFFG.ThispatternisfurtherdiscussedintheResultssectionTheneuralcorrelatesofageanddevelopingtoolrecognitionduringandthesupplementarymaterial.B:Clustersthatcorrelatedsignicantlywithagein44subjects,rangingfrom6to34yearsofageafteraredisplayedontheleftFreesurferaveragesurfacehemisphere.Theresponseintheposteriorcingulatedecreasedwithage(depictedinblue)whiletheresponseintheleftandrightinsularcortexincreasedwithage(notethattheinsularclusterspreadsoutwhenregisteredandpaintedontotheaverageinatedFreesurfersurface).Therewerealsoage-relateddifferencesintheV1response.T.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 InthenextsectionwethereforeadoptafunctionallydenedROIapproachtofurtherexploretheresponsepatternsinobjectsensitivecorticalregionswithacategorypreference.ROIanalysisofthecategorypreferenceinobjectsensitivecortexWithintheobjectsensitivecortex,weselectedROIsthatwereorganizedbycategoryatages6to7,ages8to10andinadulthood.Wetheninvestigatediftherewereage-relatedchangesintheresponsepatternsofthesefunctionallydenedROIs.Thisapproachallowedustodenevoxelswithacategorypreferenceusingaseparateagegroup(e.g.,aregionwithatoolpreferenceintheadultgroup)fromtheoneusedtotesttheexperimentalhypothesis(e.g.whetheragroupof6-to7-andagroupof8-to10-year-oldchildrendisplayatoolpreferenceinthisadultregion).Specically,wecontrastedtoolswithanimalsandviceversa(toolsscrambledandanimal-scrambled)foreachagegroupseparately,withinafunctionallydenedmaskoftheomnibusobjectsensitivecortex.TheomnibusobjectsensitivemaskisdepictedinFig.2A,rightpanelandwasobtainedbycontrastinganimals+toolsscrambledimages,averagingacrossallforty-foursubjects.WeidentiedallresultingclusterswithatoolandanimalpreferenceforeachagegroupandthenextractedfromtheseclusterstheaverageBOLDresponsetoanimalsandtools(bothcomparedtoascrambledpicturebaseline)foreachindividual.Forexample,inatool-preferringregionintheadultleftparietalcortex,weextractedtheresponsetotoolsandanimalsforallsubjects,includingallchildren.Subsequently,wetestedifthesameregionpreferredtoolstoanimalsin6-to7-and8-to10-year-oldchildrenusingsetsofANOVA's.Fig.4,theresponsestotoolsandanimalsineachROIaredisplayedforallagegroups.Ascanbeseeninthegraphs,allparticipantsshowedatoolpreferenceineachROI,eveniftheROIfromwhichtheresponsewasextractedwasdenedinadifferentagegroup.Atoolpreferencewasalreadypresentinwhatwilllaterbecomeanadulttool-specicregioninbothgroupsofchildren.Likewise,ROIsthathadapreferencefortoolsin6-to7-year-oldsand8-to10-year-oldsshowedapreferencefortoolsinadults.Inaddition,ANOVA'sshowedthattherewerenostimulus-by-ageinteractionsinanyoftheROIs(allF'saresmallerthan1.809,p=0.177).Wetherebyshowthatnosignicantdevelopmentalchangessuchasanincreasingfocalizationoftool-preferringresponseweretakingplaceoutsidethefunctionallydenedadultregions.Ananimalselectiveregionintheobjectsensitivecortexsurpassedthestatisticalthresholdonlyinadults.Inaccordancewithpreviousreports(Mahonetal.,2009),thisregionwaslocatedinthelateraloccipitalfusiformgyrus,somewhatanteriortowhathasbeenreportedasthefunctionaloccipitalfacearea(OFA)(Grill-Spectoretal.,2004).Inthegroupsofchildren,theaverageBOLDresponseinthisregion(seeFig.3A,rightbottomgraph)didnotshowapreferenceforanimals.Tofurtherexploretheemergenceofanimalselectivity,wecomparedtheaverageBOLDresponseof6-to7-year-olds,8-to10-year-oldsandadultsintheclusterwithananimalpreferencethatweidentiedinthemedialFFGintheomnibusanalysisofall44subjects(seeFig.3,right-mostpanel,themedialandmostanteriorblueregionintheinferiorviewoftherighthemisphere).Theregionshowedapreferenceforanimalsinadults(F(1,10)=12.190,p=0.006)butnotinchildren(F(1,14)=0.131,p=0.131,F(1,17)=0.751,p=0.398).Therewerenosignicantage×stimulusinteractionsinthegroupaveragefusiform Fig.3.Onlyclustersexceedingz=2.3,p=0.01withaclustersizeprobabilityofp0.05aredepicted.Regionswithatoolpreferencearedepictedinred/yellow(resultingfromthecontrasttoolsscrambled)andregionswithananimalpreferencearedepictedinblue(resultingfromthecontrastanimalsGroupaveragemapsaredisplayedfor6-to7-year-olds,8-to10-year-oldsandadultsseparatelyintheleftthreepanelsandforall44participantstogetherontheright.Regionswithatoolpreferenceintheomnibusanalysiswerelocateddorsallyinthebilateralsuperiorparietalcortexandextendedintotheanteriorinferiorparietalcortex,thebilateralprecentralsulcusinthefrontaleyeeldsandtherightinferiorfrontalgyrus,parsopercularisandtriangularisandfrontalpole.Moreventrally,clusterswithatoolpreferencewerelocatedintheleftandrightlateralfusiformgyrus(FFG)andlateraloccipitalcortex(LOC)adjacenttothemedialtemporalgyrus(MTG)andinthemoremedialaspectsofthebilateralfusiformgyrus(FFG).AlargeclusterwithapreferenceforanimalsovertoolswaslocatedinV1,extendingintothefusiformgyrus.Mostprominentlyontheright,thecorticalpreferenceforanimalsinV1,extendedintotheLOCandFGG.Theblueanimal-preferringregionlocatedbetweentwotool-preferringclustersontherightinferiorviewandthesmallblueanimal-preferringregionontherightlateralviewarelocatedinthefusiformfacearea(FFA)andtheoccipitalfacearea(OFA)(Grill-Spector,2003).Therewerenoregionswherethepreferencefortoolsoranimalscorrelatedwithageinawholebrainanalysisandcorrelationanalysesrestrictedtoa-prioridenedstructuralROIsinFFG,inferiorandsuperiorLOCandtheParietalcortex.T.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 animalarea(F(2,41)=1.058,p=0.357)ortheoccipitalanimalregiondetectedinadults(F(2,41)=1.763,p=0.184).However,theanimalpreferencewasdistinctlylessconsistentacrossage(inrightFFGandLOC)thanthetoolpreference,whichwasadult-likefromsixyearsonwards.TheneuralcorrelatesofageanddevelopingtoolrecognitionduringInthepreviousanalyseswealreadyestablishedthattherewereregionsinwhichtheresponsetoobjectscorrelatedlinearlywithage.However,theseanalyseswerenotsensitivetomorecomplex,non-linearpatternsofdevelopmentfromchildhoodtoadulthood.Inasetofanalyseswethereforeexploredchangesintheobjectsensitivecorticalresponseinthegroupof33childrenalone.Whenweexploredhowtheresponsetoobjects,irrespectiveofitscategorypreference(contrastingtools+animalswithscrambledimages),varieswithageacrossthewholebrainduringchildhood,weidentiedaclusterintheleftposteriorlateralsulcus.Whentheanalysiswasperformedatalowerthresholdwithina-prioridenedstructuralROIsthatencompassedregionswhereweexpecteddevelopmentaldifferencesinobjectprocessingtotakeplacebasedonpreviousreports(seeMaterialsandmethods;namelythefusiformgyrus,theparietalcortexandtheinferiorandsuperiorlateraloccipitalcortex)weidentiedaclusterintherightmedialfusiformgyrus.AscanbeseenFigs.5AandB,theresponsetoobjectsintheposteriorlateralsulcusincreasedwithageduringpassiveobjectviewing,andtheresponsewasalsopositivelycorrelatedwithperformanceontheunusualviewpointtask.Theresponseinthemedialrightfusiformgyrus,incontrast,decreasedwithage,andinthisregionareducedresponsetoobjectscorrelatedwithbetterperformanceontheindependentobjectrecognitiontask.WeperformedpartialcorrelationanalysestodisentangletheuniquecontributionofageandobjectrecognitiontotheBOLDresponsetoobjectsvsscrambledstimuliinrFFGandtheposteriorlateralsulcus.ActivityinbothrFFGandposteriorlateralsulcuswascorrelatedwithageaftercorrectingforrecognitionperformance Fig.4.Clusterslocatedinobjectsensitivecortex(functionallydenedbycontrastingtools+animalsscrambledpicturesinall44subjects,seeFig.2Arightpanelformask)withapreferencefortoolsoranimalswereidentiedseparatelyinadults,8-to10-year-oldsand6-to7-year-olds,atathresholdedofz=2.3,p=0.01,p0.05.Inallagegroups,welocatedobjectsensitiveregionswithatoolpreferenceintheleftparietalcortexandtheleftandrightMTG/LOC.InbothgroupsofchildrentherewereadditionalclustersintheleftmedialFFG.TheadultgroupshowedaclusterwithasignicantanimalpreferenceintherightLOC.AdultregionsofinterestwithatoolpreferencearedepictedontheFreesurferaverageinatedcortexinorange,andregionswithananimalpreferencearedepictedinblue.Intheadultmaps,theedgesofthecorrespondingregionsofinterestsinthegroup8-to10-year-oldsaredepictedingreenandthoseofthe6-to7-year-oldsaredepictedinpurple.Inthegraphs,theaveragebeta-valuestoanimalsandtoolsaredisplayedseparatelyforeachgroup,foreachregionofinterest.Blackstarsindicateasignicantcategorypreference(allF's5.6,allp's0.03).ThetworedstarsinthemedialFFGindicateamarginallycantcategorypreference(topgraph:F(1,10)=4.308,p0.065,bottomgraphF(1,10)=4.318,p0.064).Notethatnostatisticalcomparisonsoftheanimalvstoolresponseweremadeforthegroupinwhichtheregionwasdened,becausetheydiffersignicantlyperdenition.ANOVA'sshowedthattherewerenostimulus-by-ageinteractionsinanyoftheROIs(AllF'saresmallerthan1.809,p=0.177).Thegraphsshowthatfromsixyearsonwards,thetoolpreferenceinthedorsalandventralobjectsensitivecortexisstrikinglyconsistent.Conversely,intheregionwithananimalpreferencethatisonlypresentintheadultgroup,neitherofthechildagegroupsshowedasignicantpreferenceforanimalscomparedtotools.T.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 (correlationpartialwithrFFGresponse=0.394,p=0.026andwithposteriorlateralsulcusresponse=0.583,p=0.001),butonlytherFFGwascorrelatedwithrecognitionperformanceaftervarianceduetoagedifferenceswaspartialledout(correlationpartialperformancewithrFFGresponse=0.348,p=0.051andwithposteriorlateralsulcusresponse=0.067,p=0.714).Fromthis,weconcludethatprotracteddevelopmentofcomplexobjectprocessingduringpassiveviewing,withadecreasingresponsetoobjectsintherightfusiformgyrusandanincreasingresponsetoobjectsinposteriorlateralsulcus,iscorrelatedwithageandthedevelopmentofobjectrecognitionfromunusualviewpointsthatoccursbetweenthe6thand10thyearoflife.However,becauseobjectrecognitionperformanceexplainednovarianceinthe Fig.5.A:InawholebraincorrelationanalysisoftheBOLDresponsetoobjects(passivelyviewedanimals+tools)andageinthegroupof33children,activationinaclusterintheleftposteriorlateralsulcuscorrelatedpositivelywithage.WelocatedaclusterintherightmedialFFGthatcorrelatednegativelywithagewhenrestrictingtheanalysistotheFFG.BothregionsaredepictedonanaverageFreesurferSurface.B:AverageindividualresponsestoobjectsversusscrambledobjectsintheclustersdepictedinpanelAareplottedagainstage(left)andperformanceontheunusualperspectivetask(right).RegressionanalysesshowthattheBOLDresponseinbothregionsexplainsasignicantpartofthevarianceinperformanceontheunusualperspectivetaskthatparticipantsperformedoutsidethescanner(posteriorlateralsulcusgyrus:R=0.139,F(1,31)=4.99,p=0.033;medialFFG:=0.279,F(1,31)=12,01,p=0.002;notethatthecorrelationbetweenageandtheBOLDresponseintheseregionsissignicantperdenition).However,incontrasttotheresponseintherFFG,theleftposteriorlateralsulcusclusternolongercorrelatedwithperformanceaftercorrectingforage.C:TwobilateralinferotemporalregionsandtwobilateralintraparietalregionsinwhichtheBOLDresponsetoobjectscorrelatednegativelywithaccuracyontheunusualperspectivetaskin33childrenaredisplayedontheFreesurferaveragesurface.D:TheaverageindividualresponsetoobjectsintheregionsinpanelCisplottedagainstageandperformance.Intheventralinferotemporalregions,therewasatrendtowardsanage-relateddecreaseinresponse(rightR=0.09,F(1,31)=3.05,p=0.09;leftR=0.112,F(1,31)=3.90,p0.057,trendindicatedbyredstar)andtherewasacantage-relateddecreaseintheintraparietalregions(rightR=0.160,F(1,31)=5.90,p=0.021;leftR=0.712,F(1,31)=6.46,p=0.01).NotethatthecorrelationsbetweenperformanceandtheBOLDresponseinthesefourregionsaresignicantbydenition.Partialcorrelationsconrmedthataftercorrectingforagethesefourregionsstillcorrelatedwithperformanceontheusualperspectivetask.Agedidnotexplainsignicantadditionalvarianceoverandaboveperformance.T.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 activationoftheposteriorlateralsulcusoverandaboveage,itispossiblethatthedevelopmentalchangesinthisregionreectmoregenericcorticalchangesbetween6and10years,andisnotnecessarilydirectlyrelatedtoobjectrecognition.InlinewiththefunctionalROIanalysisintheprevioussection,wefoundnoregionswhereatool-oranimalpreferenceintheBOLDresponsecorrelatedwithageduringchildhoodinawholebrainorstructurallydenedROIanalysis.Finally,weinvestigatedwhereinthebrain,theresponsetoobjectscorrelatedwithperformanceontheunusualperspectivetaskbetweenthe6thand10thyearoflife.Inawholebrainanalysis,weidentifourregionswheretheBOLDresponsedecreasedsignicantlywithimprovingperformance.Twowerelocatedintheleftandrightventralfusiformregionandtwointheleftandrightintraparietalsulcus(seeFigs.5CandD).Therewasastrongtrendtowardsadecreasewithageintheleftandrightfusiformresponse,andtheresponsetoobjectsintheleftandrightintraparietalcortexdecreasedsignicantlybetween6and10yearsofage(trendsareindicatedbyredstarsinFig.5Objectrecognitionskillsexplainedasubstantialamountofvarianceintheresponseintheseregionsoverandaboveage(partialcorrelationwerebetween0.369(leftintraparietal)and0.636(rightlFFG)withpvaluessmallerthan0.04).Thissuggeststhatthedecreasingresponseintheseregionsmaynotsimplyreectgenericdevelop-mentaleffectsbutisindeedrelevanttoimprovingobject-processingability.Agenolongercorrelatedsignicantlywiththeresponsesintheseregionsaftercorrectingforobjectrecognitionperformance(allpartialcorrelationsweresmallerthan0.220withpvalueslargerthan0.226).Aregressionmodelthatcombinedtheaverageresponsetoanimals+toolsscrambledinallfourregionsexplained50%ofthevarianceinperformanceontheunusualperspectivetask(R=0.495),whileonly25%ofthevariationinscoreswasaccountedforbyincreasingage(R=0.251).Weconcludethatchildrenwhoperformedbetterontheunusualperspectivetaskoutsidethescannerprocessedthepassivelyviewedstimuliinthescannerdifferently,whichexpresseditselfasalowerBOLDresponsetoobjectsinintraparietalandinferotemporalcortex.TherewerenosexdifferencesintheaverageresponseinanyoftheclustersreportedinFig.5.Wealsofoundnoage×stimulusinteractionsinthefourdorsalandventralregionsthatcorrelatedwithperformanceontheunusualperspectivetask(allfourF's0.596,p0.523).Moreover,inawholebraincorrelationanalysiswedidnotidentifyanyregionswhereatool(oranimal)preferenceintheBOLDresponsewascorrelatedwithperformanceontheunusualperspectivetaskduringchildhood.WethusshowthattheBOLDresponsetopassivelyviewedtoolsandanimalspredictperformanceontheunusualperspectivetaskandageofthechildandinasimilarway.Thissuggeststhatthecorticalprocessesthatunderliethedevelopmentofthisobject-processingskillbetweenage6and10,arenotspecictotoolsbutaremoreobject-general.Contrarytowhatwefoundwithchildren,adultperformanceontheobjectrecognitiontaskwasnotrelatedtothebold-responseintheinferotemporalandintraparietalregionsinFig.5C(Pearson'sr0.145)andthedecreasingresponsetoobjectsthatwereportedbetween6and10yearsofagedidnotpersistintoadulthoodeither.Instead,theresponsetopassivelyviewedobjectscomparedtoscrambledpicturesincreasedagain,resultinginaU-shapeddevelop-mentaltrajectory(seeSupplement1).ThispatternisvisualizedintheleftthreepanelsofFig.2A,wheretheresponsetoobjectsvsscrambledimagesisdepictedseparatelyforeachagegroup.Thereisavisibledecreaseinresponseintheintraparietalcortexandfusiformregionsfromtheyoungertooldergroupsofchildren,andasubsequentincreaseinresponseinadults.Itishighlyunlikelythatthispatternissimplyduetomotiondifferencesfortworeasons.Firstlywewouldexpectmotionartifactstonotonlyappearinregionsthataretypicallyassociatedwithobjectprocessing,andsecondlybecauseinchildrentheresponseintheseregionsishighlypredictiveofperformanceonacompletelyindependentobject-processingtaskthatparticipantsperformedoutsidethescanner.InthepresentstudyweexploredwhichaspectsoftheBOLDresponsetoobjects(tooloranimal-preferringand/orobject-general)indorsalandventralcorticalregionsshowedprotracteddevelopmentduringchildhood.Inaddition,weexploredwhichdevelopmentalchangesincorticalobjectprocessingrelatedtothedevelopingabilitytorecognizeobjectsfromunusualviewpoints.rstreplicatedBovaetal.'snding(2007)thattheabilitytorecognizegraspableobjectsfromanunusualperspectiveimprovesrapidlyfromthe6thuntilafterthe10thyearoflifeandweshowedthattheabilityhasimprovedevenmorebytheearlytwenties.Performancedidnotimproveafterthe20thyearoflife,suggestingthattheabilitytorecognizefamiliargraspableobjectsfromanunusualviewpointstabilizessometimeduringtheseconddecade.However,itremainspossiblethattheadultsperformedatceilingwiththesetestitemsthatwerespecicallydesignedtocapturedevelop-mentalchangesinchildhood.Therefore,theseresultsdonotexcludethepossibilitythattheabilitytorecognizeobjectsfromunusualperspectivescanstillimprovewithincreasingobjectexperienceinadulthood.Wedidnotidentifyanyage-relatedchangesintool-selectivetissueinanyof(1)awholebrainanalysis,(2)astructurallydeROIanalysis,or(3)afunctionallydenedROIanalysis.Weclearlyshowthatatanageasyoungassix,merepassivetoolviewingpreferentiallyactivatesparietalregionsofthecortex,includingAIP,withoutanyexplicitrequirementstograsp.Wedidnotndanyevidenceforprotractedtuningofanyregionswithatoolpreference(Hypothesis1).Wethereforeconcludethatthedorsalandventralstreamtoolnetworkshowsanadult-likespatialdistributionandtoolpreferencerelativelyearly(Hypothesis2).Thisisinlinewithpreviousthatapreferenceforothernon-socialobjectsintheventralstreamisadult-likeby5yearsofage(Golaraietal.,2007;Scherfetal.,).Becauseatoolpreferenceduringpassiveobjectviewingwasalreadypresentandorganizedinanadult-likewayinouryoungestagegroup,wecouldnotexplorethedevelopmentalrelationshipbetweenthedorsalandventralvisualstream.Futurestudieswillneedtotrackthedevelopmentofthetoolpreferenceinthecortexatyoungeragestoestablishwhethercategoryspecializationinthetwostreamsdevelops(1)intandemor(2)whetherdevelopmentoforganization-by-categoryinthedorsalstreamprecedesandperhapsdrivesthedevelopmentoftheventralstream(Mahonetal,2007;Mounoudetal.,2007)or(3)whetherthedevelopmentoforganization-by-categoryintheventralstreamprecedesandperhapsdrivesthedevelopmentofthedorsalstream(Braddicketal,2003;Klaveretal.,2008Theconsistentorganizationofatoolpreferencefrom6yearsonwardscontrastswithpreviousreportsonthedevelopingcorticalpreferenceforfaces(Cohen-KadoshandJohnson,2007).ThetworegionswithananimalpreferencethatwedetectedintherightFFG(inanomnibusanalysisofall44subjects)andtheLOCinadultsareinfactoftenreportedasregionswithafacepreferenceandhavebeenshowntokeeptuningtofacesuntiltheteenageyears(Golaraietal.,).Comparedtothedevelopmentalcontinuitythatwefoundinregionswithapreferencefortools,therewasadistinctlylessconsistentpreferenceforanimalsacrossageinthesetwoanimal-preferringregions.Inadults,butnotinchildrentheFFGandLOCshowedasignicantanimalpreferenceintheBOLDresponse.However,directcomparisonofagegroupdifferencesdidnotreachstatisticalsignicance.Thispatternindicatesthatapreferenceforanimalsinthesecorticalregionsonlygraduallyemerges.FurtherexplorationofthepossibleprotracteddevelopmentofcorticalanimalprocessingandanimalfaceprocessinginLOCandFFGmayhaveimportantimplicationsfortheoriesonthedevelopmentofcorticalT.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 faceprocessing.Suchresearchmay,forexample,provideinsightintowhethertheprotracteddevelopmentofafacepreferenceintheseregionsisrestrictedtohumanadultfaces,orextendstotheprocessingofanimalfacesaswell.InthebilateralintraparietalcortexandtheFFG,wefoundregionswherethesensitivitytoobjects(irrespectiveofcategory)wascorrelatedwithperformanceontheunusualperspectivetaskthatimprovesbetweenage6and10.Specically,theresponsetoobjectsintheleftandrightintraparietalsulcusandregionsintheleftandrightinferotemporalcortexdecreasedwhenchildrengotbetteratrecognizinggraspableobjectsfromunusualviewpointsoutsidethescanner.Thestorybecomesmorecomplicatedwhencomparingobjectprocessinginchildrenandadults.Inadults,theresponseintheabove-mentionedregionswasnotrelatedtoperformanceontheunusualperspectivetask.Inaddition,theresponsetoobjectsintheintraparietalcortexandfusiformregionsincreasedafter10,resultinginaU-shapedpatternofBOLDresponseacrossage.Wesuggestthatthiscomplexdevelopmentalpatternmaybetheresultofdifferentcognitivestrategiesemployedbyadultsandchildren.Indeed,U-shapedpatternsofbehavioraretypicallyassociatedwithchangesincognitivestrategy(Karmiloff-Smith,1992;Karmiloff-SmithandInhelder,1978;Seigler,2004).Here,forexample,adultsmaybemoreprocientatthepassiveviewingtaskandallocatefewerresourcestopassivelyviewedscrambledstimulivsobjects.Develop-mentalpatternsinregionsthataretypicallyassociatedwithregulationofattentionsuchastheposteriorcingulatelendsomesupporttothisidea.Possibly,thedecreasingresponsetoobjectswithimprovingperformanceinchildhoodreectsthatchildrenwhoarebetteratrecognizingobjectsprocessthestimuliinthescannermoreefleadingtoadecreaseintheinferotemporalandintraparietalBOLDresponsetoobjectsingeneral.Adevelopmentaldecreaseinactivationwithageandimprovingperformanceisfrequentlyreportedinthedevelopmentalliteratureandisgenerallyattributedtomoreeforsparseprocessing(Caseyetal.,2005;Durstonetal.,2006;Poldrack,).Althoughtheneuralmechanismsthatresultindecreasingactivationwithdevelopmentremaintobeelucidated,Peelenetal.usedanadaptationparadigmtoshowthatdevelopmentalchangesintheFFAwerelinkedtoincreasinglyselectiveinternalrepresentations,andthusprovidesomeevidencethatthisisaplausibleexplanation.Thevolumeofthebraindoesnotchangemuchafterage6,butdynamicchangesinlongandshortrangeconnectivityandtheratioofgraytowhitematterdensitycontinueuntillateinadolescenceandaremostpronouncedintheprefrontalcortex,theinferotemporalcortexandtheintraparietalcortex(Caseyetal.,2005;Fairetal.,2007;Gieddetal.,1999;Gogtayetal.,2004).Inlinewiththeprotracteddevelopmentofparietalandinferotemporalregionsduringchildhoodourndingssuggestthattheintra-andinter-regionalnetworksthatsupportobjectrepresentationsindorsalandventralhigh-levelvisualregionscontinuetone-tuneduringchildhoodJohnson,2001,2010).Achallengeforfutureresearchwillbetodisentangletherolethatexperienceandbrainmaturationplayintuningthesenetworksforobjectprocessing.Insummary,weinvestigatedthreehypotheseswithregardtothedevelopmentofobjectprocessingduringhumandevelopment.Weobtainedsomeevidencefortheprotracteddevelopmentaltuningofcertainanimalselectiveregions(Hypothesis1),butwedidnotobserveanysuchdevelopmentalchangesinthoseregionswithatoolpreference,whichlookedadult-likefromatleast6years(Hypothesis2).Therelativestabilityofneuralrepresentationsoftoolsacrossagesisconsistentwiththecriticalrolethatobjectsandactionsplayinmanytheoriesoflearningandeducation,namelybootstrappingthedevelopmentofknowledgefromearlyagesonwards(Bruner,1966;Mounoudetal.,2007;Piaget,1952,1954;Vygotsky,1934,1978).Inthelightofthesetheoriesonemightindeedexpectthatbyage6,childrenhavetheneuralmachineryinplacetolearnabouttheworldthroughtheirinteractions.Theearlymatureparietalspecializationfortoolsthatwereporthereisinlinewiththis.Toexamineourthirdhypothesisweinvestigatedwhethertheresponsetobothtoolsandanimalsinobjectsensitivecortexshowedaprotracteddevelopmentwithageandincreasingobject-processingprociency.Theresultsforgeneralobjectsensitivecortexwereconsistentwiththishypothesisinthatweobservedadecreasingcorticalresponsetobothtypesofobjectsinthedorsalandventralhigher-levelobjectsensitivevisualcortexthatcorrelatedwithdevelopingobjectrecognitionabilitybetweenage6and10.Supplementarymaterialsrelatedtothisarticlecanbefoundonline16/j.neuroimage.2010.11.005AcknowledgmentsWewouldliketothankthechildrenandfamilieswhotookpartinourstudyfortheirenthusiasmandtime.WewouldalsoliketothankRobertLeechforhisvaluablehelpwiththedataanalyses.ThisworkwasfundedbyaMarieCurieCentreofExcellencegrantMEST-CT-2005-020725fromtheEuropeanCommission.MarkJohnsonisfundedbytheUKMedicalResearchCouncilG0701484.ReferencesArbib,M.A.,2005.Frommonkey-likeactionrecognitiontohumanlanguage:anevolutionaryframeworkforneurolinguistics.Behav.BrainSci.28,105Augustine,J.R.,1996.Circuitryandfunctionalaspectsoftheinsularlobeinprimatesincludinghumans.BrainRes.Rev.229Aylward,E.H.,Park,J.E.,Field,K.M.,Parsons,A.C.,Richards,T.L.,Cramer,S.C.,Meltzoff,A.N.,2005.Brainactivationduringfaceperception:evidenceofadevelopmentalchange.J.Cogn.Neurosci.17(2),308Barsalou,L.W.,2008.Groundedcognition.Annu.Rev.Psychol.59,617Beauchamp,M.S.,Lee,K.E.,Haxby,J.V.,Martin,A.,2003.fMRIresponsestovideoandpoint-lightdisplayofmovinghumansandmanipulableobjects.J.Cogn.Neurosci.15(7),991Beckmann,C.F.,Jenkinson,M.,Smith,S.M.,2003.Generalmulti-levellinearmodelingforgroupanalysisinFMRI.Neuroimage20,1052Bova,S.M.,Fazzi,E.,Giovenzana,A.,Montomoli,C.,Signorini,S.G.,Zoppello,M.,Lanzi,G.,2007.Thedevelopmentofvisualobjectrecognitioninschool-agechildren.Dev.Neuropsychol.31,79Braddick,O.,Atkinson,J.,Wattam-Bell,J.,2003.Normalandanomalousdevelopmentofvisualmotionprocessing:motioncoherenceanddorsal-streamvulnerabilityNeuropsychologia41,1769Bruner,J.S.,1966.Oncognitivegrowth.In:Bruner,J.S.,Olver,R.R.,Greeneld,P.M.(Eds.),StudiesinCognitiveGrowth.Wiley.Casey,B.J.,Tottenham,N.,Liston,C.,Durston,S.,2005.Imagingthedevelopingbrain:whathavewelearnedaboutcognitivedevelopment?TrendsCogn.Sci.9,104Chao,L.L.,Martin,A.,2000.Representationofmanipulableman-madeobjectsinthedorsalstream.Neuroimage12487Cohen-Kadosh,K.,Johnson,M.H.,2007.Developingacortexspecializedforfaceperception.TrendsCogn.Sci.11,367Crookes,K.,McKone,E.,2009.Earlymaturityoffacerecognition:nochildhooddevelopmentofholisticprocessing,novelfaceencoding,orfacespace.Cognition111(2),219Diedrichsen,J.,Shadmehr,R.,2005.DetectingandadjustingforartifactsinfMRItimeseriesdata.Neuroimage27(3),624Durston,S.,Davidson,M.C.,Tottenham,N.T.,Galvan,A.,Spicer,J.,Fossella,J.A.,Casey,B.J.,2006.Ashiftfromdiffusetofocalcorticalactivitywithdevelopment.Dev.Sci.9(1),1Fair,D.A.,Cohen,A.L.,Dosenbach,N.U.F.,Church,J.A.,Miezin,F.M.,Barch,D.M.,Raichle,M.E.,Petersen,S.E.,Schlaggar,B.L.,2007.Thematuringarchitectureofthebrain'sdefaultnetwork.PNAS105(10),4028Fransson,P.,Marrelec,G.,2008.Theprecuneus/posteriorcingulatecortexplaysapivotalroleinthedefaultnetwork:evidencefromapartialcorrelationnetworkanalysis.Neuroimage42,1178Gathers,A.D.,Bhatt,R.,Corbly,C.R.,Farley,A.B.,Joseph,J.E.,2004.Developmentalshiftsincorticallociforfaceandobjectrecognition.NeuroReport15(10),1549Giedd,J.N.,etal.,1999.Braindevelopmentduringchildhoodandadolescence:alongitudinalMRIstudy.Nat.Neurosci.10,861Gogtay,N.,Giedd,J.N.,Lusk,L.,Hayashi,K.M.,Greenstein,D.,Vaituzis,A.C.,etal.,2004.Dynamicmappingofhumancorticaldevelopmentduringchildhoodthroughearlyadulthood.Proc.NatlAcad.Sci.USA101,8174Golarai,G.,Ghahremani,D.G.,Whiteld-Gabrieli,S.,Reiss,A.,Eberhardt,J.L.,Gabrieli,J.D.,Grill-Spector,K.,2007.Differentialdevelopmentofhigh-levelvisualcortexcorrelateswithcategory-specicrecognitionmemory.Nat.Neurosci.10(4),512522.T.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005 Golarai,G.,Liberman,A.,Yoon,J.M.D.,Grill-Spector,K.,2009.Differentialdevelopmentoftheventralvisualcortexextendsthroughadolescence.Front.Hum.Neurosci.3doi:10.3389/neuro.09.080.2009Grill-Spector,K.,2003.Theneuralbasisofobjectperception.Curr.Opin.Neurobiol.13,Grill-Spector,K.,Knouf,N.,Kanwisher,N.,2004.Thefusiformfaceareasubservesfaceperception,notgenericwithincategoryidentication.Nat.Neurosci.7(5),555562.Grill-Spector,K.,2009.Objectperception:physiology.In:Goldstein,B.(Ed.),EncyclopediaofPerception.SagePublications.Hayden,B.Y.,Nair,A.C.,McCoy,A.N.,Platt,M.L.,2008.Posteriorcingulatecortexmediatesoutcome-contingentallocationofbehavior.Neuron60,19Hasson,U.,Harel,M.,Levy,I.,Malach,R.,2003.Large-scalemirror-symmetryorganizationofhumanoccipito-temporalobjectareas.Neuron37,1027Haxby,J.V.,Gobbini,M.I.,Furey,M.L.,Ishai,A.,Schouten,J.L.,Pietrini,P.,2001.Distributedandoverlappingrepresentationoffacesandobjectsinventraltemporalcortex.Science293,2425Jenkinson,M.,Bannister,P.R.,Brady,J.M.,Smith,S.M.,2002.Improvedoptimisationfortherobustandaccuratelinearregistrationandmotioncorrectionofbrainimages.Neuroimage17(2),825Johnson,M.H.,2001.Functionalbraindevelopmentinhumans.Nat.Rev.Neurosci.2,Johnson,M.H.,2010.DevelopmentalCognitiveNeuroscience,3rdEd.Wiley-BlackwellPublishing,Oxford.Johnson,M.H.,Grossmann,T.,Cohen-Kadosh,K.,2009.Mappingfunctionalbraindevelopment:buildingasocialbrainthroughInteractiveSpecialization.Dev.Psychol.45,151Johnson-Frey,S.H.,2004.Theneuralbasisofcomplextooluseinhumans.TrendsCogn.Sci.8(2),71Joseph,J.E.,Gathers,A.D.,Bhatt,R.S.,2010.Progressiveandregressivedevelopmentalchangesinneuralsubstratesforfaceprocessing:testingspecicpredictionsoftheInteractiveSpecializationaccount.Dev.Sci,1doi:10.1111/j.1467-7687.2010.00963.xJuttner,M.,Muller,A.,Rentschler,I.,2006.Adevelopmentaldissociationofview-dependentandview-invariantobjectrecognitioninadolescence.Behav.BrainRes.175,420Kang,H.C.,Burgund,E.D.,Lugar,H.M.,Petersen,S.E.,Schlagger,B.L.,2003.Comparisonoffunctionalactivationfociinchildrenandadultsusingacommonsterotacticspace.Neuroimage19,16Karmiloff-Smith,A.,1992.BeyondModularity:ADevelopmentalPerspectiveonCognitiveScience.Mass.MITPress/BradfordBooks,Cambridge(Reprinted1995).Karmiloff-Smith,A.,Inhelder,B.,1978.Ifyouwanttogetahead,getatheory.In:Johnson-Laid,Ph.,Wason,P.(Eds.),Thinking:ReadingsinCognitiveScience.CambridgeUniversityPress,Cambridge,pp.293Klaver,P.,Lichtensteiger,J.,Burcher,K.,Dietrich,T.,Loenneker,T.,Martin,E.,2008.Dorsalstreamdevelopmentinmotionandstructure-from-motionperception.Neuroimage39,1815Klingberg,T.,Forssberg,H.,Westerberg,H.,2002.Increasedbrainactivityinfrontalandparietalcortexunderliesthedevelopmentofvisuo-spatialworkingmemorycapacityduringchildhood.J.Cogn.Neurosci.14(1),1Konen,C.S.,Kastner,S.,2008.Twohierarchicallyorganizedneuralsystemsforobjectinformationinhumanvisualcortex.Nat.Neurosci.11(2),224Mahon,B.Z.,Milleville,S.,Negri,G.A.L.,Rumiati,R.I.,Martin,A.,Caramazza,A.,2007.Action-relatedpropertiesofobjectsshapeobjectrepresentationsintheventralstream.Neuron55(3),507Mahon,B.Z.,Anzellotti,S.,Schwarzbach,J.,Zampini,M.,Caramazza,A.,2009.Category-corganizationinthehumanbraindoesnotrequirevisualexperience.Neuron63(3),397Mondloch,C.J.,LeGrand,R.,Maurer,D.,2002.Confaceprocessingdevelopsmoreslowlythanfeaturalfaceprocessing.Perception31,553Mondloch,C.J.,Geldart,S.,Maurer,D.,LeGrand,R.,2003.Developmentalchangesinfaceprocessingskills.J.Exp.ChildPsychol.86,67Mondloch,C.J.,Ahola,S.,Maurer,D.,2006.Becomingafaceexpert.Psychol.Sci.17,Mounoud,P.,Duscherer,K.,Moy,G.,Perraudin,S.,2007.Theinuenceofactionperceptiononobjectrecognition:adevelopmentalstudy.Dev.Sci.10,Muzik,O.,Chugani,D.C.,Juhasz,C.,Shen,C.,Chugani,H.T.,2000.Statisticalparametricmapping:assessmentofapplicationinchildren.Neuroimage12(5),Nelson,E.E.,McClure,E.B.,Monk,C.S.,Zarahn,E.,Leibenluft,E.,Pine,D.S.,Ernst,M.,2003.Developmentaldifferencesinneuronalengagementduringimplicitencodingofemotionalfaces:anevent-relatedfMRIstudy.J.ChildPsychol.Psychiatry44(7),Nishimura,M.,Scherf,S.,Behrman,M.,2009.Developmentofobjectrecognitioninhumans.F1000Biol.Rep.1,56.Passarotti,A.M.,Paul,B.M.,Bussiere,J.,Buxton,R.,Wong,E.,Stiles,J.,2003.Developmentoffaceandlocationprocessing:afMRIstudy.Dev.Sci.6(1),Passarotti,A.M.,Smith,J.,DeLano,M.,Huang,J.,2007.Developmentaldifferencesintheneuralbasesofthefaceinversioneffectshowprogressivetuningofface-selectiveregionstotheuprightorientation.Neuroimage34(4),1708Peelen,M.V.,Glaser,B.,Vuilleumier,P.,Eliez,S.,2009.Differentialdevelopmentofselectivityforfacesandbodiesinthefusiformgyrus.DevelopmentalScience12(6),doi:10.1111/j.1467-7687.2009.00916.xPelphrey,K.A.,Lopez,J.,Morris,J.P.,2009.Developmentalcontinuityandchangeinresponsestosocialandnonsocialcategoriesinhumanextrastriatevisualcortex.Front.Hum.Neurosci.3(25).doi:10.3389/neuro.09.025.2009Piaget,J.,1952.TheOriginsofIntelligenceintheChild.InternationalUniversitiesPress,NewYork.Piaget,J.,1954.TheConstructionofRealityintheChild.BasicBooks,NewYork.Poldrack,R.A.,2010.Interpretingdevelopmentalchangesinneuroimagingsignals.Hum.BrainMapp.31,872Reiss,A.L.,Abrams,M.T.,Singer,H.S.,Ross,J.L.,Denckla,M.B.,1996.Braindevelopment,genderandIQinchildrenavolumetricimagingstudy.Brain119,1763Rizzolatti,G.,Matelli,M.,2003.Twodifferentstreamsformthedorsalvisualsystem:anatomyandfunction.Exp.BrainRes.153,146Saygin,A.P.,Sereno,M.I.,2008.Retinotopyandattentioninhumanoccipital,temporal,parietal,andfrontalcortex.Cereb.Cortex18,2158Scherf,K.S.,Behrmann,M.,Humphreys,K.,Luna,B.,2007.Visualcategory-selectivityforfaces,placesandobjectsemergesalongdifferentdevelopmentaltrajectories.Dev.Sci.10(4),F15Siegler,R.S.,2004.U-shapedinterestinU-shapeddevelopmentandwhatitmeans.J.Cogn.Dev.5,1Sereno,A.B.,Maunsell,J.H.,1998.Shapeselectivityinprimatelateralintraparietalcortex.Nature395,500Smith,S.M.,2002.Fastrobustautomatedbrainextraction.Hum.BrainMapp.17(3),Thesen,S.,Heid,O.,Mueller,E.,Schad,L.R.,2000.Prospectiveacquisitioncorrectionforheadmotionwithimage-basedtrackingforreal-timefMRI.Magn.Reson.Med.44,Valyear,K.F.,Cavina-Pratesi,C.S.,Stiglick,A.J.,Culham,J.C.,2007.Doestool-relatedfMRIactivitywithintheintraparietalsulcusreecttheplantograsp?Neuroimage36,VygotskyToolandsymbolinchilddevelopment,1934.Vygotskymindinsociety:thedevelopmentofhigherpsychologicalprocesses,1978.Woolrich,M.W.,Ripley,B.D.,Brady,J.M.,Smith,S.M.,2001.TemporalautocorrelationinunivariatelinearmodellingofFMRIdata.Neuroimage14(6),1370T.Dekkeretal./NeuroImagexxx(2010)xxx Pleasecitethisarticleas:Dekker,T.,etal.,Dorsalandventralstreamactivationandobjectrecognitionperformanceinschool-agechildren,NeuroImage(2010),doi:10.1016/j.neuroimage.2010.11.005