BehaviouralNeurology25(2012)23 - PDF document

BehaviouralNeurology25(2012)23–34DOI10.3233/BEN-2012-0346IOSPressUnpickingthesemanticimpairmentinAlzheimer’sdisease:QualitativechangeswithdiseaseseverityFayeCorbett,ElizabethJefferies,AlistairBurnsandMatthewA.LambonRalphUniversityofManchester,Manchester,UK Correspondingauthor:NeuroscienceandAphasiaResearchUnit(NARU),SchoolofPsychologicalSciences,ZochonisBuilding,Uni-versityofManchester,OxfordRoad,Manchester,M139PL,UK.Tel.:+441612752551;Fax:+441612752873;E-mail:matt.lambon-ralph@manchester.ac.uk.impairedneurologicalpopulationshaveallowedusto F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:Qualitativechangeswithdiseaseseverityearlystagesofthedisorder[5–7].AlthoughthereisavastliteratureexaminingsemanticmemoryinAD,noconsensusregardingthenatureofthedecithasyetbeenreached.Twopossiblecauseshavebeenidenti-ed;ontheonehand,ADhasbeencharacterisedasa‘storage’disorderinwhichrepresentationsarelostfromsemanticmemory.Proponentsofthisviewre-portapatternofgradualdeclineofknowledgeinAD,leadingtoaconsistentimpairmentofthesamecon-ceptsacrossdifferenttasks[8–13].Alternatively,oth-ershaveshownthatthelevelofsemanticimpairmentinthisconditionisnotstablebutvariesinlinewithchangingtaskdemands.Performanceonsemantically-demandinglexicalandconceptualdecisiontasks,forexample,hasbeenshowntoimprovesignicantlywiththeprovisionofconstrainingcues[14–17].ThisingisnotconsistentwiththenotionthatknowledgehasbeenpermanentlylostinADbutcanbeenexplainedintermsofaproblemwithaccessingthatinformation.Therearesomehintsintheliteraturethatthedispar-itybetweentheseviewpointscouldreectaseverity-relateddifference.Forexample,whileearlyADpa-tients’levelofsemanticknowledgeseemstovaryinlinewithexecutivedifculty[17],thesameisnottrueofmoreseverely-impairedpatientswhoexhibitgreaterconsistencyacrossdifferenttasks[18].Therelation-shipbetweenseverityandthenatureofsemanticim-pairmentinADhasnot,however,beenexamineddi-rectlyinpreviousstudies.Itseemslikelythatgradu-alspreadingofpathologyoverthecourseofADwillcompromiseseparablecomponentsofthesemanticnet-workatdifferenttimes.ThisviewanticipatesthatthequalitativenatureofsemanticimpairmentinADwillchangethroughoutthedisease.Bytakingseverityin-toaccount,therefore,itmightbepossibletoresolvethelongstandingconundrumregardingthenatureofsemanticimpairmentinthisdisorder.Wehavepreviouslyemployedacase-series,compar-ativedesigntoinspectsemanticimpairmentsarisingindifferentneurologicalgroups(e.g.,HSVEvs.semanticdementia(SD):[4];semanticaphasia(SA)vs.SD:[1].)SDisaneurodegenerativedisordercharacterisedbyaselectivesemanticdecitaffectingknowledgeacrossafullrangeofmodalitiesinbothexpressiveandrecep-tivetasks[19–23].Thisconditionoccursinthecon-textofatrophyfocussedontheATLbilaterally[24,25].Multimodalsemanticdecitsarealsoobservedinsomepatientsfollowingastroke[1–3].IncontrasttoSD,however,patientswithSAhavedamagetoleftpre-frontaland/ortemporoparietalregionswithcompletesparingoftheATL(see[3]).Inadirectcomparison,JefferiesandLambonRalph[1]foundthatthesetwogroupsfailedsemantictasksfordifferentreasons.TheATLatrophyinSDwasassociatedwithapatternofimpairmentindicativeofdegradedsemanticrepresen-tationsandtheirsemanticperformancewasimpervi-oustochangesintaskdemandandmodality.Greatestimpairmentwasobservedfortheleastfamiliar/lowfre-quencyconceptssuggestingthattheseitemsweremostvulnerabletodegradation.Incontrast,theSApatients’performancesuggestedthatconceptualknowledgewasnotdegradedinthesecases,butafailureofregulatorycognitivecontrolunderpinnedtheirimpairedsemanticcognition.TheSApatients’performancewasdistinctinthefollowingkeyways:a)theyweremuchlesscon-sistentacrosstasksandstronglyinuencedbytaskde-mands;b)inadditiontocoordinateandsuperordinatenamingerrors(e.g.,squirrel“rabbit”or“animal”),theSApatientsproducedassociativeerrors(e.g.,squir-“nuts”)whichwerealmostneverobservedinSD;(c)namingandcomprehensionscoreswereimprovedordiminishedbytheprovisionofcuesorotherformsoftaskconstraint(whichhadlittleornoeffectinSD);(d)theSApatientsexhibitedsemantic“access”andrefrac-torysymptomsincontrasttothe“storage”decitsofSD[26–28]and(e)theirperformanceoncomprehen-siontaskswaspredictedbythecontrolrequirementswithinanindividualtrial.Thesendingsareconsistentwiththeviewthattwointeractingcomponentssupportsemanticcognition:1)amodality-invariant,represen-tationalhubsupportedbytheATLbilaterally[29–32]and,2)anexecutivecontrolcomponentunderpinnedbyanetworkencompassingtheleftprefrontalcortexandregionsinthetemporoparietalcortex[1–3,33].For-malmeta-analysesoffunctionalneuroimaginginnor-malparticipantsindicatethat,inadditiontothesetworegions,cingulatecortexalsocontributestosemanticcognition[34].Asnotedbelow,thesethreeareasareimplicatedintheearliestphasesofAD.Aconsiderablevolumeofresearchhasbeendedicat-edtotracingdiseasepathogenesisoverthecourseofAD.Thesestudiesindicatethatthefunctionallysepa-rablecomponentsofsemanticcognitionarelikelytobecomecompromisedatdifferentstagesofthedisease.IntheearlieststagesofAD,amyloidplaquesandneu-brillarytanglesoccupythemedialtemporallobe(MTL)–acriticalregionsupportingepisodicmemo-ry[35–37].ThebroaderMTLnetworkincorporating,amongstotherregions,theposteriorcingulatecortexisalsofoundtobehypometabolicatthisearlystageofthedisease,withprobableimpactonepisodicmemoryaswellasexecutivecontrolfunctions[38–40].Beyond F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:QualitativechangeswithdiseaseseverityTable1Performanceonattentional/executiveandsemantictasks TaskMaxControlSAmeanMildADSevereADSDmeanmean(s.d.)meanmean Executive/attentionalWCST(categories)61*1.22.10.4NTElevatorcounting76*564NTElevatorcounting(withdistraction)102*‡251NTRavens’Matrices(percentiles)4/10cases50,4/10cases25,7/10cases5,9/10casesremainingcases1case50,remainingcasesremainingcase10remainingcasesPPTpicture5251.2(1.4)40423441PPTword5251.1(1.1)4145353964itembatteryNaming6462.3(1.6)21494027WPM6463.7(0.5)50574646CCTpicture6458.9(3.1)36382640CCTword6460.7(2.06)37463337EnvironmentalsoundsSpokenword–picture4847.8(0.6)39403033Sound–picture4841.2(2.5)28191122Sound–word4840.8(3.8)25171120Categoryuency(6)–95.7(16.5)14371031Letteruency–44.2(11.2)526821 *denotesnormalcut-off;NTnottested.‡normalcut-offforparticipantsaged65–80.theMTL,earlyADneuropathologycriticallyunder-minesthefrontal-temporoparietalcontrolnetworkintwokeyways:1)cellswithinthesuperiorlongitudinalfasciculusareselectivelyaffectedcausingdisconnec-tionoffrontalandtemporoparietalregions[3];and2)spreadingofpathologyintothebasalforebrain,par-ticularlythebasalnucleusofMeynert,underminesthecholinergicsystemthatnormallyprovidescriticalin-nervationtoareasoftheneocortexsupportingexecutivecontrol,includingprefrontalandparietallobesaswellasthethalamus[41].Consistentwiththesechanges,patientsintheearlystagesofADhaveexecutiveandattentionaldecits[42–45].Asthediseasereachesmoremoderatestages,generalisedhypometabolismofthetemporallobe(TL)isjoinedbyplaquesandtanglesinvadinginferolateralaspectsofthisregion[7,46]–notablythesameareasthatunderliethesemanticim-pairmentinSD[24,25,47,48].ThepresentstudyexaminedthehypothesisthattheevolutionofpathologyinADgeneratesqualitativelydifferentsemanticimpairmentsoverthecourseofthedisease.Intheearlystages,damagetothefrontal-temporoparietal-cingulatecontrolnetworkshouldpro-ducederegulationofsemanticcognition.Asneu-ropathologyspreadsintotheinferolateralaspectsoftheATL,inthemoreseverephaseofAD,semanticrepre-sentationsthemselveswillbegintodegrade,thusexac-erbatingthealreadyderegulatedsemanticsystem.As-sumingthishypothesisistrue,earlyADcasesshouldexhibitsemanticimpairmentthatischaracteristicallysimilartothatobservedinSA,whileadditionaltraitstypicalofSDshouldemergewhenthediseasereachesalaterstage.WetestedthishypothesisbyextendingJefferiesandLambonRalph’s[1]case-series,compar-ativedesigntoincludepatientswithAD.Indoingsowewereabletocomparethenatureofsemanticim-pairmentatmildandseverestagesofADagainstthesetwo“reference”groups.2.Method2.1.ParticipantsThisworkwasapprovedbylocalhealthauthorityethicscommitteesandinformedconsentwasobtainedforallparticipants.TwentypatientswithAlzheimer’sdisease(10mild,10severe)wererecruitedfromdementiaservicesinManchester,UK.Allcas-eswerediagnosedbyanexperiencedOldAgePsy-chiatrist(AB)withrespecttothestandard,interna-tionalconsensuscriteria.Patientswithahistoryofstrokeorotherneurologicalconditionswereexclud-ed.AllparticipantsinthemildADgroupscored19ormore(outof30)ontheMini-MentalStateExamination(MMSE[49]),whereasallseverecasesobtained14or F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:Qualitativechangeswithdiseaseseverityless(meanscore:mild21.3;severe0.001,seeTable1).Themildgroupwerenumericallybutnotsignicantlyolderthantheseveregroup.ThetwoADgroupswerecomparedwiththeSAandSDpatientgroupspreviouslyreportedbyJefferiesandLambonRalph[1]andBozeatetal.[50].ThetenSApatientswererecruitedfromstrokeclubsandspeechandlanguagetherapyservicesintheManchesterarea.SApatientswereidentiedonthebasisthattheyfailedbothwordandpictureversionsoftheCamelandCactussemanticassociationtask[50].Everycasehadchronicimpairmentfromastrokeatleastayearpriortotesting.Allpatientshaddamagetotheleftprefrontalcortexand/orareasinthetemporoparietalregion.Thesec-ondcomparativegroupcomprisedtenpatientswithSDwhowereidentiedthroughtheMemoryandCogni-tiveDisordersclinicatAddenbrooke’sHospital,Cam-bridge.ThesepatientsmetallpublishedcriteriaforSD[51];theyhadwordndingdifcultiesdespiteentspeechandshowedimpairedsemanticknowledgeandsinglewordcomprehensionbutphonology,syn-tax,visuo-spatialabilitiesandday-to-daymemoryre-mainedrelativelywellpreserved[50].Focalbilater-alatrophyoftheinfero-lateralaspectsoftheanteriortemporallobewasidentiedineverycaseusingMRI.2.2.Assessments2.2.1.Attentional/executiveAllfourgroupswereassessedontheRaven’scolouredprogressivematricestestofnon-verbalrea-soning[52].Inaddition,theADandSAgroups’atten-tional/executiveskillwasassessedwiththeWisconsinCardSortingTest[53]andtheElevatorCountingtaskwithandwithoutdistraction(takenfromtheTestofEverydayAttention[54]).2.2.2.SemanticmemoryassessmentArangeoftestswereusedtoassesssemanticcog-nition.ThePyramidsandPalmTreestestexaminessemanticassociativeknowledge(PPT[55]).Thistaskrequiresparticipantstodecidewhichoftworesponseoptions(e.g.,palmtree/pinetree)ismostcloselyre-latedtotheprobeitem(e.g.,pyramid).Thetaskwasassessedinbothpictureandwordformats.Fluencywasexaminedforsixcategories(animals,birds,fruit,householditems,toolsandvehicles).Patientsweregivenoneminutetoproduceasmanymembersofeachcategoryaspossible.Categoryuencywascomparedtoperformanceonaletteruencytask(participantshadtoproduceasmanywordsbeginningwiththelettersF/A/S).ThreetasksfromtheCambridgesemanticbatterywereusedtoassessknowledgeofsixty-fourconcreteconceptsdrawnfromthesamesixcategorieslistedabove.Thetaskswere:i)TheCamelandCactustest(CCT[50]):thisisasemanticassociationtasksimilartothePPT.Participantsarerequiredtoselectwhichoffourresponseoptionsismostrelatedtoaprobeitem.Thetestwasadministeredinbothwordandpic-tureformats.ii)Spokenword-to-picturematching:participantsaregiventhenameofanitemandaskedtoselectthecorrespondingpicture.Tenpicturestimuli(thetargetandninefoils)drawnfromthesamesemanticcategorywerepresentedoneachtrial.iii)Picturenaming:patientswerepresentedwithalinedrawingofeachiteminthebattery(andaskedtonameit.Theenvironmentalsoundbattery[50]employschar-acteristicsoundstoassessknowledgerelatingto48itemsdrawnfromsixcategories(domestic/foreignan-imals,humansounds,householditems,vehiclesandmusicalinstruments).Thebatteryrequiresthepar-ticipantstomatchaprobetoatargetwithinanarrayofeightpossiblewithin-categoryalternatives.Threeidentity-matchingconditionswereassessed:sound-picture,sound-wordandspokenword-picturematch-ing.TwotestswereusedtoexaminetheeffectofcueingonpicturenaminginAD.(a)Cuedandmiscuednaming:Patientswereaskedtonamethe60picturesfromtheBostonNamingTest[56].Eachitemwasassessedundercued,miscuedanduncuedconditionsonseparateoccasions.Acueconsistedofthersttwophonemesofthetargetre-sponsewhereasthephonemesofasemanticcompetitorwereprovidedonthemiscuedtrials.(b)Progressivephonologicalcueing:participantswereaskedtonamethirtyitemsselectedfromtheSnodgrassandVander-wart[57]stimulusset.Ifanitemwasnotnamedspon-taneously,thentherstphonemeofthetargetresponsewasprovidedasacue.Cuesincreasedinlength,onephonemeatatime,untileithertheitemhadbeennamedcorrectly,orthenalphonemehadbeenprovided. F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:Qualitativechangeswithdiseaseseverity3.Results3.1.Backgroundneuropsychologicalexamination3.1.1.Attentional/executiveJefferiesandLambonRalph[1]foundthatSApa-tientsperformedpoorlyacrossarangeofattention-al/executivetasks(seeTable1formeanscores).Con-versely,patientswithSDrarelyshowadecitonthistypeofassessment.AllSDcasesre-reportedinthepresentstudy,forexample,scoredwithinthetop25thpercentileoftheRaven’scolouredprogressivematricesWeexaminedtheADpatientsonanumberofatten-tional/executivemeasurespreviouslyusedtoexaminetheSApatients.InlinewithpreviousassessmentsofthisdomaininearlyAD[42,44],themildADgroupexhibitedsomeevidenceofimpairmentonthesetaskswhiletheseverelyimpairedADpatientsattainedscoresbelowthenormallevelonallattentional/executivemea-sures(seeTable1formeanscoresandTable1aforin-dividualADpatientdata).ThetwogroupsperformedequivalentlyontheWCST;6/10mildlyimpairedcasesand7/10severelyimpairedpatientsdidnotcompleteasinglecategory(1.83,n.s).Theelevatorcountingtask(withoutdistraction)waspassedby9/10casesinthemildgroupcomparedtoonly5/10patientsintheseveregroupbutthegroups’overallscoresdidnotdiffersignicantly(2.06,n.s.).Whenel-evatorcountingwasexaminedwithdistractionasig-cantgroupdifferencewasobservedwithonly3/10mildcasesscoringbelowthenormalcut-offcomparedto7/10patientsintheseveregroup(0.026).SignicantlybetterperformanceinthemildgroupwasalsoobservedfortheRaven’sprogressivematricestask(3.1.2.SemanticAllgroupsexhibitedimpairmentacrossafullrangeofpicture/wordandsound-basedtaskssuggestingamultimodalsemanticimpairmentineverycase(seeTa-bles1and1a).TheseverityofsemanticimpairmentwaslargelyequivalentfortheSAandSDgroupswhosescoresonlydifferedsignicantlyontwotasks:letteruency(SD0.001)andsound-picturematching(�SASD:0.04).TheseverelyimpairedADgroupperformedsubstantiallyworsethanthemildgrouponalltasksinthesemanticbattery(0.018).RelativetotheSApatients,themildADgroupobtainedsubstantial-lyhigherscoresonfoursemantictasks(naming,wordCCT,categoryuencyandletteruency:0.03)butsignicantlypoorerscoresontwooftheenvironmentalsoundstasks(sound-picture/wordmatching:0.01).TheSDgroupweresubstantiallyworsethanthemildADgroupinpicturenaming(0.01).Incontrast,these-vereADgroupobtainedsignicantlylowerscoresthantheSAgrouponsevensemanticassessments(naming,picture/wordPPT,pictureCCTandallthreesubtasksfromtheenvironmentalsoundsbattery:0.05)andwereworsethantheSDpatientsforvetasks(pictureCCT,categoryuency,letteruency,sound-picturematchingandsound-wordmatching:3.2.ExaminingthenatureofsemanticimpairmentIntheircase-seriescomparativestudy,JefferiesandLambonRalph[1]identiedanumberofwaysinwhichsemanticimpairmentsduetoafailureofregulatorycontrol(SA)couldbedistinguishedfromthoseduetodegradedsemanticrepresentations(SD).Thesameanalyseswereemployedinthepresentstudytoexam-inethenatureofsemanticimpairmentatmildandse-verestagesofAD.WehypothesisedthatafailureofregulatorycontrolprocessesintheearlystageofADwouldleadtoderegulatedsemanticcognition.Specif-ically,weexpectedmildcasestoexhibitthefollowingthreecharacteristicsofsemanticderegulationprevious-lyobservedinSA:1)anassociationbetweenperfor-manceonsemantictasksandattentional/executiveas-sessments;2)astrongrelationshipbetweentheabilitytomakesemanticassociativejudgementsintheCCTandthedegreeofconceptualcontrolrequiredbyapar-ticulartrial;and3)namingaccuracycanbebothim-proved(bytheprovisionofcorrectcues)anddimin-ished(bymiscues).DuetotheprogressivenatureofAD,however,itwasexpectedthataqualitativelydif-ferentsemanticimpairmentwouldemergelaterinthedisorderassemanticderegulationwasjoinedbygradu-aldegradationofsemanticrepresentations.Theimpactofthisdualimpairmentwouldbeexpectedtobeespe-ciallypronouncedontasksthatplacealargedemandonbothconceptualknowledgeandsemanticcontrol,suchascategoryuency.Furthermore,relativetomildADpatients,severecasesshouldbegintoshowcharacter-isticsofastorageproblemasexhibitedbySDpatients,including:1)moreconsistentimpairmentofconceptu-alknowledgeacrossdifferentsemantictasks;and2)astrongerinuenceofconceptualfamiliarity/frequencyonperformance.Giventhatwehaveformallydemon- F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:QualitativechangeswithdiseaseseverityTable1a,IndividualADpatientdata TaskMaxControlMildADSevereAD mean(s.d.)EMLLMHDSMOJMDMJHTAEAJLERBLADLAMLEBJM Executive/attentionalWCST(categories)61*66300000060100000021Ravens’Matrices9525905082.5802592.525251055551055510(percentiles)Elevatorcounting76*76746767676203177747Elevatorcounting103*103601449082111005300(withdistraction)PPTpicture5251.2(1.4)4443523247374046424029422526323245432536PPTword5251.1(1.1)454749344640464645493942242630394146273364itembatteryNaming6462.3(1.6)4942614449455749573952483939364237294132WPM6463.7(0.5)6350644057586156615648594827485242513448CCTpicture6458.9(3.1)4639562341302646433237302320192826341728CCTword6460.7(2.06)5253563153444540434642471618254739472128EnvironmentalsoundsSpokenword–picture4847.8(0.6)48394730434040404432264026212340323032NTSound–picture4841.2(2.5)221926101114252021198227561711NT10NTSound–word4840.8(3.8)251325111412211421151013115111515NTNTNTCategoryuency(6)95.7(16.5)5441481831353131503321800928191250Letteruency44.2(11.2)481719191020916603910560015241290 *denotesnormalcut-off.NTnottested.TATestabandoned.stratedthesebehaviouralpatternspreviouslyforSDandSApatients,theresultsbelowconcentratespeciuponthenewADdata.3.3.Indicatorsofderegulation3.3.1.CorrelationsbetweensemanticandexecutivetasksFortheADgroups,weassessedtherelationshipbe-tweenperformanceonfourexecutivetasks(elevatorcountingwith/withoutdistraction,WCSTandRaven’smatrices)andalltentasksfromthesemanticbattery(excludingletteruency).ThemildADgroupexhibit-edsignicantassociationsbetweentheelevatorcount-ingtaskandsixsemantictasks,including:WPM,wordandpictureversionsofthePPTandCCTaswellascat-egoryuency(0.046).Twofurthercor-relationsapproachedsignicance(sound-pictureandspokenword-picturematchingfromtheenvironmentalsoundsbattery:0.059).Signicantas-sociationswerenotobservedfortheremainingexec-utivetasks,whichisprobablyattributabletoooref-fectsontheseassessments.Notably,however,associ-ationsthatapproachedsignicancewereobservedforthefollowingtaskcombinations:1)elevatorcountingwithdistractionandwordPPT/sound-picturematching0.087);WCSTandwordPPT/wordCCT/categoryuency(0.088)and,3)Raven’smatricesandpicturePPT/pictureCCT/spokenword-picturematching(Themoreseverelyimpairedpatientsalsoexhibitedadegreeofassociationbetweentheirperformanceonsemanticandexecutivetasks.Signicantcorrelationsoccurredfortheelevatorcountingtask(withoutdistrac-tion)andthefollowingsemanticassessments:wordPPT,bothversionsoftheCCTandcategoryuency0.052).Associationsapproachingsig-cancewereobservedforelevatorcountingwithdis-traction/wordPPTandRavensmatrices/pictureCCT3.3.2.FactorsaffectingperformanceonCCTJefferiesandLambonRalph[1]foundthatmeasuresofthesemanticcontrolrequiredtocompleteataskwereabletopredictitem-by-itemperformancedifferentiallyinSAandSDpatients.Theeasewithwhichdistractorresponsescouldberejected(asratedbycontrolpar-ticipants)intheCCTwasparticularlysensitivetoSApatients’controldecit.ThepatternsexhibitedbythemildandsevereADgroupsreectedthoseobservedpreviouslyinSAandSDrespectively.PerformanceontheCCT(meanword/picturescore)correlatedwitheaseofrejectingdistractorsforthemildADgroup(0.0001;two-tailed)butnottheseverepa-tients(0.20,n.s.).Usingsimultaneouslogisticre-gression,acombinedanalysisrevealedasignicantin-teractionbetweengroupandsemanticcontroldemands F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:QualitativechangeswithdiseaseseverityTable2Breakdownofpicturenamingerrors ErrortypeSAMildADSevereADSD Semantic0.330.720.600.45Phonological0.1100.060.02Perseveration0.150.080.070.11Omission0.320.140.20.37Unrelated0.100.060.070.05 AllperseverativeerrorsmadebytheADgroupsandSDpatientswerealsosemanticallyrelatedtothetarget.(Wald0.011),whichwasindicativeofagreaterinuenceofthisfactoronthemildADgroup’sperformance(ExpB:mild1.51;severe3.3.3.Namingerrors/CueingOnthe64setpicturenamingtaskallgroupsmademostlysemanticandomissionerrors(seeTable2).TheSAandSDgroupsexhibitednodifferencesinthefre-quencyofsemantic,unrelated,perseverativeoromis-sionerrors.ThesamepatternwasobservedwhenthemildandsevereADgroupswerecompared;theonlysignicantdifferenceoccurredforphonologicalerrors,whichconstitutedagreaterproportionoferrorsforthesevereADgroup(SApatients’impoverishedpicturenamingcanbeimprovedsignicantlywiththeprovisionofexternalsupportintheformofphonemiccues.InlinewiththehighlyconsistentnatureofsemanticimpairmentinSD,however,verylittledifferenceisobservedbetweenspontaneousandcuednamingscores[58].This,cou-pledwiththehighomissionrateinSD,isindicativeofdegradationofthecorerepresentationsinthesemanticrepository.3.3.4.CuedandmiscuednamingInarepeatedmeasuresANOVAcomparingthemildandseverelyimpairedADgroups,thereweresignif-icantmaineffectsofcuetype(cue/miscue/nocue;F0.0001)andgroup(mild/severe;F0.02)aswellasasignicantinter-actionbetweenthetwo(F(2,36)seeFig.1).Themildgroupperformedsignibetterthantheseverelyimpairedgroupinthecuedanduncuedconditions(cued:0.02)butnodifferencewasobservedforthemiscuedcondition(n.s.).Forthemildpatientssignicantdifferenceswereobservedbetweenallthreeconditions(cued�miscued:0.015)whereastheseveregroupdidnotshowasignicantdifferencebetweentheuncuedandmiscuedconditions(cued0.016;uncuedmiscued: Fig.1.PerformanceontheBostonNamingTestunderdifferentcueconditions.*indicatessignicantdifference.Errorbarsdenotestandarderrorofthemean(SE).3.3.5.ProgressivephonologicalcueingInarepeatedmeasuresANOVAthereweresignicantmaineffectsofcuesize(nocue/twophonemes/morethantwophonemesprovided:F(2,36)0.0001)andgroup(mild/severe:F(1,18)0.019)butnointeractionbetweenthetwo(F(2,36)1.93,n.s.).Weexaminedtheproportionofphonemesrequiredtocueanitemthatwasnotnamedsponta-neously.Onaveragethemildgrouprequired37%thetarget’sphonemesforsuccessfulcueing,whichwassubstantiallylessthantheseveregroup(56%;0.004;patientBLwasexcludedfromthisanalysisasnoitemswerecuedsuccessfully).3.4.IncreasingindicationofsemanticdegradationinsevereAD3.4.1.NamingandcategoryTherelativecontributionofsemanticcontrolvs.rep-resentationvariesacrossdifferentsemantictasksandthustherelativeperformanceacrossthesetaskscanbeusedtodiagnosewhichaspect(s)ofsemanticcogni-tionareaffectedinaparticulardisorder.Weexaminedpatients’performanceacrosspicturenamingandcat-egoryuency.Thesetasksdrawonknowledgefromthesamesemanticcategoriesbutdifferinimportantways.Picturenamingisfundamentallydependentontheintegrityofaconcept’srepresentation.Adegreeofcontrolis,arguably,alsorequiredtoinhibitcompetingnames.Likepicturenaming,categoryuencyrequiresintactsemanticknowledgebutalsoplacessubstantialdemandoncontrolprocesses.Participantsmustper-formanonlinesearchofthesemanticrepository,se- F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:Qualitativechangeswithdiseaseseverity Fig.2.Namingandcategoryuencyperformance.Verbaluencytasks,unlikeconfrontationalnaming,donothaveamaximumscore.Thustocompareperformancedirectlyweexpressedthepatients’performanceasaproportionofthecontrolmeanoneachtask.ErrorbarsdenoteSE.lectingitemsthatpossesskeycharacteristicsofagiv-encategoryandrejectingclosecompetitors,aswellasmonitoringpreviousresponses.Thedemandoncontrolprocessesbecomesgreaterasrepresentationsdegradeandnormalcategoryboundariesbecomedistorted.Allgroupsperformedwithintheimpairedrangeforbothtasksbutthepatternacrosstasksdifferedacrossthegroups(Table1).Whileallfourgroupsperformedworseonthecategoryuencytaskcomparedtonaming,thisdeclinewasexaggeratedforthesevereADgroup(rawscoresdividedbycontrolmean;seeFig.2).InrepeatedmeasuresANOVAs,signicanttaskbygroupinteractionswerefoundwhentheseverelyimpairedADpatientswerecomparedwitheachoftheotherpa-tientgroups(mildAD:F(1,18)SA:F(1,17)0.008;SD:F(1,18)0.022).BorderlineinteractionswerealsoobservedwhenthemildADgroupwerecomparedwiththeSApatients(F(1,17)0.094)andtheSDgroup(F(1,18)0.098).Performanceoncate-uencywassignicantlyworsethannamingintheseverelyimpairedADpatients(0.0001)andthemildADpatients(0.0001),butnottheremainingtwogroups(SA:2.23,n.s.;SD:0.64,n.s.).Theseresultsareconsistentwiththeviewthatthereisacombinationofregulatorycontrolfailureandde-gradingsemanticrepresentationsinthelaterstagesofAD.Themilderpatientsshowedasimilarbutlessacutedeclineinperformanceacrossthetwotasks,reectingagradualtransformationfromapurelysemanticcontroldisorderintoacombinedstorageandcontroldisorderacrossthecourseofthedisease.3.4.2.ConsistencySimultaneouslogisticregressionwasusedtodeter-minewhetheritemspassedorfailedinonetaskwouldpredictperformanceonthesameitemsinadifferenttask.Familiaritywasalsoincludedasapredictorbe-causeitcontributestosomeoftheconsistencyexhib-itedbytheSDpatients.Inadditiontoseparateanaly-sesofthedifferentpatientgroups,combinedanalyseswereusedtocompareconsistencyacrossthedifferentgroups.Patientswhoscoreatoor/ceilingonataskwereexcludedfromtheanalysis(i.e.,10%orcorrect).Previousstudiesfoundsignicantitemcon-sistencyforSDandSApatientsforwithintaskcom-parisonsbutonlyconsistencyfortheSDpatientswhencomparingacrossdifferenttasktypes.3.4.2.1.ConsistencyacrossdifferentmodalitieswithinthesamesemantictaskThemildandsevereADgroupsalsoperformedcon-sistentlyacrosswordandpictureversionsoftheCCT(Mild:10,Wald0.0001;Severe:9,Wald0.0001).Familiaritywasasignicant,ornearlysignicant,predictorofperfor-manceonbothCCTp/CCTwtasksforthemildandseverelyimpairedgroups(Mild:Wald4.18and9.97,0.04and0.002;Severe:Wald14.87and2.8,0.0001and0.094,CCTpandCCTwrespec-tively).Interactiveanalysesdidnotreachsigni(Wald F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:Qualitativechangeswithdiseaseseverity Fig.3.Familiarityeffectsinpicturenaming.ErrorbarsdenoteSE.Consistencywasalsoexaminedacrossthreetasksfromtheenvironmentalsoundsbattery:sound-picture,sound-wordandspokenword-picturematching.ThemildADcasesexhibitedconsistencyfor9/12analyses(sound-picture/sound-wordn=10,spokenword-picture7;Wald0.0001)withexceptionsarisingforcombinationsofsound-wordandspokenword-picturematching(Wald1).Theseveregroupwereconsistentacrossalltwelvecomparisons(sound-picture7,sound-word6,spokenword-picture9;Wald0.0001).Interactionsapproachedsig-canceforthreetaskpairs,whichwereindicativeofagreaterdegreeofconsistencyintheseverelyimpairedgroup(spokenword-picturevs.sound-picture;sound-picturevs.spokenword-picture;sound-wordvs.spo-kenword-picture:WaldmeanExpB:Mild1.71;Severe3.4.2.2.ConsistencyacrossdifferentsemantictasksThemildADcasesonlyexhibitedconsistencyforonetaskpair(naming/WPM,5,Wald0.007).Incontrasttheseverelyimpairedgroupperformedhighlyconsistentlyacross6/10taskcom-binations(naming/WPM:9,Wald0.01;naming/CCTp:10,WaldWPM/CCTw,8,Wald3.4.3.FamiliarityNamingperformanceonthe20mostfamiliaritemsinthe64setbatterywascomparedwiththatofthe20leastfamiliaritems(20ineachset).TheSDpa-tientsshowedsubstantiallylargereffectsoffamiliari-tythantheSAgroup.WhenthemildandsevereADgroupswerecompared(seeFig.3),asignicantgroupbyfamiliarityinteractionwasfound(F(1,16)0.054;seeFig.3b).Themildgroupweremoreimpairedwhennaminglowerfamiliarityitems(0.023)andahighlysignicanteffectoffamiliaritywasobservedfortheseverelyimpairedADgroup(0.001).Whilethetwogroupsperformedcomparablyonthehighfrequencyitems,themoreseverelyimpairedgroupperformedsigniworseonthelowfrequencyitems(4.DiscussionWeextendedJefferiesandLambonRalph’s[1]case-seriescomparisonofSAandSDtoexaminethenatureofsemanticimpairmentinAD.BothmildandsevereADpatientsperformedbelowthenormallevelacrossafullrangeofsemanticassessments.Whenpatientsatmildandmoreseverestagesofthediseasewerecon-trasteddirectly,however,qualitativedifferencesinthenatureoftheirsemanticimpairmentsemerged.ThesendingssuggestthattheevolutionofAlzheimerpathol-ogyoverthecourseofthediseaseaffectsseparatecom-ponentsofsemanticcognitionatdifferenttimes.Whileafailureofcontrolprocessesleadstosemanticderegu-lationinitially,bythelaterstageofthediseaseseman-ticrepresentationsthemselveshavealsobeguntode-grade.Ourresultscouldexplainwhyithasnotprevi-ouslybeenpossibletodeterminethenatureofsemanticimpairmentinADwithoutexplicitlyconsideringtheimpactofdiseaseseverity.Intheiroriginalcase-seriescomparativestudy,Jef-feriesandLambonRalph[1]identiedanimportantlinkbetweenthesemanticimpairmentexhibitedbypa-tientswithSAandtheirimpairedexecutivecontrolpro-cesses.Inthepresentstudy,thesamepatternofdereg-ulatedsemanticcognitionwasalsoexhibitedbymildlyimpairedADpatients.LikeSApatients,themildAD F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:Qualitativechangeswithdiseaseseveritygroupshowedanassociationbetweenperformanceonattentional/executivetasksandassessmentsofconcep-tualknowledge.Theseverelyimpairedgroupalsoex-hibitedadegreeofcorrelationbetweensemanticandexecutivemeasuresbutfewerreachedsignicance.Wespecicallyexploredtherelationshipbetweenperfor-manceontheCCTsemanticassociationtaskandthecognitivedemandsofanindividualtrial.ItwasfoundthattheeasewithwhichdistractorresponsescouldberejectedonlycorrelatedwithperformancefortheSAandmildADpatientssuggestingthatexecutiveburdenwasparticularlyimportantforthesegroups.Moreover,weinvestigatedtheeffectofmanipulatingthedegreeofexternalconstraintprovidedinapicturenamingtaskandfoundstrikinggroupdifferences.AspreviouslyobservedinthecontextofSA[1,58],themildADgroupshowedsubstantialimprovementwhenprovid-edwiththerstphonemeofthetargetwhileperfor-mancediminishedwhenpatientsweremiscuedwiththephonemeofasemanticcompetitor.Conversely,thesevereADgroupshowedonlyalimitedimprovementwithhighlyconstrainingcuesandnoeffectofmis-cues.TheparallelnatureofsemanticdecitsobservedinmildADandSAsuggeststhat,likeSApatients,themildADcasesperformedpoorlyonsemantictasksduetoafailuretocontrolorshapetheirknowledgeinataskappropriatefashion.Arelativelackofcorrespon-dencebetweenSAandsevereAD,however,suggeststhataqualitativelydifferentimpairmentunderliespoorperformanceonsemantictasksatalaterstageofAD.WeexaminedthehypothesisthattheexecutivedeunderlyingsemanticimpairmentintheearlystagesofADnotonlyworsensovertimebutiseventuallyjoinedbyanadditional,gradualdegradationofthesemanticrepositoryitself.ThisviewanticipatesthatlaterstageADpatientsshouldbelesscomparablewithSAcasesandbegintoexhibitcharacteristicsofdegradedknowl-edgealreadydescribedinthecontextofSD.Toinves-tigatethenatureofsemanticimpairmentinsevereADweexaminedpatients’performanceacrosstasksthatmadedifferentdemandsonindividualcomponentsofsemanticcognition:picturenamingandcategorycy.Bothtasksdependonalargerepositoryofconcep-tualknowledge.Inaddition,however,categoryuencyalsoplacesamuchgreaterdemandoncontrolprocess-esthandoespicturenaming.Inordertogeneratecate-gorymembers,aparticipantmustbeabletosearchthesemanticrepository,selectappropriateexemplarsandrejectinappropriateitemsaswellasmonitorpreviousresponses.Whileallfourpatientgroupsobtainedlow-erscoresincategoryuencythanpicturenaming,thisdifferencewasparticularlyexaggeratedforthesevere-lyimpairedADgroup.Thegrowinginuenceofse-manticdegradationoverthecourseofADwasalsore-ectedbytheincreasinglyconsistentnatureofpatients’impairments.Whilethemildpatientsperformedmostconsistentlyacrossthesametaskpresentedindiffer-entmodalities(i.e.,whenthetaskdemandsremainedconstant),theseverelyimpairedgroupalsoexhibitedadegreeofconsistencyacrosstaskstappingthesameconceptsbutindistinctways(e.g.,naming/CCT).ThedifferentpatternsofconsistencyobservedinmildandsevereADechoedthosealreadyobservedinSAandSDrespectively.Moreover,agrowingtendencytowarditemconsistencyinlaterADcoincideswiththeviewthatrepresentationsaregraduallybutpermanentlyde-gradedinthiscondition.Inlinewiththishypothesis,weobservedaninuenceofconceptfamiliaritynoton-lyinSDbutalsointhesevereADpatients.ConverselytheSAandmildADpatientsshowednodifferenceinperformanceforlowandhighfrequencyconcepts.BydirectlycomparingpatientswithmildandsevereAD,wehavedemonstratedthatthenatureofseman-ticimpairmentisnotstablebutsubjecttoqualitativechangesoverthecourseofthedisease.Disruptionofthefrontal-temporoparietal-cingulatecontrolnetworkleadstoderegulationofsemanticcognitionintheearlystagesofthedisease.Thisproblemnotonlypersistsandworsensovertimebutalsobecomescompoundedbydegradationofamodalsemanticrepresentationssup-portedbytheATL.OurresultsrepresentakeyadvanceinunderstandingthenatureofsemanticimpairmentinAD.WeassertthatitisnolongerusefultocharacteriseADaseitherastoragedecitoranaccess/controldis-order.Instead,researchinthisareashouldbedirect-edtowardtracingchangesinthenatureofsemanticimpairmentwithdiseaseprogression.AcknowledgementsWewouldliketothankallofthepatientsandtheircarersfortheirsupportofourstudies.WearealsogratefultoDr.SarahCraigforidentifyingsuitablecasesforthestudy.ThisworkwassupportedbyanMRC-ESRCstudentship(toF.Corbett),andanMRCgrant(G0501632).WearegratefultoProfs.JohnHodgesandKaralynPattersonforcontinuingtosharetheoriginalSDdatawithus. F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:QualitativechangeswithdiseaseseverityReferences[1]E.JefferiesandM.A.LambonRalph,Semanticimpairmentinstrokeaphasiaversussemanticdementia:acase-seriescom-parison,Brain(2006),2132–2147.[2]F.Corbettetal.,Qualitativelydifferentconceptualimpair-mentsinsemanticdementiaandsemanticaphasia:Converg-ingevidencefromthenon-verbaldomain,Brain[3]K.A.Noonanetal.,Elucidatingthenatureofderegulatedse-manticcognitioninsemanticaphasia:Evidencefortherolesofprefrontalandtemporoparietalcortices,JournalofCogni-tiveNeuroscience(2010),1597–1613.[4]M.A.LambonRalph,C.LoweandT.T.Rogers,Neuralbasisofcategory-specicsemanticdecitsforlivingthings:evidencefromsemanticdementia,HSVEandaneuralnetworkmodel,Brain(2007),1127–1137.[5]S.Ahmedetal.,Namingofobjects,facesandbuildingsinmildcognitiveimpairment,Cortex(2008),746–752.[6]J.R.Hodges,S.ErzincliogluandK.Patterson,Evolutionofcognitivedecitsandconversiontodementiainpatientswithmildcognitiveimpairment:Avery-long-termfollow-upstudy,DementiaandGeriatricCognitiveDisorders(2006),380–[7]J.R.HodgesandK.Patterson,Issemanticmemoryconsistent-lyimpairedearlyinthecourseofAlzheimer’sdisease?Neu-roanatomicalanddiagnosticimplications,Neuropsychologia(1995),441–459.[8]J.R.Hodges,D.P.SalmonandN.Butters,Semanticmemo-ryimpairmentinAlzheimer’sdisease:Failureofaccessordegradedknowledge?Neuropsychologia(1992),301–314.[9]P.Garrardetal.,Semanticfeatureknowledgeandpicturenam-ingindementiaofAlzheimer’stype:Anewapproach,BrainandLanguage(2005),79–94.[10]J.R.Hodgesetal.,NamingandknowingindementiaofAlzheimer’stype,BrainandLanguage(1996),302–325.[11]J.R.Hodges,D.P.SalmonandN.Butters,RecognitionandnamingoffamousfacesinAlzheimer’sdisease:Acognitiveanalysis,Neuropsychologia(1993),775–788.[12]H.ChertkowandD.Bub,SemanticmemorylossindementiaofAlzheimerstype-Whatdovariousmeasuresmeasure?Brain(1990),397–417.[13]H.Chertkow,D.BubandD.Caplan,ConstrainingTheoriesofSemanticMemoryProcessing-EvidencefromDementia,CognitiveNeuropsychology(1992),327–365.[14]R.NebesandE.M.Halligan,SentencecontextinuencestheinterpretationofwordmeaningbyAlzheimerpatients,BrainandLanguage(1996),233–245.[15]R.NebesandE.M.Halligan,ContextualconstraintfacilitatessemanticdecisionsaboutobjectpicturesbyAlzheimerpa-tients,PsychologyandAging(1995),590–596.[16]R.Nebes,SemanticmemoryinAlzheimer’sdisease,Psycho-logicalbulletin(1989),377–394.[17]K.Bayles,Alzheimer’sdiseaseeffectsonsemanticmemory:Lossofstructureorimpairedprocessing?Journalofcognitiveneuroscience(1991),166–182.[18]K.Bayles,PerformanceofAlzheimer’sdiseasepatientsinjudgingwordrelatedness,JournaloftheInternationalNeu-ropsychologicalSociety(1999),668–675.[19]S.Luzzietal.,DistinctpatternsofolfactoryimpairmentinAlzheimer’sdisease,semanticdementia,frontotemporalde-mentia,andcorticobasaldegeneration,Neuropsychologia(2007),1823–1831.[20]M.Cocciaetal.,Semanticmemoryisanamodal,dynamicsys-tem:Evidencefromtheinteractionofnamingandobjectuseinsemanticdementia,CognitiveNeuropsychology[21]S.Bozeatetal.,Whenobjectslosetheirmeaning:Whathap-penstotheiruse?Cognitive,Affective&BehaviouralNeuro-science(2002),236–251.[22]M.A.LambonRalphetal.,Naminginsemanticdementia:whatmatters?Neuropsychologia(1998),775–784.[23]M.A.LambonRalphetal.,Isapictureworthathousandwords?Evidencefromconceptdenitionsbypatientswithsemanticdementia,BrainandLanguage(1999),309–335.[24]C.J.Mummeryetal.,Avoxel-basedmorphometrystudyofsemanticdementia:Relationshipbetweentemporallobeat-rophyandsemanticmemory,AnnalsofNeurology[25]J.S.Snowden,P.J.GouldingandD.Neary,Semanticdemen-tia:Aformofcircumscribedcerebralatrophy,BehaviouralNeurology(1989),167–182.[26]S.J.GottsandD.C.Plaut,Theimpactofsynapticdepres-sionfollowingbraindamage:Aconnectionistaccountof“ac-cess/refractory”and"degraded-store"semanticimpairments,Cognitive,Affective&BehaviouralNeuroscience[27]E.K.WarringtonandR.A.McCarthy,Categoryspecicaccessdysphasia,Brain(1983),859–878.[28]E.K.WarringtonandL.Cipolotti,Wordcomprehension:Thedistinctionbetweenrefractoryandstorageimpairments,Brain(1996),611–625.[29]M.A.LambonRalphetal.,Takingbothsides:dounilater-alanteriortemporallobelesionsdisruptsemanticmemory?Brain(2010),3243–3255.[30]M.A.LambonRalph,G.PobricandE.Jefferies,ConceptualKnowledgeIsUnderpinnedbytheTemporalPoleBilaterally:ConvergentEvidencefromrTMS,CerebCortex[31]M.A.LambonRalphetal.,Coherentconceptsarecomputedintheanteriortemporallobes,ProceedingsoftheNationalAcademyofSciences(2010),2717–2722.[32]G.Pobric,E.JefferiesandM.A.LambonRalph,Category-SpecicversusCategory-GeneralSemanticImpairmentIn-ducedbyTranscranialMagneticStimulation,CurrentBiology(2010),964–968.[33]C.Whitneyetal.,TheNeuralOrganizationofSemanticCon-trol:TMSEvidenceforaDistributedNetworkinLeftInfe-riorFrontalandPosteriorMiddleTemporalGyrus,CerebralCortexOnlineearly-view[34]J.R.Binderetal.,Whereisthesemanticsystem?Acriti-calreviewandmeta-analysisof120functionalneuroimagingstudies,CerebralCortex(2009),2767–2796.[35]B.Deweeretal.,MemorydisordersinprobableAlzheimer’sdisease:theroleofhippocampalatrophyasshownwithMRI,JournalofNeurology,NeurosurgeryandPsychiatry[36]N.C.Foxetal.,PresymptomatichippocampalatrophyinAlzheimer’sdisease:AlongitudinalMRIstudy,Brain(1996),2001–2007.[37]B.C.Dickersonetal.,IncreasedhippocampalactivationinmildcognitiveimpairmentcomparedtonormalagingandAD,Neurology(2005),404–411.[38]P.J.Nestoretal.,LimbichypometabolisminAlzheimer’sdis-easeandmildcognitiveimpairment,AnnalsofNeurology(2003),343–351. F.Corbettetal./UnpickingthesemanticimpairmentinAlzheimer’sdisease:Qualitativechangeswithdiseaseseverity[39]P.J.Nestor,T.D.FryerandJ.R.Hodges,DeclarativememoryimpairmentsinAlzheimer’sdiseaseandsemanticdementia,Neuroimage(2006),1010–1020.[40]S.Minoshimaetal.,Metabolicreductionintheposteriorcin-gulatecortexinveryearlyAlzheimer’sdisease,AnnalsofNeurology(1997),85–94.[41]A.D.LawrenceandB.J.Sahakian,AlzheimerDisease,atten-tion,andthecholinergicsystem,Alzheimerdiseaseandasso-ciateddisorders(1995),37–49.[42]R.J.PerryandJ.R.Hodges,AttentionandexecutivedeinAlzheimer’sdisease–Acriticalreview,Brain[43]M.C.Silverietal.,Attentionandmemoryinthepreclinicalstageofdementia,JournalofGeriatricPsychiatryandNeu-rology(2007),67–75.[44]R.J.Perry,P.WatsonandJ.R.Hodges,Thenatureandstag-ingofattentiondysfunctioninearly(minimalandmild)Alzheimer’sdisease:relationshiptoepisodicandsemanticmemoryimpairment,Neuropsychologia(2000),252–271.[45]S.A.Wylieetal.,Inefcientresponseinhibitioninindividualswithmildcognitiveimpairment,Neuropsychologia[46]H.BraakandE.Braak,StagingofAlzheimer’sdisease-relatedbrillarychanges,NeurobiologyofAging[47]C.J.Galtonetal.,DifferingpatternsoftemporalatrophyinAlzheimer’sdiseaseandsemanticdementia,Neurology(2001),216–225.[48]R.Daviesetal.,DevelopmentofanMRIratingscaleformultiplebrainregions:comparisonwithvolumetricsandwithvoxel-basedmorphometry,Neuroradiology(2009),491–[49]M.F.Folstein,S.E.FolsteinandP.R.McHugh,“Mini-mentalstate”:Apracticalmethodforgradingthementalstateofpatientsfortheclinician,JournalofPsychiatricResearch(1976),189–198.[50]S.Bozeatetal.,Non-verbalsemanticimpairmentinsemanticdementia,Neuropsychologia(2000),1207–1215.[51]J.R.Hodgesetal.,SemanticDementia:Progressiveaphasiawithtemporal-lobeatrophy,Brain(1992),1783–[52]J.C.Raven,ColouredProgressiveMatricesSsetsA,AB,B1962:London:H.K.Lewis.[53]D.T.Stussetal.,WisconsinCardSortingTestperformanceinpatientswithfocalfrontalandposteriorbraindamage:effectsoflesionlocationandteststructureonseparablecognitiveprocesses,Neuropsychologia(2000),388–402.[54]I.H.Robertsonetal.,TheTestofEverydayAttention,Flempton:ThamesValleyTestCompany.[55]D.HowardandK.Patterson,PyramidandPalmTrees:ATestofSemanticAccessfromPicturesandWords,1992,BuryStEdmunds:ThamesValleyTestCompany.[56]E.Kaplan,H.GoodglassandS.Weintraub,BostonNamingTest,1983,Philadelphia:Lea&Febiger.[57]J.G.SnodgrassandM.Vanderwart,StandardizedSetof260Pictures–Normsfornameagreement,imageagreement,familiarity,andvisualcomplexity,JournalofExperimentalPsychology-HumanLearningandMemory(1980),174–215.[58]E.Jefferies,K.PattersonandM.A.LambonRalph,Deofknowledgeversusexecutivecontrolinsemanticcognition:Insightsfromcuednaming,Neuropsychologia(2008),649–