Dissociating Cortical Regions Activated by Semanti

DissociatingCorticalRegionsActivatedbySemanticandPhonologicalTasks:AfMRIStudyinBlindandSightedPeopleH.Burton,1J.B.Diamond,1andK.B.McDermott21DepartmentofAnatomyandNeurobiologyand2DepartmentofPsychology,WashingtonUniversitySchoolofMedicine,St.Louis,Missouri63110Submitted24March2003;acceptedinfinalform1June2003Burton,H.,J.B.Diamond,andK.B.McDermott.Dissociatingcorticalregionsactivatedbysemanticandphonologicaltasks:AfMRIstudyinblindandsightedpeople.JNeurophysiol90:1965–1982,2003.FirstpublishedJune4,2003;10.1152/jn.00279.2003.Previousneuroim-agingstudiesoflanguageprocessinginblindindividualsdescribedcor-ticalactivationofprimary(V1)andhighertiervisualareas,irrespectiveoftheageofblindnessonset.Specifically,participantsweregivennounsandaskedtogenerateanassociatedverb.Theseresultsconfirmedthepresenceofadaptationsinthevisualcortexofblindpeopleandsuggestedthattheseresponsesrepresentedlinguisticoperations.Thepresentfunctionalmagneticresonanceimagingstudyattemptedtofur-thercharacterizetheseresponsesasbeingpreferentialforsemanticorphonologicalprocessing.Threegroupsofparticipants(sighted,early-onset,andlate-onsetblind)heardlistsofrelatedwordsandattendedtoeitheracommonmeaning(semantictask)orcommonrhyme(phonolog-icaltask)thatlinkedthewords.Inallthreegroups,thesemantictaskelicitedstrongeractivityintheleftanteriorinferiorfrontalgyrusandthephonologicaltaskevokedstrongeractivitybilaterallyintheinferiorparietalcortexandposterioraspectsoftheleftinferiorfrontalgyrus.Onlyblindindividualsshowedactivityinoccipital,temporal,andpari-etalcomponentsofvisualcortex.Thespatialextentofvisualcortexactivitywasgreatestinearlyblind,whoexhibitedactivationinallventralanddorsalvisualcortexsubdivisions(V1throughMT)forbothtasks.Preferentialactivationappearedforthesemantictask.Lateblindindivid-ualsexhibitedresponsesinventralanddorsalV1,ventralV2,VPandV8,butonlyforthesemantictask.Ourfindingssupportpriorevidenceofvisualcortexactivityinblindpeopleengagedinauditorylanguageprocessingandsuggestthatthisactivitymayberelatedtosemanticprocessing.INTRODUCTIONWhenblindpeoplereadBraille,neuralregionstraditionallyinvokedforvisualprocessingareactivated(Bu¨cheletal.1998a,b;Burtonetal.2002a;Cohenetal.1997,1999;Ham-iltonetal.2000;Melzeretal.2001;Sadatoetal.1996,1998).Somepriorstudiesshowthatindividualsbornblindexhibitactivityinprimaryvisualcortex(V1)andmultiplehighertiervisualareas(Cohenetal.1997,1999;Sadatoetal.1996,1998,2002).Conversely,somestudiesreportthatindividualswholosesightafterhavinglearnedtoreadprintshowstrongerV1activitythanthosebornblind(Bu¨cheletal.1998a,b;Melzeretal.2001).Theconclusionsthatcanbedrawnfromtheobserveddif-ferencesinV1activitythatvarywithageofblindnessonsetaretemperedbytaskdifferencesacrossstudies.SomestudiesrequiredorthographicprocessingofindividualBrailleletters(Cohenetal.1997,1999).ThestudythatreportedgreateractivityinV1inlateblindrelativetoearlyblindindividualsusedanorthographictaskthatrequiredparticipantstodetectwhetheranyBraillecellshaddot6elevatedinreadwordsorconsonantstrings(Bu¨cheletal.1998a,b);thedegreetowhichthisorthographictaskwasaccompaniedbysemanticprocessesinherentinwordreadingisunclear.StrongerV1activityinearlyblind(relativetolateblind)wasreportedinastudythathadearlyandlateblindparticipantsreadBraillewords(Sadatoetal.1996,1998).AnotherstudythatalsoinvolvedreadingBraillewordsreportedtheoppositefindingofgreateractivityinV1oflateblindpeople(Melzeretal.2001).We(Burtonetal.2002a)employedaverb-generationtasktoBraille-readnouns.Thesemanticdemandsinherentinthistask(Gabrielietal.1998;Segeretal.1999)mightthusunderliefindingprimaryandhighervisualcortexresponsesinallblindparticipantsirrespectiveoftheageofblindnessonset.Thefindingofdifferentpatternsofactivationwithinvisualcortexinblindandsightedindividualsshouldnotovershadowthecommonalitiesobserved.Forexample,leftinferioranddorsolateralfrontalcorticallanguageregionsrespondedsimi-larlyinblindpeoplereadingBrailleandsightedparticipantsreadingprint(Burtonetal.2002a).Thisresulthasseveralimplications.Reorganizationoflanguagewithinblindpeopleisselectivetocertaincorticalregions,andpossiblyjustwithinthevisualcortex.Thecontributionthatthevisualcortexmakestolanguagewithinblindpeoplemightbeacross-modalsensoryprocessthatoccursinBraillereadingbeforelanguageanalysesinfrontalcortex.Wepreviouslysuggestedthatvisualcortexactivityinblindpeoplereflectslexicalprocessesbecauseoffindingsimilarresponsepatternsinallblind,butnotsighted,peoplewhentheygeneratedverbsthatwereassociatedtoheardnouns(Burtonetal.2002b).VisualcortexresponsesinthisstudylargelyresembledthosenotedpreviouslywithBraillereading.ThecomparabilityofresultsacrossstudiesusingverbgenerationtoauditorilypresentednounsandBraillereadingAddressforreprintrequestsandothercorrespondence:H.Burton,Depart-mentofAnatomyandNeurobiology,CampusBox8108,WashingtonUniver-sitySchoolofMedicine,660SouthEuclidAve.,St.Louis,MO63110(E-mail:[e-mail]).Thecostsofpublicationofthisarticleweredefrayedinpartbythepaymentofpagecharges.Thearticlemustthereforebeherebymarked‘‘adverti