潜意识暗示提高人机交互能力Word格式文档下载.docx
《潜意识暗示提高人机交互能力Word格式文档下载.docx》由会员分享,可在线阅读,更多相关《潜意识暗示提高人机交互能力Word格式文档下载.docx(20页珍藏版)》请在冰豆网上搜索。
EnhancedLearning
PierreChalfounandClaudeFrasson
Dé
partementd'
InformatiqueetdeRechercheOpé
rationnelle,Université
deMontré
al,Office2194,Montré
al,Qué
bec,H3T1J4,Canada
Received2June2010;
Accepted25September2010
AcademicEditor:
KennethRevett
Copyright©
2011PierreChalfounandClaudeFrasson.ThisisanopenaccessarticledistributedundertheCreativeCommonsAttributionLicense,whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
Abstract
Thispaperpresentsresultsfromanempiricalstudyconductedwithasubliminalteachingtechniqueaimedatenhancinglearner'
sperformanceinIntelligentSystemsthroughtheuseofphysiologicalsensors.Thistechniqueusescarefullydesignedsubliminalcues(positive)andmiscues(negative)andprojectsthemunderthelearner'
sperceptualvisualthreshold.Apositivecue,calledanswercue,isahintaimingtoenhancethelearner'
sinductivereasoningabilitiesandprojectedinawaytohelpthemfigureoutthesolutionfasterbutmoreimportantlybetter.Anegativecue,calledmiscue,isalsousedandaimsatobviouslyattheopposite(distractthelearnerorleadthemtothewrongconclusion).Thelatestobtainedresultsshowedthatonlysubliminalcues,notmiscues,couldsignificantlyincreaselearnerperformanceandintuitioninalogic-basedproblem-solvingtask.Nonintrusivephysiologicalsensors(EEGforrecordingbrainwaves,bloodvolumepressuretocomputeheartrateandskinresponsetorecordskinconductivity)wereusedtorecordaffectiveandcerebralresponsesthroughouttheexperiment.Thedescriptiveanalysis,combinedwiththephysiologicaldata,providescompellingevidenceforthepositiveimpactofanswercuesonreasoningandintuitivedecisionmakinginalogic-basedproblem-solvingparadigm.
1.Introduction
TheuseoftechnologytobuildIntelligentUserInterfaces(IUI)hasrevolutionizedthewaycomputersinteractwithhumanbeings.ExamplesoftheseIUIscanbefoundinvirtuallyeveryaspectofourlives,suchasmobilemediasharing[1,2],intelligentvehicularassistance[3],andmentalspellingmachines[4].OnegrowingareaofresearchwithintheHCIcommunityrecentlyhasbeenfocusingtowardsthedesignofsmartinterfacesforlearning.Indeed,asmartinterfaceshouldbeable,inalearningcontext,todetecttheuser’semotionalandcognitivestatesinordertoadjustandadapttheteachingmaterialaccordingly.SuchadaptivesystemsrelyingonefficientIUIsareknownasintelligenttutoringsystemsandarecomprisedofmultiplegoal-specificmodulestoaidthelearner.Oneofthesemodules,calledthetutor,makesuseofcognitivepedagogicalstrategiesandaffectivestatestoproperlymodelhumancognitionaccordingtothelearner’sconstantneedsandevolution.ThetutortakesadvantageofrecentdevelopmentinIUItoefficientlycommunicatewiththelearner.Thismultidisciplinaryfieldofresearchtriestoaccomplishthisdauntinggoalofuser-modelingandadaptationbyimplementingthemostrecentevolutionsandadvancesinvariousresearchareassuchasartificialintelligence,neuroscience,andcognitivescience.Oneofthoserecentadvancesthispastdecadehasbeentheshiftinemphasisforcognitivesciencefromlearner’sperformancetolearner’sexperienceduringlearning.Indeed,affectivestates,motivation,andknowledgeconstructionhavebeenextensivelymeasuredandexplored[5–8]andhaveshownthatemotionsareanimportantcomponentinskillacquisitionandplayapivotalroleinlearning.Indeed,researchesinneurosciencesandcognitivepsychologyhaveshownthatemotionsarewidelyrelatedtodiversecognitiveprocesses,suchasattention,problemsolving,anddecision-making[9,10].Emotionsinfluenceourbehaviorandplayanimportantroleinourevery-daydecision-makingprocesses[11].Cognitiveactivityisalsofundamentallyrelatedtoemotions[12].Cognitiveprocesssuchasproblemsolvinganddecisionmakingnotonlydependbutaregreatlyintertwinedwiththeindividual’semotionalstate[13].Moreover,emotionsareessentialactorsforcreativethinking,inspiration,aswellasconcentrationandmotivation[10,14].ItthenbecomesvitalforanHCIsystemtodetectandrecognizetheseemotionalandcognitivestates,viaphysiologicalsensorsorotherwise,andrelaythemtothepertinentITSmodules.Hence,learningsystemswouldbeabletointelligentlyadapttheircommunicationandinteractionwithlearnersthroughadaptiveHCIsystems.
Nevertheless,amajorcomponentoflearninganddecisionmakingwhensolvingproblemshasbeenmostlyneglectedinthisresearchfield:
humanunconsciouscognition.Indeed,thecognitiveunconsciousasamajorplayeristheintegrationandinterpretationofcomplexmaterialwithregardstodecisionmakingandpossiblylearning.Tothatend,alargebodyofworkinneuroscienceandotherfieldshasputforthcompellingevidencethatlearningsimple-to-complexinformationcanbedonewithoutperceptionorcompleteawarenesstothetaskathand[15–18].Theideathatanyinformationprojectedtoofasttobeconsciouslyperceivedbyalearner(calledsubliminalprojection)hasbeenthefocusofmuchresearchinneuroscienceandcognitivepsychology.AvarietyofIUIshavebeendesignedforsuchapurpose,rangingfromsimpletwo-dimensionalinterfaces(2D)tocomplex3Dimmersiveones[19–23].Furthermore,therecordingoftheneuralelectricalactivity,calledEEGactivity,presentinthebrainduringtheoccurrenceofthesementalprocessesisalsoanactiveresearchareainHCI.Indeed,anIUIthatdetectsbrainactivitycanrecognizeandquantifytheuser’sengagementlevelinspecificactivities[24].However,intheHCIcommunity,EEGactivityismainlyusedtoconstructbraincomputerinterfacesmainlyaimedatcharacterspellingoritemrecognition[4,18,25–27].
Unfortunately,veryscarceresearchintheHCIcommunityhasemployedaffect,subliminalprimingandEEGforeducationalpurposes.Indeed,theexistenceofperceptuallearningwithoutperceptionhasnotonlybeenproven,butalsoreplicatedinonestudywedidtwoyearsago[19].Inthisstudy,wepresentedanovelsubliminalprimingtechniquebuiltintoanadaptiveIUIaimedatboostingperformancebyenhancingthelearner’sdeductivereasoningcapabilities.Ourfindingsconcurwiththeliterature:
subliminalstimuli,whichisstimuliprojectedbelowthethresholdofconsciousperception,canenhancelearningandincreasethedeductivereasoningoflearnersonlyifcarefullyandintelligentlyconstructedbeforebeingintegratedintoanIUI.However,thisinitialstudysufferedfromamajorlimitation:
theexclusiveuseofpositivesubliminalpriming(cuesdesignedtohelplearning,nothinderorinterferewithit).Thus,wedidnotknowwhateffectmightnegativecues,ormiscues,haveonlearningandperformancewhenintegratedintoourIUIsystem.
Inthepresentpaper,weintendtofollowuponourrecentstudybydesigninganevolvedversionofouradaptiveIUIwithournovelsubliminallearningtechniqueaimedatenhancingthelearner’sinductivelearningcapabilities.ThisnewIUIimplementedinourtutoringsystemwillmarrypositiveandnegativeprimesaswellasaffectivesensorsinthestringent2Denvironmentresemblingonlinetests.Affectivesensorswillserveasimportantindicatorofthelearner’semotionalandcerebralstatewhenfacedwiththedecisionofcorrectlyansweringaquestion.Whenfacingaquestion,learnerscaneither(a)guesstheansweror(b)correctlyinferthesolution(sincethisisanexamandthesolutionisnotpresentedtothem).WebelievethatIUIswouldbeabletodetectbothoutcomes((a)and(b))byrecordingandanalyzingemotionalandcerebraldata.Wealsobelieve,basedontheextensiveliteraturetothateffect,thatsubliminalprimingshouldinfluencebothoutcomes.WethusintendtoinvestigatetherelevanceofaugmentingourIUIwithcerebralsensorsandusingoursubliminalteachingtechniquebystatingtwoHCI-researchquestions.First,doesintegratingdifferentsubliminalcuingtypes(positive,control,andmiscue)intothisnewlydesignedIUIenhanceorhamperthelearner’sinductivereasoningcapabilitiesandperformance?
Second,whatsignificantphysiologicalimpactscanthisnewlydesignedinterfacewithoursubliminalteachingtechniquehaveonlearner’sperformanceduringquestionansweringaswellasontheentirelearningprocess?
Theorganizationofthispaperisasfollows:
inthefirstsection,wepresentpreviousworkrelatedtovariousaspectsofourresearch.ThesecondsectionwilldiscusspriminginHCI,morespecificallyfromaneducationalstandpoint.Thethirdsectionwilllaythebasesofphysiologicalrecordingsandimportancetoeducationandlearningingeneral.Thefourthsectiondescribestheexperimentsetupanddepictsthevariousaspectsrelatedtosubliminalcuesandmiscues.Thefifthsectionpresentstheobtainedresultswhicharediscussedinsectionsixleadingtothelastsectionwhereweconcludeandpresentfuturework.
2.RelatedWork
Thepresentworkemployssubliminalpriming,affect,andEEGtoinvestigatethepossiblecontributionofprimingtoenhancelearningandmorespecificallyinductivereasoning.Tothebestofourknowledge,thereareonlytwosimilarpapersintheITS/AIED(Artificialintelligenceineducation)/HCIcommunitythatemploysubliminalprimingandEEGinaneducationalcontext[28,29].Thefirstisourownworkwherewepresentedcerebralevidence,namely,aP300component(seebackgroundonEEGandaffectivedatabelow),toconfirmthatbothcuesandmiscueswereinterpretedinthebrain,butwithslightl