14.82www

Neuropsychologiaxxx(2005)xxx–xxxDyslexicsareimpairedonimplicithigher-ordersequencelearning,butnotonimplicitspatialcontextlearningJamesH.
HowardJr.
a,b,,DarleneV.
Howardc,d,KarinC.
Japiksea,d,GuinevereF.
Edend,eaDepartmentofPsychology,TheCatholicUniversityofAmerica,Washington,DC20064,USAbDepartmentofNeurology,GeorgetownUniversity,Washington,DC,USAcDepartmentofPsychology,GeorgetownUniversity,Washington,DC,USAdInterdisciplinaryPrograminNeuroscience,GeorgetownUniversity,Washington,DC,USAeDepartmentofPediatrics,GeorgetownUniversity,Washington,DC,USAReceived1July2005;receivedinrevisedform25September2005;accepted16October2005AbstractDevelopmentaldyslexiaischaracterizedbypoorreadingabilityandimpairmentsonarangeoftasksincludingphonologicalprocessingandprocessingofsensoryinformation.
Somerecentstudieshavefounddecitsinimplicitsequencelearningusingtheserialreactiontimetask,butothershavenot.
Otherskills,suchasglobalvisuo-spatialprocessingmayevenbeenhancedindyslexics,althoughdecitshavealsobeennoted.
Thepresentstudycompareddyslexicandnon-dyslexiccollegestudentsontwoimplicitlearningtasks,analternatingserialresponsetimetaskinwhichsequentialdependenciesexistacrossnon-adjacentelementsandaspatialcontextlearningtaskinwhichtheglobalcongurationofadisplaycuesthelocationofasearchtarget.
Previousevidenceindicatesthattheseimplicitlearningtasksarebasedondifferentunderlyingbrainsystems,fronto-striatal-cerebellarcircuitsforsequencelearningandmedialtemporallobeforspatialcontextlearning.
Resultsrevealedadoubledissociation:dyslexicsshowedimpairedsequencelearning,butsuperiorspatialcontextlearning.
Consistentwiththisgroupdifference,therewasasignicantpositivecorrelationbetweenreadingability(singlerealandnon-wordreading)andsequencelearning,butasignicantnegativecorrelationbetweenthesemeasuresandspatialcontextlearning.
Testsofexplicitknowledgeconrmedthatlearningwasimplicitforbothgroupsonbothtasks.
Thesendingsindicatethatdyslexiccollegestudentsareimpairedonsomekindsofimplicitlearning,butnotonothers.
Thespecicnatureoftheirlearningdecitisconsistentwithreportsofphysiologicalandanatomicaldifferencesforindividualswithdyslexiainfrontalandcerebellarstructures.
2005ElsevierLtd.
Allrightsreserved.
Keywords:Implicitlearning;Developmentaldyslexia;Serialreactiontimetask;Contextualcueing;Learningdisorders;LearningdisabilityFluentreadingisachievedinaseriesofstagesorphasesoveraprotractedperiodinchildhoodviaregularinstructionandpractice(forreviews,seeEhri,1999)andisaccompa-niedbybrain-basedchanges(Simosetal.
,2001;Turkeltaub,Gareau,Flowers,Zefro,&Eden,2003).
However,evenwithadequateeducationalopportunity,somechildrendonotbecomeuentreaders;5–12%ofschool-agedchildrenareidentiedwithdevelopmentaldyslexia(Lyon,1995;Vellutino,Fletcher,Snowling,&Scanlon,2004).
Themostprominentweaknessesofdevelopmentaldyslexiaarefoundinwordidentication,phono-logical(letter-sound)decodingandspelling.
AlthoughadultsCorrespondingauthor.
Tel.
:+12023194791;fax:+12023196263.
E-mailaddress:howard@cua.
edu(J.
H.
HowardJr.
).
withdevelopmentaldyslexiamaycompensateinsomeareasofreading,thecardinalmarkersobservedinchildhood,suchaspoorphonologicalawarenessskills(Bradley&Bryant,1981),frequentlypersistintoadulthood(Ransby&Swanson,2003;Shaywitzetal.
,1999).
Behavioralstudiesconductedinchildrenandadultswithdyslexiahavefocusedonadiversesetoflanguageandnon-languageskills.
Inadditiontofaultyphonologicalprocessing,developmentaldyslexiahasbeendescribedasareadingdisor-derattributabletootherdecits,includingimpairedtemporalprocessing,magnocellularprocessingorrapidnaming,aswellasalackofautomatizationoracombinationoftheabove(forreviews,seeEden&Zefro,1998;Rayner,Foorman,Perfetti,Pesetsky,&Seidenberg,2001;Stein&Walsh,1997;Vellutinoetal.
,2004;Wolff&Lundberg,2002).
Theresultisanongoing0028-3932/$–seefrontmatter2005ElsevierLtd.
Allrightsreserved.
doi:10.
1016/j.
neuropsychologia.
2005.
10.
015NSY-2144;No.
ofPages142J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxxdiscussiononthecontributionoftheseobservedlanguageandsensorimotordecitsandtheirpotentialroleintheetiologyofdyslexia.
Anatomical(Eckert&Leonard,2000;Eckertetal.
,2003;Galaburda,Sherman,Rosen,Aboitiz,&Geschwind,1985)andfunctionalstudiesinindividualswithdyslexiahaverevealeddifferencesinregionsofoccipital–temporal,temporo-parietalandfrontalregionsofthelefthemispherewhencomparedtotypicalreaders.
Thesevariationsinbrainfunctionhavebeendemonstratedwhileparticipantsengageincognitivelinguistic(Brunswick,McCrory,Price,Frith,&Frith,1999;Edenetal.
,2004;Flowers,Wood,&Naylor,1991;Rumseyetal.
,1992;Shaywitzetal.
,1998)aswellassensorimotortasks(Demb,Boynton,&Heeger,1998;Eden&Zefro,1998).
Takentogether,bothbehavioralandbrain-basedresearchindicatesthatthemanifestationsobservedindyslexiaarecomplex,makingitdifculttoprovideaunitaryaccountoftheetiologyofthiscom-monandheritablelearningdisability(Eden&Zefro,1998).
Despitetheapparentdiscrepanciesintheeld,itiswidelyacceptedthatchildrenwithdyslexiahaveimpairedphonologi-calawareness.
Phonologicalawarenessistheabilitytoisolateandmanipulatetheconstituentsoundsoforallanguage,andprociencyinphonologicalawarenessiscrucialinlearningtomapalphabeticsymbolstosound,leadingtosuccessfulphono-logicaldecodingoftext(Vellutinoetal.
,2004).
Further,thereisstrongevidenceofbenecialeffectsofinterventionusingphonologicalawarenesstraining,suggestingadirectcausalrela-tionshipbetweenphonologicalawarenessskillsandlearningtoread(Alexander&Slinger-Constant,2004;Torgesenetal.
,2001).
Yet,littleisknownaboutwhydyslexicchildrenstruggletolearnthecodewhichlinksgraphemeswithphonemes,andwhethertheirinabilitytolearnthemappingofalphabeticsym-bolstosoundsisevidentinnon-linguisticdomainsoflearning.
Dyslexiaisrarelystudiedintheframeworkofthecontemporarylearningliterature.
Thepresentresearchdoessoviaafocusonimplicitlearning.
Unlikethedeliberateandconsciousprocessesthatoccurinexplicit(declarative)learning,implicitlearningoccursautomat-icallywithouttheintentiontolearnortheresultingexplicitknowledgeofwhatwaslearned(e.
g.
,Reber,1989).
Learningtoreadinvolvesbothexplicitandimplicitprocesses;childrenini-tiallylearngrapheme–phonememappingsexplicitlyafterwhichtheyapplyandcontinuetolearnthemimplicitly(Gombert,2003).
Theyalsolearntheorthography–meaningcorrespon-denceexplicitlythroughpicture–wordmatchingandimplicitlythroughcontext.
Onecouldcontemplateseveralmechanismsbywhichadecitinimplicitlearningcontributestodifcultiesassoci-atedwithdyslexia,butthesmallliteratureonimplicitlearninganddyslexiahasyieldedmixedresults.
Fivestudieshaveusedtheserialreactiontimetask(SRTT)introducedbyNissenandBullemer(1987)inwhichpeoplerespondtoeachofaseriesofstimulibypressingacorrespondingkey.
Sequencelearningisrevealedbyadeclineinperformancewhenthepredictablerepeatingpatternisreplacedbyarandomsequence.
Threeofthesestudiesreportedanimplicitlearningdecitinpoorreaders(Stoodley,Harrison,&Stein,inpress;Vicarietal.
,2005;Vicari,Marotta,Menghini,Molinari,&Petrosini,2003),whereastheothertwodidnot(Kelly,Grifths,&Frith,2002;Waberetal.
,2003).
Twoadditionalstudieshaveusedotherimplicitlearningtasks,againwithmixedresults.
PothosandKirk(2004)foundnoreading-relateddecitsinoneversionofanarticial-grammarlearningtaskandasignicantadvantagefordyslexicpeopleontheotherversion.
Yetanotherstudyfoundarelationshipbetweenimplicitcategoricallearningandreadingabilitysuchthatpoorreaderswereimpairedinimplicitlearning,butnotexplicitlearn-ing(Sperling,Lu,&Manis,2004).
Andnally,incontrasttomostearlierstudiesthatfocusedononlyonekindofimplicitlearning,Vicarietal.
(2005)examinedtwoimplicitlearningtasksthatengagedifferentcognitiveskills,serialreactiontimeandmirrordrawing.
Theyfoundthatdyslexicchildrendidmorepoorlythancontrolsonbothtasks,leadingthemtoconcludethatdyslexiaischaracterizedbyageneraldecitinimplicitlearning.
Thesendingssuggestthatitisnotenoughtocompareimplicitversusexplicitlearningortoinvestigateasingleimplicitlearningtask.
Thus,inthepresentstudyweusedtwoimplicitlearningtasksthatweexpecttobedifferentiallyaffectedbydyslexia.
TherstisanalternatingSRTTinwhichsequen-tialdependencesexistacrossnon-adjacentelements(Howard&Howard,1997;Howard,Howard,Japikseetal.
,2004).
Thesecondisaspatialcontextlearningtaskinwhichtheglobalcongurationofadisplaycuesthelocationofasearchtarget(Chun&Jiang,1998).
Thesetwoimplicitlearningtasksappeartorelyondifferentcognitiveskillsanddifferentbrainregions(Howard,Howard,Dennis,Yankovich,&Vaidya,2004).
Learn-ingofnon-adjacent,higher-order,sequentialregularitiescallsonfronto-striatal-cerebellarcircuitrywhereasspatialcontex-tuallearningdependsonmedialtemporallobestructures(Chun&Phelps,1999;Prull,Gabrieli,&Bunge,2000).
Cerebellar(Nicolson,Fawcett,&Dean,2001b)aswellasstriatal(Vicarietal.
,2005)decitshavebeenassociatedwithdyslexia,butthereisnoevidencetosuggestmedialtemporallobedysfunc-tionindevelopmentaldyslexia.
Wethereforepredictedthatpoorreaderswouldbeimpairedonimplicitsequencelearning,butnotimplicitspatialcontextlearning.
Furthermore,adissocia-tionbetweenthetwotypesofimplicitlearningtaskswouldhelptoestablishthatthedecitsshownononeimplicitlearningtaskareunlikelyduetogeneralattentiondecitswhichwouldpre-sumablyinuencebothtasks.
Bothoftheimplicitlearningtasksusedinthepresentstudyarestructuredsothatpredictableandunpredictabletrialsoccurineveryblock,makingitpossibletomeasurepatternlearningcontinuouslythroughouttraining.
Thisapproachisanimprove-mentoverthestudiesdescribedaboveinwhichlearningisnotmeasureduntilasinglerandomblockoccursneartheendoftraining.
Hence,thepresentdesignshouldbemoresensitivetoanygroupdifferencesintherateofimplicitlearning.
Inaddi-tion,bothtaskshavebeenshowntoresultinrelativelypureimplicitlearning:subjectsareunabletoconsciouslyrecognizeorproducetheregularitiestheyhavelearnedatabovechancelevels(Howard&Howard,2001;Howard,Howard,Dennisetal.
,2004;Howard,Howard,Japikseetal.
,2004).
Anothernewaspectofthepresentstudyisthatthesequencelearningtaskusedhererequiresthatpeoplelearnhigher-orderstructure.
UnlikethesimplerepeatingsequencesinthepreviousJ.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxx3studieswithdyslexics,thepredictiverelationshipstobelearnedspanatleastthreeevents(i.
e.
,thestimulusontrialnpredictsthestimulusontrialn+2).
Thisisimportantbecausethelevelofsequentialstructureinuencesthecognitiveandneuralsystemsengaged.
Forexample,healthyagingischaracterizedbydecitsinlearninghigher-order,butnotlower-orderrepeatingsequences(Curran,1997a;Howard&Howard,1997;Howard,Howard,Japikseetal.
,2004).
Therefore,itislikelythatsequencescontainingonlyhigher-orderstructurewillbemoresensitivetosequencelearningdecits.
Inthepresentstudy,wetestedcollegestudentswithandwith-outahistoryofdyslexiaconsistentwithpreviousresearch(e.
g.
,Kellyetal.
,2002;Pothos&Kirk,2004).
Wecomparedtheirperformanceontwoimplicitlearningtasks.
Onthebasisoftheknowndissociablefunctionalcorrelatesunderlyingthesetwotasksandthewell-documentedweaknessinlearningreading-relatedskills,wepredictedthatstudentswithahistoryofpoorreadingskillwouldshowimpairedimplicitlearningonthehigher-ordersequencelearningtask,butunimpairedimplicitlearningonthespatialcontextualcueingtask.
1.
Method1.
1.
GeneralprocedureParticipantswerescheduledforeither2(13people)or3(10people)test-ingsessionsonseparatedays.
Ontherstday,participantssignedaninformedconsentapprovedbytheInstitutionalReviewBoardsofbothGeorgetownandCatholicUniversities.
Theythencompletedthecontextualcueingtask(SCCTdescribedbelow).
Thealternatingserialreactiontimetask(ASRTT)wascom-pletedeitherintwosessionsontheseconddayorinindividualsessionsontwoseparatedays.
Arestbreakofatleast15minwasgivenbetweensessionsforthosetestedonasingleday.
Severalstandardizedneuropsychologicaltests(seebelow)wereadministeredfollowingtheexperimentaltasksoverthetwoorthreedaysofthestudy.
1.
2.
ParticipantsTwenty-threecollegestudentvolunteers(11dyslexicand12non-dyslexic)participatedintheexperiment.
Eightofthedyslexicparticipantsrespondedtoane-mailsolicitationsenttodyslexicstudentsbytheCatholicUniversityDis-abilitySupportServicesOfce.
Theseindividualshadadocumentedhistoryofdyslexiaasrequiredtoreceivedisabilitysupport.
Theremainingdyslexicsandthetypicalcontrolsrespondedtoyersplacedoncampus.
Documentationwasnotobtainedforthethreeadditionaldyslexics;however,theydidnotdifferfromthedocumenteddyslexicsonastatisticalcomparisonofscoresonthestandard-izedtestsassociatedwithreadingability(seebelow).
Noneoftheparticipantshadbeeninasimilarstudyandeachwaspaidforparticipating.
Participantsunderwentbehavioraltestinginordertocharacterizetheirhand-edness,singlewordreading,verbalworkingmemory,spelling,phonemicaware-ness,rapidnamingandvocabulary(describedbelow).
AsmaybeseeninTable1,thetwogroupsdidnotdifferinageordigitspan,butasexpectedthedyslexicgroupscoredmorepoorlyontestsofsinglerealandnon-wordreading(wordidenticationandwordattack)aswellasskillsrelatedtoreading,suchasrapidautomatizednaming(RAN),phonologicalawareness,spellingandvocabulary.
Fourofthedyslexicsubjectsandnoneofthecontrolsreportedapriordiagnosisofattentiondecithyperactivitydisorder(ADHD).
Allparticipantswereright-handed,asdeterminedbytheEdinburghHandednessInventory(Oldeld,1971)andwerenativespeakersofEnglish.
Singlerealwordreadingandpseudoword(phoneticallyregularnon-word)decodingskillswereassessedusingsubtestsfromtheWoodcock–JohnsonPsycho-EducationalBattery:letter–wordidenticationandwordattack(Woodcock&Johnson,1990).
AuditoryworkingmemorywastestedbyaskingTable1ParticipantcharacteristicsControlDyslexicGender8F,4M5F,6MAge20.
25(1.
14)20.
59(1.
46)Handedness*93.
82(11.
80)61.
34(52.
25)Wordidentication(W–JWI)***133.
00(17.
04)102.
09(14.
61)Wordattack(W–JWA)**116.
67(15.
23)98.
54(8.
85)Digitspancombined(WAIS-III)108.
33(14.
82)99.
09(10.
44)Spelling(TWS)***119.
58(10.
41)99.
18(9.
68)Phonemeawarenessa(TAAS)*12.
75(.
45)11.
73(1.
42)Rapidautomatizednaming*102.
00(10.
32)88.
00(16.
05)WASIvocabulary*70.
00(8.
15)59.
18(10.
84)aRawscore,othersstandardscores.
*pcombined).
ThevocabularyportionoftheWechslerAbbreviatedScaleofIntelligence(WASI)requiredsubjectstoorallydenewordsandwasusedtoassesswordknowledge(Wechsler,1999).
Publishednormswereusedwhenavailable,withtheexceptionoftheRosnerTestofAuditoryAnalysisSkillsandtheRapidAutomatizedNamingTest.
Fortheformer,wepresentrawscoresandforthelatternormswerebasedonthedistributionofscoresfromalargesampleofadultsubjectsinwhomdevelopmentalreadingdisabilityhadbeenruledoutbychildhoodtesting(Felton,Naylor,&Wood,1990;Flowers,1995).
Allstan-dardizedscoreshaveameanof100andastandarddeviationof15,excepttheWASIvocabularythathasameanof50andstandarddeviationof10.
1.
3.
Alternatingserialreactiontimetask1.
3.
1.
DesignThedesignwasa2*2*8(Group*TrialType*Epoch)mixedfactorial,withGroup(dyslexicversuscontrol)asabetween-subjectsvariableandTrialType(patternversusrandom)andEpoch(1–8)aswithin-subjectsvariables.
1.
3.
2.
StimuliandapparatusFouropencircles(.
5each)weredisplayedhorizontallyontheiMaccom-puterscreen.
ThetoppartofFig.
1showsaschematicofthedisplay.
Theentiredisplaysubtended12ofvisualangleatthe56cmviewingdistance.
Aneventoccurredwhenoneoftheopencirclesbecamesolidblack.
Fourlabeledkeyswereusedforrespondingwiththemiddleandindexngerofeachhand(z,xandTargetlocationsweredeterminedbyarepeatingeight-elementstructureinwhichxedandrandomlocationsalternated.
Participantswereassignedoneofthesixuniquepermutationsofthexedsequencelocations(i.
e.
,ArBrCrDr,ArBrDrCr,ArCrBrDr,ArCrDrBr,ArDrBrCr,ArDrCrBr,inwhichA–Drepre-sentspatialpositionsorderedfromlefttoright).
Eachpermutationwasusedtwiceforthecontrolgroup,whereasveofthepermutationsoccurredtwiceandoneonlyonceforthedyslexicgroup.
Onrandomtrials,theeventsweresampledfromauniformdistributionsuchthatthefourlocationswereequallylikely.
Hence,unlikemanyprevioussequencelearningstudies,inthisstudythesameeventcouldrepeatonimmediatelysuccessivetrials.
1.
3.
3.
ProcedureParticipantsweretoldthattheyweretopressthekeyunderthetargetcircleasquicklyaspossiblewhilemaintainingapproximately92%accuracy.
Thesequenceregularitywasnotmentioned.
Participantscompletedtwo20-block4J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxxFig.
1.
Schematicofcomputerdisplayforthesequencelearning(upperpart)andcontextualcueing(lowerpart)tasks.
sessions.
Eachblockbeganwith10randomtrialsfollowedby80learningtrials,i.
e.
,10repetitionsofthe8-elementlongpattern.
Oneachtrialoneofthecircleslledinandremainedsountilacorrectresponseoccurred.
Reactiontimewasmeasuredfromtargetonsettotherstresponse.
Thenextstimulusfollowedthecorrectresponseafteraxed120msdelay.
Feedbackwaspresentedonthecomputerscreenaftereachblockaskingpeopletofocusmoreoneitherspeedoraccuracytoencouragerespondingatabout92%accuracy.
Inall,eachpersonrespondedto3200trialsor400repetitionsofthepattern.
Aftercompletingthese40blocks,participantscompletedseveraltasksdesignedtoassesstheirexplicitknowledge.
Intherst,theyrespondedtoasinglerecognitionblockinwhichtheyobservedasequenceof16eventsoneachof20trials.
Afterobservingthesequencetheywereaskedtoevaluateifithadoccurredduringtheresponsetrialsusingascaleof1(certainitdidnot)to4(certainitdid).
Onhalfthetrialstheeventsconsistedoftwopassesthroughthealternatingsequencethatwasusedontheresponsetrials(e.
g.
,BrCrArDr,beginningatarandomstartingpoint).
Ontheremainingtrialstheeventswereproducedbyafoilsequencemadeupofeitherthealternatingresponsesequenceinreverse(i.
e.
,DrArCrBr,againfromarandomstartingpoint)orarandomlygeneratedsequence.
Thereversesequenceensuredthattherst-andsecond-orderstatisticsofthetargetandfoilsequenceswereidenticalwhereasthesedifferedfortherandomfoils.
Followingrecognition,peopleundertookasortingtask.
Theyweregivenadeckof64cardseachofwhichportrayed3successivetrialsas3rowsof4circleseach,with1circledarkenedoneachtrial(theevent).
Therewasonecardforeachofthe64possiblethree-trialsequencesortriplets.
Participantswereaskedtoexamineeachcardcarefullyandsortitintooneofthreecategoriesreectingthefrequencywithwhichthattripletoccurredduringtheexperiment("mostfrequent,""somewhatfrequent"or"leastfrequent").
Inpreviousresearch,wehaveshownthissortingtasktobeasensitiveindicatorofexplicitknowledgeintheASRTtask(Japikse,Howard,&Howard,2001).
Theexperimentconcludedwithaninterviewtoprobedeclarativeknowledgeofthesequence.
Peoplewereaskedaseriesofincreasinglyspecicquestions,rangingfrom"Doyouhaveanythingtoreportregardingthetask"to"didyounoticeanyregularityinthewaythestimuluswasmovingonthescreen"Finally,theyweretoldthattherewasinfactaregularitythatoccurredoneveryothertrialandtheywereaskedagaintoidentifyit.
1.
4.
Spatialcontextualcueingtask(SCCT)1.
4.
1.
DesignThedesignwasa2*2*6(Group*Conguration*Epoch)mixedfacto-rial,withGroup(dyslexicversuscontrol)asabetween-subjectsvariableandConguration(repeatedversusnew)andEpoch(1–6)aswithin-subjectsvari-ables.
1.
4.
2.
StimuliandapparatusThestimuliconsistedof12-elementarraysof11distractorsandasingletargetshownonanAppleiMac15in(38cm)monitoraswhitecharactersonagraybackground.
AsshowninthelowerhalfofFig.
1,thetargetwasahorizontalTwiththetailpointingeitherleftorrightandthedistractorswereL'srandomlyrotatedby0,90,180or270.
FollowingChunandPhelps(1999)Experiment2,theLlegwasoffsetbythreepixelstoincreasesimilaritywiththetarget.
Eachelementsubtendedapproximately1.
1ofvisualangleataviewingdistanceof56cm.
Arraysweregeneratedbyrandomlyplacingthe12itemsintocellsofaninvisible6*8(rows*columns)grid.
Acrossarrays,targetlocationwasbalancedforeccentricitywithrespecttothecenterofthescreenaswellasforleftandrightscreenhalf.
Targetsneverappearedinthefourcentercellsorattheextremecornersofthedisplaygrid.
Everyelementwasrandomlyrepositionedwithinitscellby±2pixelsalongeachaxistoavoidco-linearitywithotherelements.
Asetof12arrayswasconstructedforrepeatedpresentationacrosstheexperiment(detailsbelow).
Individualswithineachgroupreceivedadifferentsetofnewandrepeatedcongurations,butthesamesetswereusedacrossgroupswiththeirpresentationorderrandomized.
1.
4.
3.
ProcedureTheexperimentaltaskwasthesameasthatusedbyChunandJiang(1998).
Subjectscompleteda24-trialpracticeblockafterreceivinginstructions.
Trialsbeganwithawhitexationdot(approximately.
5)centeredonthescreen.
After1sthedotwasreplacedbyasearcharrayandparticipantshadtopressakeyindicatingthetargetorientation("z"forleftand"/"forrightpointing).
Theyweretoldto".
.
.
locatethe"T"onthescreen,determinewhichwayitisfacingandpressthekeythatcorrespondstothatdirectionasquicklyandasaccuratelyaspossible.
Anoccasionalerrorisacceptable(e.
g.
,oneerrorperblockof24trials).
"Auditoryfeedbackwasprovidedaftereveryresponse(abeeportonetosignalcorrectorerrorresponses,respectively).
Adifferentsearcharraywaspresentedoneachtrialinthepracticeblock.
Afterfurtherquestions,participantscompleted30blocksof24trials.
Theseblocksweresimilartothepracticeblockexceptthatonly12ofthesearcharraysineachblockwerenewcongurations.
Theremaining12arrays(repeatedcongurations)wererepeatedacrossblocks,appearingonceineachblock.
Therepeatedcongurationspredictedthelocationofthetargetelement,butnotitsorientation.
Presentationorderwasrandomizedwithinblocks,andpeoplewereencouragedtotakeashortbreakbetweenblocks.
Asinpreviousstudies(Chun&Phelps,1999;Manns&Squire,2001),trialsonwhicharesponsedidnotoccurwithina6-stime-outintervalwereaborted—atonesoundedandthenexttrialbegan.
Aftercompletingthese30blocks,participantswereaskedaseriesofpro-gressivelymorefocusedquestionstoobtaininsightsintotheirstrategyandtheirdeclarativeknowledgeofthetask.
Next,subjectsweregivenasingle24-trialrecognitiontest,consistingofthe12repeatedcongurationsand12othersnotpresentedduringlearning,inrandomorder.
Oneachrecognitiontrialsubjectsjudgedwhethertheyhadseen".
.
.
adisplaywithitemsinthesamescreenposi-tionsasthisearlierintheexperiment.
"Theyrespondedbypressingeitherakeylabeled"yes"oronelabeled"no.
"Theywereurgedtoguessiftheywereunsure.
Nofeedbackwasprovided.
2.
Results2.
1.
AlternatingserialreactiontimetaskInthisanalysis,wetestthehypothesisthatdyslexiccollegestudentswillshowimpairedhigher-orderimplicitsequencelearningcomparedtoage-matchedcontrols.
ThisJ.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxx5Fig.
2.
Meanaccuracy(uppergraph)andmeanofmedianRT(lowergraph)onpredictable(pattern)andunpredictable(random)trialsasafunctionofepochforbothgroupsontheserialreactiontimetask.
wasaccomplishedbycomparingtheaccuracyandspeedofrespondingontherepeating"pattern"trialstotheunpredictable"random"trials(seeHoward&Howard,1997).
Medianreactiontimeswererstdeterminedseparatelyforcorrectpatternandrandomtrialsoneachblockforeachperson.
Followingtheconventionestablishedinpreviouswork,thesemedianswerethenaveragedacrosssuccessiveblockstoobtainavalueoneachof8ve-block(400-trial)epochsforeachindividualandTrialType(patternorrandom).
Asimilardatareductionwasperformedonaccuracy.
Astatisticalcriterionof.
05wasusedinallsignicancetests.
2.
1.
1.
TrialTypeeffectsonaccuracyandspeedFig.
2plotsthemeanaccuracy(uppergraph)andmeanofmedianRT(lowergraph)dataforbothgroups.
ThesedataweresubmittedtoGroup(dyslexicversuscontrol)*TrialType(patternversusrandom)*Epoch(1–8)mixeddesignANOVAswithrepeatedmeasuresontheTrialTypeandEpochfac-tors.
Althoughthedyslexicgrouprespondedsignicantlymoreslowlythanthecontrolgroup(422msversus365msoverall),F(1,21)=13.
50,MSE=22379,theydidnotdifferfromthecon-trolsinoverallaccuracy(93and91%,respectively).
Hence,thefeedbackprovidedtoequatethetwogroupsat92%accu-racywaseffective.
TherewerealsosignicantmaineffectsofEpochonbothmeasuresreectingoverallmotorskilllearning,F(7,147)=5.
16,MSE=.
003andF(7,147)=53.
89,MSE=427,foraccuracyandreactiontime,respectively.
Moreimportantly,however,thereisevidenceofhigher-ordersequencelearninginthatthepatternandrandomtrialsdivergeinbothaccuracyandspeedwithpractice.
Thiswasdemon-stratedbysignicantmaineffectsofTrialTypeforbothaccu-racy,F(1,21)=69.
85,MSE=.
001,andspeed,F(1,21)=25.
77,MSE=227,aswellasTrialType*Epochinteractionsforbothaccuracy,F(7,147)=4.
76,MSE=4.
09E4,andspeed,F(7,147)=6.
03,MSE=55.
37.
Thisindicatesthatresponsesonrandomtrialsareslowerandlessaccuratethanthoseonpat-terntrials,andthatthisdifferenceincreasesacrossepochsforbothgroups.
Furthermore,accuracyishighandrelativelycon-stantonpatterntrials,butdeclinesacrossepochsonrandomtrials.
Thispatternofincreasingerrorsontheunpredictable,randomtrialswithpracticeistypicalwhenprobabilisticregu-laritiesareused(Curran,1997a;Howard,Howard,Japikseetal.
,2004;Schvaneveldt&Gomez,1998).
Participantsoftenreportthattheirngersseemtotakeoverleadingthemtomakemore"oops"errors.
Unbeknownsttothem,theseerrorsoccurprimar-ilyontheunpredictablerandomtrials,andhencereectlearningofthesequencestructure.
Thus,whenprobabilisticsequencesareused,errorsareassensitivetolearningasareresponsetimes.
Althoughbothgroupsshowsequencelearning,thedyslexicsshowsignicantlylesslearningthancontrolsonbothmea-sures.
ThisissupportedbysignicantTrialType*Groupinter-actionsforaccuracy,F(1,21)=25.
64,MSE=.
001,andspeed,F(1,21)=4.
61,MSE=226.
58.
Fig.
2alsosuggeststhatlearningoccursmoreslowlyfordyslexicsinthatthedifferencebetweenpatternandrandomtrialsappearstodevelopmoregraduallywithpracticeforthedyslexicthancontrolgroups.
Despitethis,thethree-wayinteractionsdidnotreachsignicanceforeithermeasure.
Thismayreecttherelativelylowpowerfromoursmallsamplesizeaswellasthedirectionalinsensitivityofthisomnibustest.
Follow-uptwo-way(TrialType*Epoch)ANOVAswerecarriedoutseparatelyonthetwogroups.
Resultsindicatedthatbothgroupsshowedsignicantlearningonbothmea-sures.
Forthedyslexicgroup,themaineffectofTrialTypewassignicantfortheaccuracymeasure,F(1,10)=17.
57,MSE=2.
6E4,andmarginallysignicantforthespeedmeasure,F(1,10)=3.
31,MSE=280.
99,pcomeawareofthealternatingstructureofthesequenceevenafterpracticingformorethan10,000trials(Howard,Howard,Japikseetal.
,2004).
Wehavealsoshownintrial-by-trialanalysesthatperformancebecomesincreasinglysensitivetothelocalsequencecontext—specicallytothefre-quencywithwhichrunsoflengththree,i.
e.
,triplets,occur(Howard,Howard,Japikseetal.
,2004).
Tripletsarerelevant6J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxxbecauseinthealternatingsequencesweuse(e.
g.
,ArBrCrDr,.
.
.
)thelowestlevelofpredictiverelationshipoccursbetweeneventsthatareseparatedbyalagoftwotrials(i.
e.
,trialn2predictstrialnorBpredictsCintheaboveexample).
Inthefollowingwecomparethesensitivityofthedyslexicandcontrolgroupstothetripletstructureofthesequence.
Therearetworeasonswhythisanalysisisimportant.
First,pre-existingresponsetendenciesexistforsometriplets.
Forexample,becauseofperceptualand/ormotorprimingpeopletendtorespondveryquicklytorepetitions(e.
g.
,CCC)whereastheyrespondslowlytotrills(e.
g.
,CDC)(Howard,Howard,Japikseetal.
,2004;Remillard&Clark,2001).
Sincerepeti-tionsandtrillscanendonlyonrandomtrialsinthealternatingsequences,thesetendenciesmaycontaminatelearningmea-suresbasedexclusivelyontheTrialTypeeffect(i.
e.
,patternversusrandomtrials)reportedabove.
Second,wehaveshownthatpeopleoftenacquiredeclarativeknowledgeorinaccuratehypothesesaboutthelikelihoodofthesedistinctivetriplets.
Forexample,manypeoplereport(incorrectly)thatrepetitionsoccurfrequentlyintheASRTtask(Howard,Howard,Japikseetal.
,2004).
Hence,inthefollowinganalysiswecompareresponsestohigh-andlow-frequencytripletsafterresponsestorepetitionsandtrillshavebeenremoved.
High-frequencytripletsoccuronallpatterntrialsandonsomerandomtrialsbychance(e.
g.
,AxB,BxC,etc.
,fortheaboveexample,wherexreectsanyofthefourevents)whereaslow-frequencytripletsoccuronlyonrandomtrials(e.
g.
,AxD,DxB,etc.
).
Eachperson'seventsequencewasparsedintoaseriesofover-lappingtripletsusingaslidingthree-trialwindow(seeHoward,Howard,Japikseetal.
,2004fordetails).
Eachtripletwasthensortedintooneoffourcategories;repetitions,trills,highorlowfrequency.
Repetitionscontainthreesuccessiveidenticaleventsandtrillsbeginandendwiththesameelementbutwithadif-ferentmiddleelement.
Forthereasonsoutlinedabove,thesetwotriplettypeswerenotincludedinthefollowinganalyses.
High-frequencytripletsincludedthoseendingonapatterntrialaswellasthoseendingonrandomtrialsthatbychanceformastructure-consistenttriplet.
Low-frequencytripletsincludedalltheremainingtriplets,i.
e.
,thoseendingonrandomtrialsthatareneitherstructureconsistent,repetitions,ortrills.
Overall,thereare16possiblehigh-frequencytripletsthatoccuron62.
5%ofthetrialsand32low-frequencytripletsthatoccuron25%ofthetrials.
Theremaining25%ofthetrialsareeitherrepetitions(4kinds)ortrills(12kinds).
Three-waymixedANOVAs(Group*TripletType*Epoch)revealedanidenticalpatternofresultstothoseseeninthetrialtypedatareportedaboveforaccuracy.
However,forreactiontimetheTripletType*Groupinteractiondidnotreachsig-nicancedespiteatrendtowardlesslearningforthedyslexicgroup.
Hence,thisanalysissupportstheoverallconclusionthatwhilebothgroupsshowhigher-orderimplicitsequencelearning,dyslexicsareimpairedindoingso.
2.
1.
3.
RecognitionanalysisTherecognitiondatawereanalyzedbydeterminingthemeanratingassignedtothefoilsandtargetsforeachindividual(1:certainitdidnotto4:certainitdid).
Apreliminarycompari-Table2Meanrecognitionratings(standarddeviations),ASRTtaskFoilTargetControl2.
55(.
92)2.
71(.
81)Dyslexic2.
71(.
84)2.
71(.
95)Table3Meanproportion"mostoften"category(standarddeviations),ASRTtaskHighfrequencyLowfrequencyRepetitionsTrillsControl.
58(.
11).
51(.
10).
48(.
30).
39(.
13)Dyslexic.
57(.
11).
56(.
10).
61(.
43).
50(.
16)sonoftherandom(10people)andbackward(12people)foilsrevealednodifferencesinrecognitionperformancesothisdis-tinctionwasnotconsideredfurther.
Recognitiondatafromonedyslexicparticipantwerelostduetoacomputererror.
AsmaybeseeninTable2,themeanratingswerevirtuallyidenticalacrosssequencetypeforthetwogroupsindicatingthatneithergroupwasabletodistinguishbetweenthetargetandfoilsequences.
Thiswasconrmedbyatwo-way(Group*SequenceType)ANOVAthatyieldednosignicanteffects.
Thus,peoplewereunabletoexpressknowledgeofthesequencestructureinanexplicitrecognitiontask,despiterevealingsensitivitytoitintheirresponding.
ThisisconsistentwithpreviousndingsinrevealingthatlearningintheASRTtaskisimplicit.
2.
1.
4.
SortingtaskanalysisTodetermineifpeoplewereabletojudgeexplicitlytherelativefrequencywithwhichvarioustripletsoccurred,wecal-culatedthemeanproportionoftimeshigh-frequency(structureconsistent)andlow-frequencytripletsweresortedintothe"mostoften"category.
Forthereasonsarguedabove,wealsodistin-guishedrepetitionsandtrillsforthisanalysis.
ThesortingdataareshowninTable3.
Atwo-way(Group*TripletType)ANOVAcarriedoutonthesedatarevealednosignicanteffects,indicatingthatneitherthedyslexicnorcontrolgroupswereabletoexplicitlyevaluatethefrequencywithwhichthedifferenttriplettypesoccurred.
Despitethis,bothgroupsrevealanon-signicanttendencytoratetrillsasoccurringlessoftenthantheothertriplettypes.
Thiscouldreecteitherapre-existingbiasorlearning,butcouldnot,ineithercase,haveinuencedtheresultsofthehighfrequencyversuslowfrequencyanalysisoflearningabove,inwhichbothtrillsandrepetitionswereeliminated.
Hence,inkeepingwithourpreviousndings,subjectswerenotabletoexpresstheirknowledgeofthetemporalstructureinanexplicitsortingtaskaddingtotheevidencethatlearningisimplicit.
2.
1.
5.
InterviewResponsesonthepost-experimentalinterviewwereexam-inedforevidenceofdeclarativeknowledge.
Theyrevealednoapparentdifferencesbetweenthetwogroups.
Mostpeoplereportedthattheyfeltthattherewassomeregularity,butthatitwastoosubtleforthemtopickup.
Noonereportedtheyhadfoundapatternand,specically,nooneidentiedthealternatingstructureofthesequence.
J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxx7Whenspecicallytoldtoguess,themostfrequentdescrip-tionswereeithervague,thelightsmovedinanorderlyway,"orincorrect".
.
.
somepositionsoccurredmoreoftenthanothers,"and".
.
.
doublesortriplesofthesameitemswerecommon.
"Althoughpeoplecouldnotdescribetheregularity,somefeltthattheywerelearningsomething,perhapsuncon-sciously,therewasapattern,butIcan'tdescribewhatitis.
"Thus,consistentwithourearlierworkwiththistask,peo-plewereunabletodescribetheregularitiestowhichtheyhadbeenexposedforthousandsoftrials.
Nonetheless,mostbelievedthattherewasaregularitythatinuencedtheirperformanceinsomeway.
Thisisconsistentwiththeevidencefromtherecognitionandcard-sortingtasksthatthelearningrevealedbybothgroupswasimplicitanduncontaminatedbyexplicitlearning.
2.
1.
6.
CorrelationsbetweenimplicitlearningandindividualreadingabilityInadditiontothegroupanalysesreportedabove,weper-formedaseriesofcorrelationstoexaminetherelationshipbetweenimplicitsequencelearningandreadingabilityasmea-suredontwostandardizedtestsofsinglerealandpseudowordreading(Woodcock&Johnson,1990).
Thiswasmotivatedbypreviousstudiesthathaveexaminedtherelationshipbetweenindividualreadingabilityandimplicitlearning(e.
g.
,Waberetal.
,2003).
Wordidenticationmeasuressinglerealwordreadingandwordattackisameasureofphonologicaldecod-ing(ascontextormemorycannotbeappliedtonamethesenon-words).
Unlikethegroupcomparisons,thesecorrelationspermitustoinvestigatetherelationshipbetweenindividualreadingabilityandimplicitlearningindependentofdiagnosticcategory.
Twomeasuresofterminalsequencelearningweredeter-minedbycalculatingthedifferenceinmeanaccuracy(patternaccuracyrandomaccuracy)andmeanreactiontime(randomRTpatternRT)onthenaltestingepoch(epoch8)foreachindividual.
Thesevalueswerethencorrelatedwiththesinglerealword(WI)andsinglepseudowordreading(WA)scoresdescribedabove.
Thisrevealedsignicantpositivecorrelationsbetweenbothmeasuresofreadingabilityandtheaccuracy-basedimplicitlearningscore(r=.
59,pcomparedthemtoanage-matchedgroupof18collegestu-dentsfromanearlierstudy(Howard,Howard,Dennis,etal.
,2004,Experiment1).
Thepresentcontrolshadverysimilarnalcontextuallearningscorestotheearliergroup(epoch6differ-encesof.
116and.
127s,respectively),andathree-wayANOVAcarriedoutonthelearningdataproducednosignicantgroupdifferences.
Hence,ourcontrolsdonotappeartobeatypical.
2.
2.
2.
ErroranalysisAlthoughpeopleinbothgroupsmaderelativelyfewerrors(4.
5and4.
0%forthedyslexicandcontrolgroups,overall),wecarriedoutasimilaranalysisoftheerrordata.
ThisrevealedonlyasignicantmaineffectofEpoch,F(5,150)=7.
20,MSE=.
001,reectinganoveralldecreaseinerrorswithpractice(from6.
5to3.
6%).
2.
2.
3.
RecognitionanalysisToinvestigateiflearningwasimplicit,wecalculatedtherat-ingaccuracyfortherepeatedandnovelcongurationsontherecognitionblockforeachperson.
Accuracywasnearlyidenticalacrossgroups(52and55%forcontrolanddyslexic,respec-tively)andcongurations(53and55%fornovelandrepeated,respectively).
AGroup*CongurationANOVArevealednosignicanteffectsandaccuracywasnotsignicantlydifferentfromchanceinanycondition.
Thus,spatialcontextlearningoccursimplicitlyforboththedyslexicandcontrolgroups.
2.
2.
4.
InterviewCommentsfromthepost-experimentalinterviewwereexam-inedforevidenceofdeclarativeknowledge.
Aswithsequencelearning,nosystematicdifferenceswereevidentbetweenthetwogroups.
Ontheopen-endedquestionspeoplefrequentlyindi-catedincorrectlythatthetargetdidnotoccuronsomeofthetrialsandveryfewsaidthattheythoughtsomedisplaysrepeated.
Hence,theinterviewdataareconsistentwiththerecognitiondatareportedaboveaswellaswithpreviousndingswiththistaskinrevealingnoexplicitknowledgeoftherepeatedcongurations.
2.
2.
5.
CorrelationsbetweenimplicitspatialcontextlearningandindividualreadingabilityThedifferenceinmeanresponsetimebetweennovelandrepeatedcongurationsonthenaltestingepoch(epoch6)wasdeterminedforeachindividual.
Theselearningscoreswerethencorrelatedwiththereadingandspellingscoresasinthesequencelearninganalysisdescribedearlier.
Thisrevealedsig-nicantnegativecorrelationsbetweenbothmeasuresofreadingability(r=.
44,pcomparisons(basedprimarilyonachildhooddiagnosisofdyslexia)andtheindividualcorrelationalanalyses(basedonsinglerealandnon-wordreadingscores).
Thusthesendingsrevealadissociationbetweentwoformsofimplicitlearninginhigh-functioningdyslexicadultswithastrongtrendtowardadoubledissociation:college-agestudentswithahistoryofdyslexialearnrelationshipsamongsimultane-ousspatialstimuliatleastaswellascontrols,buttheyarepooreratlearningrelationshipsamongnon-adjacenteventsintemporalsequences.
Theseresultssupportandextendtheexistinglitera-tureonimplicitlearningindyslexiainanumberways.
First,thepresentndingsindicatethattheimpairedsequencelearningobservedinthisandsomepreviousresearch(Vicarietal.
,2005)doesnotreectageneralcognitiveorattentionaldecit.
Ifthiswerethecase,dyslexicindividualsshouldshowimpairmentonboththesequencelearningandcontextualcueingtasks,ratherthantheselectiveimpairmentobserved.
Norcanpoorlearningontheserialreactiontimetaskbeattributedtotherelativeoveralldifcultyofthetwotasks.
Thereisnoevidencethatsequencelearningwasmoredifcultthancontextualcuingforeithergroup.
Onthecontrary,mostparticipantsreportedthatcontextualcueingwasmoredifcult,anobservationconsistentwiththesubstantiallylongerresponsetimesthatoccurinthistask.
Furthermore,thesamplewithahistoryofdyslexiashowedthesamepatternofoverallperformancecomparedtocontrolsonbothtasks;theyweresloweroverall,butjustasaccurate.
Second,theresultsestablishthatdecitsinimplicitsequencelearningoccurevenwhenexplicitlearningcanberuledout.
Wehavedemonstratedhereandinearlierresearch(Howard,Howard,Japikseetal.
,2004),thatthepresenttaskstaprela-tivelypureimplicitlearning,intheabsenceofexplicitlearning.
Inthepresentstudy,neithergroupshowedevidenceofdeclara-tiveknowledgeforeithertaskininterviewsoronothersensitivemeasuresofexplicitknowledge,includingrecognitionandcardsortingforsequencelearningandforced-choicecongurationrecognitionforcontextualcueing.
Furthermore,previousstud-ieshavereportedimpairedimplicitbutsparedexplicitlearningindyslexia(Sperlingetal.
,2004)suggestingthatevenifexplicitlearninghadoccurredduringourtask,whichwebelievetobeunlikely,wewouldnotexpectittoaffectthetwogroupsdifferentially.
Asanaddedprecautionweexaminedthepossi-bilitythatsomeindividualsgainedawarenessinaseriesofposthoct-testscomparingindividualrecognitionratingsforthefoilandtargetsequencesand"mostoften"card-sortingfrequenciesforthehigh-andlow-frequencytriplets.
Therewereonlytwosignicantcomparisons:acontrolsubjecthadasignicantlyhighermeanrecognitionratingfortargetthanfoilsequencesandadyslexicparticipantsortedthelow-frequencytripletsintothe"occurredmostoften"categorymoreoftenthanthehigh-frequencytriplets—thereverseofwhatactuallyoccurred.
Thus,onlyonepersonshowedstatisticalevidenceofawarenessontherecognitiontaskandnooneshowedsuchevidenceonmorethanoneoftheexplicitknowledgemeasures.
Hence,thepresentnd-ingscannotbeexplainedbydifferencesinexplicitknowledgebetweenthetwogroups.
Whilesomeearlierstudiesofdyslexicchildrenhadreportedimpairednon-linguisticsequentialprocessingontheserialreac-tiontimetask(Vicarietal.
,2003,2005,howeverseeWaberetal.
,2003),theonlypreviousstudyusingtheserialreactiontimetaskindyslexicadults(Kellyetal.
,2002)didnotleadingtospeculationthatthisdiscrepancymightbeattributabletoage10J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxx(Vicarietal.
,2005).
Ourthirdconclusionistherefore,thatourndingsrefutethissuppositionandsuggestthatsequencelearn-ingdecitsarenotlimitedtochildhood,butdoindeedoccurinadults.
WeneedtoconsiderwhyourndingsdifferfromthosereportedbyKellyandcoworkers,whofoundnosequencelearn-ingdecits.
Sincethetwostudiesusedcomparablenumbersofcollegestudentswithachildhoodhistoryofdyslexia,itisunlikelythatthedifferentndingsrelatetosampleselection.
Asistypicalofdyslexiccollegestudents,theirreadingwasinthenormalrangeforthegeneralpopulation,butsignicantlyworsethantheirnon-dyslexicuniversitypeers.
Forbothstudies,theselectionofuniversitystudentsmeansthattherearelimitationswithregardstogeneralizingtheseresultstotheentirepopula-tion.
Futurestudiesinadultsthataremorerepresentativeofthegeneralpopulationwillneedtobeconducted.
OurdyslexicsampledidincludefoursubjectswithADHD,whereasKellyandcoworkersdonotreportiftheirsamplewasscreenedforADHD.
Toruleoutthepossibilitythatparticipantswithacomorbiddiagnostichistory(ADHDanddyslexia)weredrivingtheeffects,were-analyzedthesequencelearningdatawithouttheseindividualsandfoundanidenticalpatternofmeansandstatisticalsignicance.
Thus,wecanruleoutanyconcernthattheADHDcasesareresponsiblefortheeffectsobservedontheSRTtask.
EvenifthepresenceofADHDplayedaroleinsubject'sperformanceonimplicitlearning,thereisnoreasontobelieveitwoulddifferentiallyaffectthesequencelearningandcontextualcueingtasks.
ThemostlikelyexplanationforthediscrepancybetweenourndingsandthoseofKellyandcoworkersisbasedontheserialreactiontimetaskitself.
Forexample,thefactthattheyusedsimplerepeatingsequencescouldaccountforthediffer-ence.
Specically,bydemonstratingadecitinhigher-orderimplicitsequencelearning,i.
e.
,learningthatrequiresintegrat-ingacrossatleastthreeelements,thepresentstudycallsattentiontothepotentialimportanceofsequencecomplexityinstudiesofimplicitsequencelearningindyslexia.
Todate,sequencestruc-turehasnotbeenvariedsystematicallyinstudiesofdyslexia,andpreviousstudieshaveusedrelativelysimplerepeatingpat-terns.
Thereisevidencethatdifferentlevelsofstructurecallondifferentbrainsystems(Curran,1997b;Fletcheretal.
,2004;Howard,Howard,Dennisetal.
,2004)raisingthepossibilitythatdyslexiamayinuencethelearningofsimpleandcomplexsequencesdifferently.
Futurestudiesneedtoexaminethispos-sibilitysystematically.
SowhatdoourndingssuggestabouttheroleofpoorimplicitlearninginindividualswhoarepoorreadersFirst,itisimportanttounderscorethatourresultsdemonstratethatdyslexicsdonotsufferfromanoveralldecitinimplicitlearning.
Althoughbothtasksinvestigatedinthepresentstudyinvolveimplicitlearning,theywereselectedtocomplementeachother:thealternatingserialreactiontimetaskhasastrongsequencingcomponentwhereasthecontextualcueingtaskdoesnot.
Thisdistinctionenablesustoseparatesequencingdecitsfromotherformsofimplicitlearningdecits.
Thedifferenceswereportinimplicitlearningofsequentialversusspatialinformationareconsistentwithresultsfromacomputer-basedtestinwhichchildrenwithdyslexiashowedimpairedtemporalprocessingwithpreservedspatialprocessingofvisualstimuli(Eden,Stein,Wood,&Wood,1995).
Interestinglythereisevidenceofahigherprevalenceofdyslexiaamongartiststhannon-artistsatcollege(Wolff&Lundberg,2002)suggestingthatdyslexicsmayberelativelygoodatconguralvisualprocessing.
Itispossiblethatthedis-crepantndingsfromearlierstudiesaredueinparttoafailuretodistinguishamongthesedifferentkindsofimplicitlearning.
Ourresultssuggestthattheweaknessinimplicitlearningcanbenarroweddowntoparadigmsthatinvolvesequentialpro-cessing.
Furthermore,evenwithinoneformofimplicitlearning,suchasimplicitsequencelearning,itisimportanttoconsiderthelevelofstructurethatispresent.
Anumberofauthorshavearguedfortheimportanceofimplicitlearninginlearningtoreadandindyslexia(Gombert,2003;Sperlingetal.
,2004).
Thereareseveralplausiblemechanismsbywhichaselectiveweaknessinimplicitlearningofsequentialinformationcouldaccountforthephonologicalprocessingandreadingproblemsthatarethecardinalfeatureofdyslexia.
Theyareaddressedbelowbyconsideringtheroleof:(1)automaticity,(2)phonemicawarenessand(3)orthographicawarenessinattainingreadingskills.
Finally,thegrowingliteraturedescribingpatientpopula-tionsoremployingfunctionalbrainimagingtechnologytostudyimplicitlearningwillbeconsideredtorelateourndingstothefunctionalanatomyofimplicitlearning.
Ourresultsareconsistentwithearlierstudies(Vicarietal.
,2005)inindicatingthattheimplicitsequencelearningdecitisnotlimitedtolinguisticmaterials,butratherrepresentsamoregeneralsequencingproblem.
Theconceptofadecitinnon-linguisticprocessingindyslexiahasbeendescribedusingdifferenttheoreticalframeworks,suchasimpairedinformationprocessingattributedtolow-levelsensoryperception(Stein&Walsh,1997),alackofautomaticityinthecontextofdualtaskperformance(Nicolson&Fawcett,1999)orinthecontextofarticulatoryuency(Wolf,Miller,&Donnelly,2000).
Thecur-rentstudywasnotdesignedtotestanyofthesetheories,butfuturestudiescouldexaminetheseaspects.
Inthepresentstudy,wedidmeasurerapidautomatizednam-ingskills(Wolf,1986)andfoundthesetobeimpairedinoursamplewithchildhoodreadingdecits,consistentwiththelit-eraturearguingthatrapidnamingproblemsinchildrenwithreadingdisabilitiespersistsintoadulthood(Korhonen,1995).
Toexploreadhocwhetherpoorrapidautomatizednamingskillsarerelatedto"uency"inperformanceofsequentiallearning,par-tialcorrelationswereperformedtoexploretheinuenceofRANonthecorrelationbetweenwordreading(realandnon-word)andtheaccuracymeasureofimplicitsequencelearningdescribedintheresultssection.
ThecorrelationsremainedthesamewithRANpartialedout,providingnoevidenceofapossibleconnec-tionbetweenSRTTperformanceandrapidautomatizednaming.
OurndingsaremoreconsistentwiththeexplanationofferedbySperlingetal.
(2004)whohavearguedthatpoorimplicitlearningcouldhindertheestablishmentofgoodphonologicalprocessingaswellaslearningorthographic–phonologicalrepresentations.
Gombert(2003)proposedthatchildrenwithdyslexiahaveaphonologicaldecitthatpreventstheimplicitlearningoflinguisticregularitiesand,hence,interfereswithJ.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxx11reading.
However,thendingsfromSperlingetal.
andthepresentstudysuggestthatthecausalitymaybesomewhatmorecomplex.
Forexample,thecombinationofaphonologicaldecitwithanimpairedimplicitsequencelearningsystemcouldleadtotheobservedreadingdisabilities.
Thiscombinationcouldmanifestasafailureinapplyingimplicitorprobabilisticrulesrequiredforuentapplicationofgrapheme–phonemecorrespondences(Sperlingetal.
,2004).
Inoursample,skillsthatinvolveorrequirephonologicalprocessing(i.
e.
,soundelisionmeasuredwiththeTestofAuditoryAnalysisSkillsandnon-worddecodingmeasuredviaWordAttack),wereimpaired.
Likewise,ourdyslexicsubjectshadsignicantdecitsinspelling.
Residualdecitsinencodingareoftenseeninadultswithdyslexiaevenoncetheyhaveovercomesomeoftheirreadingproblems(Flowers,1995).
Notsurprisinglytherefore,posthocanalysisofspellingability(TWS)provedtocorrelatesignicantlywithourmostinformativelearningscore(theaccuracymeasureofsequencelearning).
Spellingsharesagreatdealofvariancewithmeasuresofreadingduetotheircommonrelianceonphonologicalandorthographicprocessing.
Theformerrelationshipwasillustratedbytheobservationthattherelationshipbetweensinglerealwordreading(WI)andsequencelearningremainedsignicantwhenthespellingwaspartialedout(usingTWS),buttherelationshipbetweensequencelearningandnon-wordreading(WA)wasnolongersignicantwhenspelling(TWS)wasaccountedfor.
Surprisingly,addingothervariablessuchasphonemicaware-ness(TAAS)andverbalworkingmemory(digitspan)hadnoeffectandthecorrelationsbetweenimplicitsequencelearningandreading(WIandWA)remainedlargelythesameundertheseconditions.
Ourdatathereforedonotprovideaclearpictureorallowforastraightforwardinterpretationoftherelationshipbetweenimplicitsequencelearningandphonologicalprocess-ing.
However,itcanbestatedthatnotonlyistherelationshipbetweensequencelearningandrealwordreadingthestrongest,butitisalsothemostrobustrelationship,survivingthepartialcorrelationswithmeasuresofspelling,phonemicawareness,rapidautomatizednamingandverbalworkingmemory.
Ithasbeensuggestedthatdyslexic,unliketypicalreaderswhousearule-basedapproachtosoundoutwords,compensatebyrecol-lectingwordsbysight(Manisetal.
,1987).
Wedidnotacquireameasureofpredictableandunpredictablewordreading,butnotedinoursampleofadultswithahistoryofdyslexiathattheirperformancewasequalonthespellingofpredictableandunpre-dictablewords.
Therelationshipbetweenorthography,phonol-ogyandsequencelearningthereforerequiresfurtherstudy.
Aparticularlyinterestingtheoreticalquestionwouldbetherela-tionshipbetweenreadingscoresandthehigher-orderSRTTwithdifferingwordtypes,totestthepredictionthatthereadingofwordswithmorecomplexpronunciationruleswouldbemorestronglypredictedbymeasuresofhigher-orderratherthansim-plesequencelearning.
Turningtotheanatomicalcorrelatesoflearning,itisknownthatdifferentformsofimplicitlearningmakedifferentcognitivedemandsandcallondifferentneuralsubstrates.
Thetwotasksweinvestigatedarethoughttorelyondifferentunderlyingbrainsys-tems(Howard,Howard,Dennisetal.
,2004).
Thisunderscoresthefactthatimplicitlearningisnotaunitaryphenomenondepen-dentonasinglebrainsystem,butratheritrepresentsarangeoftasksthatengagedifferentneuralsystems.
Evidencefrompatient,functionalneuroimaging,andtran-scranialmagneticstimulationstudiesindicatesthatsequencelearningdependsonfronto-striatal-cerebellarcircuitry(Prulletal.
,2000;Robertson,Tormos,Maeda,&Pascual-Leone,2001).
Forexample,patientswithfocalcerebellarorfrontallesionsrevealimpairedlearninginanSRTT(Gomez-Beldarrain,Garcia-Monco,Rubio,&Pascual-Leone,1998;GomezBeldarrain,Grafman,Pascual-Leone,&Garcia-Monco,1999;GomezBeldarrain,Grafman,RuizDeVelasco,Pascual-Leone,&Garcia-Monco,2002)asdoindividualswithstri-ataldisorderssuchasHuntington's(Willingham,Koroshetz,&Peterson,1996)andParkinson'sdisease(e.
g.
,Dominey&Jeannerod,1997;Helmuth,Mayr,&Daum,2000;e.
g.
,Jackson,Jackson,Harrison,Henderson,&Kennard,1995).
Functionalneuroimaging(e.
g.
,Grafton,Hazeltine,&Ivry,1995;Rauchetal.
,1997;Seidleretal.
,2002)andtranscranialmagneticstimulation(Robertsonetal.
,2001)studieshavebeengener-allyconsistentwiththepatientndings.
Althoughthereissomeevidencethatthemedialtemporallobemayalsobeinvolvedwhenhigher-ordersequencesareused(Curran,1997a;Fletcheretal.
,2004;Schendan,Searl,Melrose,&Stern,2003),theSRTTdependsprimarilyonfronto-striatal-cerebellarcircuits.
Incontrasttosequencelearning,thecontextualcueingtaskseemstodependprimarilyonmedialtemporallobestruc-tures.
Forexample,amnesicpatientswithlesionsinvolvingparahippocampalregionspossiblyinvolvingthehippocampusareimpairedincontextualcueingcomparedtohealthycon-trols(Chun&Phelps,1999;Manns&Squire,2001).
Thereisalsopreliminaryevidencethatcontextualcueingisimpairedinelderlymildcognitiveimpairmentpatientsbelievedtohavemedialtemporallobepathology(Negashetal.
,2004).
Consis-tentwiththis,functionalneuroimagingstudieshavereportedgreateractivationinmedialtemporallobestructures,includinghippocampusandparahippocampalareas,oncontextualcueingtrialswithrepeatedcomparedtonovelcongurations(Greene&Gross,2003;Preston,Saladis,&Gabrieli,2001).
Thus,thepresentndingssuggestthatdyslexiaisassociatedwithselectivedecitsinthefronto-striatal-cerebellarcircuitsthatunderliesequencelearning.
Thisisconsistentwithpreviousevidencethatoneormoreoftheseregionsareinvolvedinthefunctionalpathologyofdyslexia(Eckert,2004;Eden&Zefro,1998;Nicolson,Fawcett,&Dean,2001a).
ItalsocouldexplaintheresultsbyVicarietal.
(2005)whoshowedthatdyslexicchildrenwereimpairedontwodifferentimplicitlearningtasks,sequencelearningandmirrordrawing.
Thereisevidencethatthesetwotasksshareadependenceonfronto-striatalbraincircuits(seediscussionbelowandPoldrack&Gabrieli,2001).
Incontrast,ourndingsrevealthatdyslexicsarenotuniversallyimpairedonimplicitlearning,showingselec-tivedecitsonsome,butnotothertasksdependingonwhichbrainsystemsareengaged.
Thepresentdataarealsoconsistentwiththeargumentthatthemedialtemporallobesystemthatunderliesimplicitlearningin12J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxxthecontextualcueingtaskisintactindyslexia.
Infact,thepresentdatashowastrongtrendtowardenhancedlearningindyslexiccomparedtotypicalcollegestudentsonthistask.
Evidenceforthisoccurredinboththegroup(signicantposthoccomparisonoflearningonthenalsession)andindividualanalyses(signi-cantnegativecorrelationsbetweenindividuallearningandwordreadingscores).
Itispossiblethatthisreectscompensationforimpairedstriatal,cerebellarorfrontalfunctioninthesuccess-ful,dyslexiccollegestudentswetested.
Manyprevioushumanandanimalstudieshavesuggestedthatcompensationofthissortdevelopsinthefaceofsystem-specicpathology(Edenetal.
,2004;Poldrack&Packard,2003;Ullman,2004).
Insummary,thepresentndingsrevealthatdyslexicsareimpairedinhigher-orderimplicitsequencelearning,butsparedorevenenhancedintheimplicitlearningofspatialcontext.
Thisindicatesthatotherwisehigh-functioningcollege-studentdyslexicshavedifcultyintegratinginformationacrosstempo-rallynon-adjacentelements,butnotinthespatialconguralprocessingofinformationwithinasingledisplay.
Thisiscon-sistentwithpreviousbehavioralandneurologicalevidenceoffronto-striatal-cerebellarpathologyindyslexiaandsuggeststhatitisimportanttostudythedevelopmentofthesedifferentformsofimplicitlearningsoastodetermineifthispatternofsparedandimpairedabilityreectsdevelopmentalcompensation.
AcknowledgementsThisresearchwassupportedbygrantsfromtheNationalInstituteonAging(R37AG15450),NationalInstituteofChildHealthandHumanDevelopment(HD36461andHD40095)andbytheGeneralClinicalResearchCenterProgramoftheNationalCenterforResearchResources(MO1-RR13297),NationalInsti-tutesofHealth.
WethankEileenCahillandAliMussonifortheirhelpwithdatacollectionandanalysis.
ChandanVaidya,NancyDennis,BarbaraSchwartz,LynnFlowersandThomasZefrocontributedmanyhelpfulcomments.
WearegratefultoBonnieMcClellanandherstaffoftheCatholicUniversityDisabilitySupportServicesforassistanceinrecruitingparticipants.
Apre-liminaryreportofthesendingswaspresentedattheOctober2004meetingsoftheSocietyforNeuroscience.
ReferencesAlexander,A.
W.
,&Slinger-Constant,A.
M.
(2004).
Currentstatusoftreat-mentsfordyslexia:Criticalreview.
JournalofChildNeurology,19(10),744–758.
Bradley,L.
,&Bryant,P.
(1981).
Visualmemoryandphonologicalskillsinreadingandspellingbackwardness.
PsychologicalResearch,43(2),193–199.
Brunswick,N.
,McCrory,E.
,Price,C.
J.
,Frith,C.
D.
,&Frith,U.
(1999).
Explicitandimplicitprocessingofwordsandpseudowordsbyadultdevelopmentaldyslexics:Asearchforwernicke'swortschatzBrain,122,1901–1917.
Chun,M.
M.
,&Jiang,Y.
(1998).
Contextualcueing:Implicitlearningandmemoryofvisualcontextguidesspatialattention.
CognitivePsychology,36(1),28–71.
Chun,M.
M.
,&Phelps,E.
A.
(1999).
Memorydecitsforimplicitcontex-tualinformationinamnesicsubjectswithhippocampaldamage.
NatureNeuroscience,2(9),844–847.
Curran,T.
(1997a).
Effectsofagingonimplicitsequencelearning:Accountingforsequencestructureandexplicitknowledge.
Psycholog-icalResearch,60(1–2),24–41.
Curran,T.
(1997b).
Higher-orderassociativelearninginamnesia:Evidencefromtheserialreactiontimetask.
JournalofCognitiveNeuroscience,9(4),522–533.
Demb,J.
B.
,Boynton,G.
M.
,&Heeger,D.
J.
(1998).
Functionalmag-neticresonanceimagingofearlyvisualpathwaysindyslexia.
JournalofNeurosciences,18(17),6939–6951.
Denckla,M.
B.
,&Rudel,R.
G.
(1976).
Rapid"automatized"naming(R.
A.
N):Dyslexiadifferentiatedfromotherlearningdisabilities.
Neu-ropsychologia,14(4),471–479.
Dominey,P.
F.
,&Jeannerod,M.
(1997).
Contributionoffrontostriatalfunc-tiontosequencelearninginParkinson'sdisease:Evidencefordissociablesystems.
Neuroreport,8(5),iii–ix.
Eckert,M.
A.
(2004).
Neuroanatomicalmarkersfordyslexia:Areviewofdyslexiastructuralimagingstudies.
Neuroscientist,10(4),362–371.
Eckert,M.
A.
,&Leonard,C.
M.
(2000).
Structuralimagingindyslexia:Theplanumtemporale.
MentalRetardationandDevelopmentalDisabilitiesResearchReviews,6(3),198–206.
Eckert,M.
A.
,Leonard,C.
M.
,Richards,T.
L.
,Aylward,E.
H.
,Thomson,J.
,&Berninger,V.
W.
(2003).
Anatomicalcorrelatesofdyslexia:Frontalandcerebellarndings.
Brain,126(Pt2),482–494.
Eden,G.
F.
,Jones,K.
M.
,Cappell,K.
,Gareau,L.
,Wood,F.
B.
,Zefro,T.
A.
,etal.
(2004).
Neuralchangesfollowingremediationinadultdevelop-mentaldyslexia.
Neuron,44(3),411–422.
Eden,G.
F.
,Stein,J.
F.
,Wood,H.
M.
,&Wood,F.
B.
(1995).
Temporalandspatialprocessinginreadingdisabledandnormalchildren.
Cortex,31(3),451–468.
Eden,G.
F.
,&Zefro,T.
A.
(1998).
Neuralsystemsaffectedindevelop-mentaldyslexiarevealedbyfunctionalneuroimaging.
Neuron,21(2),279–282.
Ehri,L.
C.
(1999).
Phasesofdevelopmentinlearningtoreadwords.
InJ.
Oakhill&R.
Beard(Eds.
),Readingdevelopmentandtheteachingofreading:Apsychologicalperspective(pp.
79–108).
Oxford:BlackwellScienceLtd.
Felton,R.
H.
,Naylor,C.
E.
,&Wood,F.
B.
(1990).
Neuropsychologicalproleofadultdyslexics.
BrainLanguage,39(4),485–497.
Fletcher,P.
C.
,Zaris,O.
,Frith,C.
D.
,Honey,R.
A.
E.
,Corlett,P.
R.
,Zilles,K.
,etal.
(2004).
Onthebenetsofnottrying:Brainactivityandconnectivityreectingtheinteractionsofexplicitandimplicitsequencelearning.
CerebralCortex,bhh201.
Flowers,D.
L.
(1995).
Neuropsychologicalprolesofpersistentreadingdisabilityandreadingimprovement.
InC.
K.
L.
R.
M.
Joshi(Ed.
),Developmentalandacquireddyslexia(pp.
61–77).
Netherlands:KluwerAcademicPublishers.
Flowers,D.
L.
,Wood,F.
B.
,&Naylor,C.
E.
(1991).
Regionalcerebralbloodowcorrelatesoflanguageprocessesinreadingdisability.
ArchivesofNeurology,48,637–643.
Galaburda,A.
M.
,Sherman,G.
,Rosen,G.
D.
,Aboitiz,F.
,&Geschwind,N.
(1985).
Developmentaldyslexia:Fourconsecutivecaseswithcorticalanomalities.
AnnalsofNeurology,18,222–233.
Gombert,J.
E.
(2003).
Implicitandexplicitlearningtoread:Implicationasforsubtypesofdyslexia.
CurrentPsychologyLetters:Behavior,BrainandCognition,10(1),2003.
Gomez-Beldarrain,M.
,Garcia-Monco,J.
C.
,Rubio,B.
,&Pascual-Leone,A.
(1998).
Effectoffocalcerebellarlesionsonprocedurallearningintheserialreactiontimetask.
ExperimentalBrainResearch,120(1),25–30.
GomezBeldarrain,M.
,Grafman,J.
,Pascual-Leone,A.
,&Garcia-Monco,J.
C.
(1999).
Procedurallearningisimpairedinpatientswithprefrontallesions.
Neurology,52(9),1853–1860.
GomezBeldarrain,M.
,Grafman,J.
,RuizDeVelasco,I.
,Pascual-Leone,A.
,&Garcia-Monco,C.
(2002).
Prefrontallesionsimpairtheimplicitandexplicitlearningofsequencesonvisuomotortasks.
ExperimentalBrainResearch,142(4),578.
Grafton,S.
T.
,Hazeltine,E.
,&Ivry,I.
(1995).
Functionalmappingofsequencelearninginnormalhumans.
JournalofCognitiveNeuroscience,7,497–510.
J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxx13Greene,A.
J.
,&Gross,W.
L.
(2003).
Theroleofthehippocampus:Fmrirevealshippocampalactivationonacontext-dependentimplicittask.
PaperpresentedattheSocietyforNeuroscience.
Helmuth,L.
L.
,Mayr,U.
,&Daum,I.
(2000).
SequencelearninginParkin-son'sdisease:Acomparisonofspatial–attentionandnumber–responsesequences.
Neuropsychologia,38(11),1443–1451.
Howard,J.
H.
,Jr.
,&Howard,D.
V.
(1997).
Agedifferencesinimplicitlearningofhigherorderdependenciesinserialpatterns.
PsychologyandAging,12(4),634–656.
Howard,D.
V.
,&Howard,J.
H.
,Jr.
(2001).
Whenitdoeshurttotry:Adultagedifferencesintheeffectsofinstructionsonimplicitpatternlearning.
PsychonomicBulletin&Review,8(4),798–805.
Howard,J.
H.
,Jr.
,Howard,D.
V.
,Dennis,N.
A.
,Yankovich,H.
,&Vaidya,C.
J.
(2004).
Implicitspatialcontextuallearninginhealthyaging.
Neu-ropsychology,18(1),124–134.
Howard,D.
V.
,Howard,J.
H.
,Jr.
,Japikse,K.
,DiYanni,C.
,Thompson,A.
,&Somberg,R.
(2004).
Implicitsequencelearning:Effectsoflevelofstructure,adultage,andextendedpractice.
PsychologyandAging,19(1),79–92.
Jackson,G.
M.
,Jackson,S.
R.
,Harrison,J.
,Henderson,L.
,&Kennard,C.
(1995).
SerialreactiontimelearningandParkinson'sdisease:Evidenceforaprocedurallearningdecit.
Neuropsychologia,33(5),577–593.
Japikse,K.
,Howard,D.
V.
,&Howard,J.
H.
,Jr.
(2001).
Evaluationofadirectnonverbalmeasureofdeclarativesequenceknowledge.
JournalofCognitiveNeuroscienceSupplement,12,62.
Kelly,S.
W.
,Grifths,S.
,&Frith,U.
(2002).
Evidenceforimplicitsequencelearningindyslexia.
Dyslexia,8,43–52.
Korhonen,T.
T.
(1995).
Thepersistenceofrapidnamingproblemsinchil-drenwithreadingdisabilities:Anine-yearfollow-up.
JournalofLearningDisabilities,28(4),232–239.
Larsen,S.
C.
,&Hammill,D.
D.
(1986).
TheLarsen–Hammilltestofwrittenspelling.
Austin,TX:Pro-Ed.
Lyon,G.
R.
(1995).
Towardadenitionofdyslexia.
AnnalsofDyslexia,45,3–27.
Manis,F.
R.
,Savage,P.
L.
,Morrison,F.
J.
,Horn,C.
C.
,Howell,M.
J.
,Szeszulski,P.
A.
,etal.
(1987).
Pairedassociatelearninginreading-disabledchildren:Evidenceforarule-learningdeciency.
JournalofExperimentalChildPsychology,43(1),25–43.
Manns,J.
R.
,&Squire,L.
R.
(2001).
Perceptuallearning,awarenessandthehippocampus.
Hippocampus,11,776–782.
Negash,S.
,Petersen,L.
E.
,Geda,Y.
E.
,Knopman,D.
S.
,Boeve,B.
F.
,Smith,G.
E.
,etal.
(October2004).
Implicitsequencelearningandcontextualcueinginmildcognitiveimpairment.
PaperpresentedattheSocietyforNeuroscience.
Nicolson,R.
I.
,&Fawcett,A.
J.
(1999).
Developmentaldyslexia:Theroleofthecerebellum.
Dyslexia,5,155–177.
Nicolson,R.
I.
,Fawcett,A.
J.
,&Dean,P.
(2001a).
Developmentaldyslexia:Thecerebellardecithypothesis.
TrendsinNeurosciences,24(9),508–511.
Nicolson,R.
I.
,Fawcett,A.
J.
,&Dean,P.
(2001b).
Dyslexia,developmentandthecerebellum.
TrendsinNeurosciences,24(9),515–516.
Nissen,M.
J.
,&Bullemer,P.
(1987).
Attentionalrequirementsoflearning:Evidencefromperformancemeasures.
CognitivePsychology,19,1–32.
Oldeld,R.
C.
(1971).
Theassessmentandanalysisofhandedness:TheEdinburghinventory.
Neuropsychologia,9(1),97–113.
Poldrack,R.
A.
,&Gabrieli,J.
D.
(2001).
Characterizingtheneuralmech-anismsofskilllearningandrepetitionpriming:Evidencefrommirrorreading.
Brain,124(Pt1),67–82.
Poldrack,R.
A.
,&Packard,M.
G.
(2003).
Competitionamongmultiplememorysystems:Convergingevidencefromanimalandhumanbrainstudies.
Neuropsychologia,41(3),245–251.
Pothos,E.
M.
,&Kirk,J.
(2004).
Investigatinglearningdecitsassociatedwithdyslexia.
Dyslexia,10(1),61–76.
Preston,A.
R.
,Saladis,J.
,&Gabrieli,J.
D.
(2001).
Medialtemporallobeactivityduringimplicitcontextuallearning.
PaperpresentedattheSocietyforNeuroscience.
Prull,M.
W.
,Gabrieli,J.
D.
E.
,&Bunge,S.
A.
(2000).
Age-relatedchangesinmemory:Acognitiveneuroscienceperspective.
InF.
I.
M.
Craik&T.
A.
Salthouse(Eds.
),Thehandbookofagingandcognition(2nded.
).
Mahway,NJ:LawrenceErlbaumAssociates.
Ransby,M.
J.
,&Swanson,H.
L.
(2003).
Readingcomprehensionskillsofyoungadultswithchildhooddiagnosesofdyslexia.
JournalofLearningDisabilities,36(6),538–555.
Rauch,S.
L.
,Whalen,P.
J.
,Savage,C.
R.
,Curran,T.
,Kendrick,A.
,Brown,H.
D.
,etal.
(1997).
Striatalrecruitmentduringanimplicitsequencelearningtaskasmeasuredbyfunctionalmagneticresonanceimaging.
HumanBrainMapping,5(2),124–132.
Rayner,K.
,Foorman,B.
R.
,Perfetti,C.
A.
,Pesetsky,D.
,&Seidenberg,M.
S.
(2001).
Howpsychologicalscienceinformstheteachingofreading.
PsychologicalSciences,2(2Suppl.
),31–74.
Reber,A.
S.
(1989).
Implicitlearningandtacitknowledge.
JournalofExper-imentalPsychology:General,118(3),219–235.
Remillard,G.
,&Clark,J.
M.
(2001).
Implicitlearningofrst-,second-,andthird-ordertransitionprobabilities.
JournalofExperimentalPsychology:Learning,Memory,andCognition,27(2),483–498.
Robertson,E.
M.
,Tormos,J.
M.
,Maeda,F.
,&Pascual-Leone,A.
(2001).
Theroleofthedorsolateralprefrontalcortexduringsequencelearningisspecicforspatialinformation.
CerebralCortex,11(7),628–635.
Rosner,J.
,&Simon,D.
P.
(1971).
TestofAuditoryAnalysisSkill,TAAS.
JournalofLearningDisabilities,4(7),40–48.
Rumsey,J.
M.
,Andreason,P.
,Zametkin,A.
J.
,Aquino,T.
,King,A.
C.
,Hamburger,S.
D.
,etal.
(1992).
Failuretoactivatethelefttemporoparietalcortexindyslexia.
Anoxygen15positronemissiontomographicstudy.
ArchivesofNeurology,49(5),527–534[publishederratumappearsinArchivesofNeurology1994;51(March(3)):243].
Schendan,H.
E.
,Searl,M.
M.
,Melrose,R.
J.
,&Stern,C.
E.
(2003).
SequenceWhatsequenceThehumanmedialtemporallobeandsequencelearning.
MolecularPsychiatry,8(11),896–897.
Schvaneveldt,R.
W.
,&Gomez,R.
L.
(1998).
Attentionandprobabilisticsequencelearning.
PsychologicalResearch,61(3),175–190.
Seidler,R.
D.
,Purushotham,A.
,Kim,S.
G.
,Ugurbil,K.
,Willingham,D.
,&Ashe,J.
(2002).
Cerebellumactivationassociatedwithperformancechangebutnotmotorlearning.
Science,296(5575),2043–2046.
Shaywitz,S.
E.
,Fletcher,J.
M.
,Holahan,J.
M.
,Shneider,A.
E.
,Mar-chione,K.
E.
,Stuebing,K.
K.
,etal.
(1999).
Persistenceofdyslexia:TheConnecticutlongitudinalstudyatadolescence.
Pediatrics,104(6),1351–1359.
Shaywitz,S.
E.
,Shaywitz,B.
A.
,Rugh,K.
R.
,Fulbright,R.
K.
,Constable,R.
T.
,Mencl,W.
E.
,etal.
(1998).
Functionaldisruptionintheorganizationofthebrainforreadingindyslexia.
ProceedingsoftheNationalAcademyofSciencesofUnitedStatesofAmerica,95,2636–2641.
Simos,P.
G.
,Breier,J.
I.
,Fletcher,J.
M.
,Foorman,B.
R.
,Mouzaki,A.
,&Papanicolaou,A.
C.
(2001).
Age-relatedchangesinregionalbrainactivationduringphonologicaldecodingandprintedwordrecognition.
DevelopmentalNeuropsychology,19(2),191–210.
Sperling,A.
J.
,Lu,Z.
L.
,&Manis,F.
R.
(2004).
Slowerimplicitcategoricallearninginadultpoorreaders.
AnnalsofDyslexia,54(2),281–303.
Stein,J.
,&Walsh,V.
(1997).
Toseebutnottoread;themagnocellulartheoryofdyslexia.
TrendsinNeurosciences,20(4),147–152.
Stoodley,C.
J.
,Harrison,E.
P.
,Stein,J.
F.
(inpress).
Implicitmotorlearningdecitsindyslexicadults.
Neurosychologia,doi:10.
1016/j.
neuropsychologia.
2005.
07.
009.
Torgesen,J.
K.
,Alexander,A.
W.
,Wagner,R.
K.
,Rashotte,C.
A.
,Voeller,K.
K.
,&Conway,T.
(2001).
Intensiveremedialinstructionforchildrenwithseverereadingdisabilities:Immediateandlong-termoutcomesfromtwoinstructionalapproaches.
JournalofLearningDisabilities,34(1),33–58,78.
Turkeltaub,P.
E.
,Gareau,L.
,Flowers,D.
L.
,Zefro,T.
A.
,&Eden,G.
F.
(2003).
Developmentofneuralmechanismsforreading.
NatureNeuro-science,6(7),767–773.
Ullman,M.
T.
(2004).
Contributionsofmemorycircuitstolanguage:Thedeclarative/proceduralmodel.
Cognition,92(1–2),231–270.
Vellutino,F.
R.
,Fletcher,J.
M.
,Snowling,M.
J.
,&Scanlon,D.
M.
(2004).
Specicreadingdisability(dyslexia):WhathavewelearnedinthepastfourdecadesJournalofChildPsychologyandPsychiatry,45(1),2–40.
14J.
H.
HowardJr.
etal.
/Neuropsychologiaxxx(2005)xxx–xxxVicari,S.
,Finzi,A.
,Menghini,D.
,Marotta,L.
,Baldi,S.
,&Petrosini,L.
(2005).
DochildrenwithdevelopmentaldyslexiahaveanimplicitlearningdecitJournalofNeurology,Neurosurgery,andPsychiatry,76,1392–1397.
Vicari,S.
,Marotta,L.
,Menghini,D.
,Molinari,M.
,&Petrosini,L.
(2003).
Implicitlearningdecitinchildrenwithdevelopmentaldyslexia.
Neu-ropsychologia,41(1),108–114.
Waber,D.
P.
,Marcus,D.
J.
,Forbes,P.
W.
,Bellinger,D.
C.
,Weiler,M.
D.
,Sorensen,L.
G.
,etal.
(2003).
Motorsequencelearningandreadingability:IspoorreadingassociatedwithsequencingdecitsJournalofExperimentalChildPsychology,84(4),338–354.
Wechsler,D.
(1997).
WechslerAdultIntelligenceScale(3rded.
).
SanAntonio,TX:PsychologicalCorporation.
Wechsler,D.
(1999).
WechslerAbbreviatedScaleofIntelligence.
SanAntonio,TX:ThePsychologicalCorporation.
Willingham,D.
B.
,Koroshetz,W.
J.
,&Peterson,E.
W.
(1996).
Motorskillshavediverseneuralbases:SparedandimpairedskillacquisitioninHunt-ington'sdisease.
Neuropsychology,10(3),315–321.
Wolf,M.
(1986).
Rapidalternatingstimulusnaminginthedevelopmentaldyslexias.
BrainLanguage,27(2),360–379.
Wolf,M.
,Miller,L.
,&Donnelly,K.
(2000).
Retrieval,automaticity,vocabu-laryelaboration,orthography(RAVE-O):Acomprehensive,uency-basedreadinginterventionprogram.
JournalofLearningDisabilities,33(4),375–386.
Wolff,U.
,&Lundberg,I.
(2002).
Theprevalenceofdyslexiaamongartstudents.
Dyslexia,8(1),34–42.
Woodcock,R.
W.
,&Johnson,M.
B.
(1990).
Woodcock–JohnsonPsycho-EducationalBattery—revised(WJ-r).
Allen,TX:DLMTeachingResources.