DISS.ETHNO.21214BistablePerceptionoftheNeckerCube - PDF document

AbstractTheNeckercubeisabistablestimuluswithaverylongresearchhistory,spanningmorethanacentury.Veryearly,itstemporaldynamicsanditsstimuluspropertieswereexamined.Itwasfoundthatthenumberofre-versalswithinacertaintimeintervalwerereproducibletoahighaccuracywithinoneobserverbutcouldvarysignicantlybetweendierentobserv-ers.Alreadyearlyinthersthalfofthe20thcentury,attemptsatlinkingbistableperceptionoftheNeckercubewithpersonalityweremade.Eventhoughmuchprogressinthedescriptionofthereversaldynamicshasbeenmadesince,acomprehensiveunderstandingofinter-individualdierencesinbistableperceptionintermspersonalitytraitsandcognitiveprocessesisstilllackingtoday.TwostudiesonneutralandvoluntarilycontrolledperceptionoftheNeckercubewereconducted.Thetemporaldynamicsanditsdependenceonstim-ulusparametersaswellasitsrelationtopersonalitytraits,mindfulness,temporalprocessing,workingmemory,generalreactiontimes,attentionandperceptionofanacousticbistablestimuluswereexplored.Newresultsoninitialadaptation,goodnessoftandstationaritywithre-specttocubesizewerefound.Aquantitativeanalysisofaperceptualbiaseectwasgivenintermsofdwelltimedistributions.IndividualdierencesinvoluntarycontroloverperceptionoftheNeckercubewerefoundtoberelatedtopersonalitytraitsandmindfulness.Severalpersonalitytraitsnotrelatedtobistableperceptionandsomerelatedtoitsneutralperceptionwereidenti-ed.Furthermore,evidenceforthepresenceoftwomechanismsoftemporalprocessing,namelyprocessingspeedandtemporalintegration,inbistableperceptionwasdiscovered.SimilaritiesanddierencesbetweenperceptionoftheNeckercubeandareversiblewordstimuluswerereported.Finally,individualdierencesinworkingmemorycapacityseemlikelynottorelatetobistableperception.Inconclusion,animproveddescriptionofthetemporaldynamicsofbistableperceptionandsomelow-levelmodulatingfactorswasgiven.Furthermore,inter-individualdierencesinthedwelltimedistributionwereshowntobereectedinseveralpersonalitytraitsandcognitiveprocesses,inparticulartimeprocessing.Thisdemonstratesthatvariationsinbistableperceptionbe-tweenindividualscanindeedbebetterunderstoodandclassiedbylinkingthemtoothercharacteristicsincognitionandpersonality.2 entreindividusdanslaperceptionbistableeneetpeuventêtreconçuesetclassiéesmieuxsiellessontassociéesàdesautrestraitscharactéristiquescognitifsetdepersonnalité.4 Contents1Introduction101.1WhyLookatBistability?.....................101.2ClassicationofMultistability..................111.2.1AmbiguousFigures....................121.2.2BinocularRivalry.....................161.2.3MonocularRivalry....................171.2.4Structure-from-Motion..................181.2.5ApparentMotionQuartets................191.2.6Motion-inducedBlindness................191.2.7Non-visualMultistability.................191.3ThePsychophysicsofVisualBistability.............201.3.1MeasuringBistablePerception..............211.3.2ViewingParametersfortheNeckerCube........221.3.3ReproducibilityofDwellTimes.............241.4ThePhysiologyofVisualBistability...............251.4.1EyeMovements&Blinks.................251.4.2Neuro-Imaging......................261.4.3Lesions...........................291.5Genetics..............................301.6ThePsychologyofVisualBistability..............311.7SimilaritiesandDierences....................332ModelsofBistablePerception352.1UporDown?...........................352.2OscillatorsorAttractors?.....................382.3FurtherApproaches........................406 6BiasEect766.1QualitativeReports........................766.2QuantifyingthePerceptualBias.................776.3Results...............................786.4SeeingtheCubeFromAbove..................787VoluntaryInuence827.1VolitioninBistabilityandPsychology..............827.2MeasuringVolition........................857.3Results...............................887.4Discussion.............................928Perception&Personality978.1StudiesLinkingBistabilityandPersonality...........978.2OperationalisationofPersonalityTraits.............998.3Results...............................998.4Discussion.............................1009Mindfulness&Perception1049.1MindfulnessinScienceandPerception.............1049.2Methods..............................1069.3Results...............................1079.4MindfulnessRelatestoPerceptualVolition...........10810TemporalProcessing11210.1TimePerception,ReactionandAttention............11210.2ExploringLinksinTimeScales.................11310.3Results...............................11610.4Discussion.............................11611(Un-)RelatedProcesses12111.1TheVerbalTransformationEect................12111.1.1AcousticMultistablePerception.............12111.1.2Methods..........................12211.1.3Results...........................12311.1.4Discussion.........................12511.2WorkingMemory.........................12711.2.1MemoryinBistablePerception............1278 1.Introduction1.1WhyLookatBistability?...orhear,feelorsmellit,forthatmatter?Bistableperceptionseemstohavefascinatedhumansforalongtimeasnu-merousoccurrencesofambiguousornamentsorpiecesofartillustrate.Manychurches,forexample,showornamentsliketheoneonthetitlepage,whichcanfoundonaninteriorwallofthechurchoftheSwissmonastryKlosterKappel.1Itswhiteareascanbeseenaseitherthetopfacesofotherwiseblackcubesorasthebottomfaceofsuchcubes.Iftheattendantofthemasskeptlookingatitforawhileheorshewouldexperiencechangesfromonetotheotherandback.Onlylately,inthehistoriccontextofitsexistence,hasbistableperceptionbecometheobjectofextensive,rigorousscienticstudy.There,itprovidesauniqueopportunity.Thepeculiarityofmanymultistablestimuliisthatovertimetheyelicitsev-eraldistinctconsciousimpressions,orpercepts,intheobserver,whiletheactualstimulusremainscompletelyunchanged.Anyonewhoisinterestedintheunderstandingofconsciousnessshouldbefascinatedbythisimmediatelyandinfact,philosophersare,aprominentexamplebeingLudwigWittgen-stein.Studyingmultistableperception,onemarvelsatthefactthattherearetwoormoredierentmentalstates,whichtheobservercanvividlyexperienceandwhichareelicitedbythesameexternalstimulus.Hence,theconsciousinnerexperience,thequalia,isclearlyanddistinctlymodulatedovertimewhileexternallynothingseemstochange.Whatmakesthisevenmorein-terestingforempiricallyinclinedresearcher,isthatitisquantiable.Thetimesbetweenperceptualchangescanbemeasuredviaself-report.Thus, 1IamindebtedtoRichardDählerfortakingmeonacyclingtourtothisbeautifulplaceonaveryhotsummerdayin2012.10 thenext.I.e.dwelltimesindicatehowlongoneperceptisseenatatime,beforeperceptionchangestotheotherpercept.Afewothertermsareusedsynonymouslytodwelltimesintheliterature,e.g.stabilitydurations,re-versaltimes,switchingtimes,perceptualdurations.Thereareseveraldierentcategoriesofmultistablestimuli.Manymultistablevisualstimuliareinfactbistable,i.e.thereareonlytwopossiblepercepts,atleastapproximately.Inthefollowing,anoverviewoverthemostimportantclassesofvisualbistablestimuliwillbegiven.Also,afewexamplesofbi-andmultistablestimuliinothersensualmodalitieswillbepresented.Itshouldbenotedthattherearedierentnomenclaturesfortheclassic-ationofmultistablestimuli.Thus,termsmightvaryfrompublicationtopublicationandalternativeexpressionsaregiveninthefollowingwhenevercommon.1.2.1AmbiguousFiguresManydierentambiguousguresgainedpublicattentioninthelasttwocenturies.Someofthosehavebeenstudiedextensivelyinscienticresearch.Ambiguousguresareonesubclassofmultistablevisualstimulicharacter-isedbytheirabilitytoelicittwoormoremutuallyexclusiveperceptsinanobserverwhilethegureitselfstaysconstant.Bistableambiguousguresareaspecialcaseofambiguousgures,havingonlytworivalingpercepts.Incontrasttobinocularrivalry,botheyesviewthesamestimulus.Thistypeofbistablestimulusisalsocalledperceptualrivalrysometimes.Note,thatmulti-andbistabilityiscontingentontwofactors.Therstisknowledgeofreversibility.RockandMitchener(1992)showedforseveralcommonbistableguresthatreversalsarelargelyabsentiftheobserverisunawareofoneofthepossibleinterpretationsofthestimulus.Thesecondone,whichconcernsinparticularbistableperception,istheneglectofpos-siblebutinfrequentorimprobableperceptualalternatives.Aperspectivereversinggure,liketheNeckercubeortheMachbook,forexample,cannotonlybeseeninitstwothree-dimensionalperspectives,butcanalsobeperceivedasanabstracttwo-dimensionaldrawing.Moststudiesdonottaketheseconsiderationsintoaccountinordertoexplorecertainaspectsofthestimuliinisolation,namelythereversalsbehaviourbetweenthemostdom-inantinterpretationsofthegure.Thisapproachwillbefollowedhere,too.Someexamplesofdierentambiguousgureswillbepresentedinthefollow-ing.12 (a)Neckercube (b)Schröderstaircase (c)Machbook (d)Vase/faces (e)Canadianag (f)Batman&theJokerFigure1.1:Dierentambiguousguresofthecategoriesperspectivereversalandgure-groundreversal:(a)theNeckercube(owndrawing),(b)theSchröderstaircase(repro-ducedfromStrüberandStadler(1999)),(c)Machbook(owndrawing),(d)vase/faces(drawingbyBryanDerksen),(e)Canadianag,(f)Batman&theJoker(papercutbyeyez2theskiezonickr,2007).14 (a)Duck/rabbit (b)Wife/motherinlawFigure1.2:Dierentambiguousguresofthecontentreversalcategory:(a)Duckandrabbit(publishedinFliegendeBlätter,October23,1892,München),(b)wife/motherinlaw(advertisementbytheAnchorBuggycompany,1890).Interestingly,alsopartsoftheCanadianagfeaturesomegure-groundambiguity.Theupperboundariesofthemapleleavecanbeseenastheprolesoftwopeoplearguing,clashingattheirforeheads,thelowerpartofthemapleleaveoutliningtheirshoulders(Fig.1.1e).AheroofpopularcultureandhisenemyarepicturedintheBatman&theJokerpapercutpresentedinFig.1.1f.Also,theartistM.C.Eschermadeuseofgure-groundperceptioninafascinatingwayinseveralofhisdrawings.ContentReversalStimuliIntheclassicationemployedhere,contentreversalstimulishalldenoteg-ures,forwhichthereversalsareduetoitscontentandnotduetoperspectiveorgure-ground.Afamousrepresentativeofthisclassistheduck/rabbitgurerstpub-lishedintheGermanhumormagazine,FliegendeBlätter3(Fig.1.2a).Itfeatureseitheraduckwiththebeakfacingtotheupperleftorarabbitwithitsmouthontherighthandsideoftheimage.Later,theAmericanpsycho- 3FliegendeBlätter,p.17,October23,1892,München15 (a)Twooverlaidgratings (b)House/faceFigure1.4:Twomonocularrivalrystimuli:(a)MonocularrivalrystimuluscreatedbyAlexanderMaierconsistingoftwosuperimposedsinewavegratings.(b)MonocularrivalrystimulusreproducedfromO'Sheaetal.(2009).Theauthorsgavethefollowingviewinginstructions:Toexperiencemonocularrivalrystareapproximatelyatthecentreoftheimage,sayatthebridgeoftheface'sglasses.Bepatient!Monocularrivalrytakesawhiletodevelop.Butafteratime,10-30orso,youwillnoticeuctuationsintherelativeclarityofthetwoimages.Youmayevenseeoneofthetwoimagesbecomeexclusivelyvisiblebriey,alongwithbriefcompositesinwhichdierentpartsoftheimageappearindierentpartsofthevisualeld.LeopoldandLogothetis(1999)reviewedevidencethatbinocularrivalryiscloselyrelatedtootherformsofmultistableperceptionasitisnotcharacter-isedbyspecialisedinterocularinhibitoryprocessesbutratherbycompetitionbetweencentralstimulusrepresentations.Thisndingissupportedbystrongsimilaritiesinthedwelltimedistributionsofbinocularrivalryandambiguousgures(Brascampetal.,2005).Buttherearealsoconsiderabledierencestotheperceptionofambiguousgures,liketheamountofvoluntarycontrolthatcanbeexercised(MengandTong,2004)ortheroleofeyemovementsininstigatingperceptualreversals(vanDamandvanEe,2006).Similaritiesanddierencesbetweenthesetwotypesofstimuliwillbedis-cussedinmoredetailinSec.1.7.1.2.3MonocularRivalryInmonocularrivalry,twovisualstimuliaresuperimposedandviewedwithbotheyes.AnexampleofsuchastimulusisgiveninFig.1.4.O'Sheaetal.(2009)comparedmonocularrivalrytobinocularrivalryandfoundseveralcommonalities.Inparticular,theauthorsreportedgammadistributeddwell17 withthepatternrhythm.Formostobserversthehightonewasalwaysheardinoneearwhileitwaspossibleunderprolongedlistening,thatthispreferenceswitchedsponteously(Deutsch,1975).Thisresultindicatesabinauralrivalryprocesssimilartobinocularrivalryinthevisualmodality.Furthermore,thereistheverbaltransformationeect,achangeintheperceivedwordwhentherecordingofasuitablewordisplayedinaloop.ItwasrstdescribedbyWarrenandGregory(1958)forwordslikesayandstress.Radilovaetal.(1990)reportedsomequantitativendingsfordierentwordsanddrewacomparisontobistablevisualperception.InSec.11.1adetailedaccountofthetemporaldynamicsoftheverbaltransformationeectanditsrelationtoperceptionoftheNeckercubewillbegiven.HapticCarteretal.(2008)describedthetactilecomplementtothevisualapparentmotionquartet(seeabove).Inthissetup,participantswerestimulatedonasmallsquareareaoftheirngertips.Whenstimuliperiodicallyalternatedbetweentheopposingendpointsofthetwodiagonalsofthesquare,parti-cipantsreportedswitchesbetweenperceivedleft-rightandup-downmotion.OlfactoryAnolfactoryanaloguetobinocularrivalrywasreportedbyZhouandChen(2009).Inthisstudy,participantswerepresentedwithtwodierentsmellsforeachnostrilandtheyreportedperceivingonesmellatatime,switchingspontaneouslybetweenthem.1.3ThePsychophysicsofVisualBistabilityAfterthisoverviewovermultistabilityindierentmodalities,thefocuswillnowbedirectedonvisualbistability,inparticularontheNeckercube.Itconstitutesthemostimportantstimulusforthisthesisasalltheexperimentspresentedlateremployedthisambiguousgure.Thus,inthefollowing,amoredetaileddescriptionofthepsychophysicsofbistableperceptionoftheNeckercubewillbegiven,includingthemeasure-mentofdwelltimesandanoutlineofkeyparameters.20 astimulusliketheNeckercube,itsunambiguousversionsareverydierentformtheambiguousone,sothatthetrainingeectmightnotdirectlyrelatetotheactualexperiment.Someworkhasbeendonetowardstherecognitionofperceptualreversalswiththeuseofphysiologicalmeasures.Forseveraltypesofbistablestim-uli,Einhäuseretal.(2008)foundthatpupildilationprecededperceptualreversals.Similarly,fordiscontinuouspresentationoftheNeckercube,brainactivityintherightinferiorparietalcortexhasbeenidentiedasaprecursorofperceptualreversals(Britzetal.,2009).Inbothinstances,thesechangescanonlybedetectedinaveragesovermanytrials.Hence,theyarenotsuitedasmarkersforasingleperceptualreversal.1.3.2ViewingParametersfortheNeckerCubeTheresearchonbistableperceptioninthesecondhalfofthelastcenturywasstronglyinuencedbytheclassicationofexplanationsintermsoftop-downandbottom-up(cf.LongandToppino(2004)forareview).Thetop-downapproachassumesactive,volitionalprocessesnearperceptualawarenessasresponsibleforgurereversals,whileinthebottom-upapproachpassive,automaticandlocallyadaptablemechanismsduringearlyvisualpro-cessingcreatethereversals(Kornmeieretal.,2009).Psychophysicalndingsonstimulusparameterswerecitedmainlyinfavourofthelattermodel.Thus,inthefollowing,somestimulusparametersthatplayaroleintheperceptionoftheNeckercubewillbediscussed.Evidencefortop-downinuencesaswellastheneedforanintegrationofbothperspectivesofthedebatewillbegiveninSec.2.1.SizeTherelationofthesizeoftheNeckercubetoitsperception,inparticularthetemporaldynamicsintermsofthedwelltimesanditsstatisticaldistribution,wasstudiedbyseveralresearchgroups(BergumandFlamm,1975;Borsellinoetal.,1982;DuggerandCourson,1968;Toppino,2003;Washburnetal.,1931).Theresultsofallgroupsindicatethatdwelltimesarelongerthelargerthecubeis.ThisisanimportantndingregardingthecomparabilityofdierentstudiesontheNeckercube.AmoredetailedoverviewoverthestudiescitedaboveandownresultswillbegiveninSec.5.1.22 parameters,apartfromillumination,isindicativeofbottom-upcomponentsofbistableperception,astheymainlyinvolveearlyprocessesofvisualper-ception.Thebottom-upvs.top-downclassicationofbistableperceptionaswellasthemodelofadaptationwillbepresentedindetailinSec.2.1.There,alsotheotherclassofevidence,namelythatforthetop-downinuence,willbeoutlined.1.3.3ReproducibilityofDwellTimesAnimportantaspectofthetemporaldynamicsofbistableperceptionisthereproducibilityofdwelltimes.Whethermeandwelltimes,orthenumberofreversals,couldbereproducedwithinonepersonwasaskedalreadyveryearlyintheresearchhistoryofbistableperception,inordertondouthowmuchoftrait-characterthismeasurehad.Ifthenumberofreversalswithinagivenamountoftimewereindeedmoreorlessstablewithinaperson,thestudyofpotentialrelationstootherratherstablecharacteristics,likeperson-alitytraits,wouldbesensible.GuilfordandHunt(1931)seemtohaveconductedtherstrigoroustestofthehypothesisthatreversalratesareinfactreproduciblewithininonepersontoahighaccuracy.Fiveindividualsreportedperceptualreversalsforthreeminutes,atthreetimesofthedayforsixdays.Theauthorsfoundthatbothwithinonedayandbetweendays,thenumberofreversalswerenotsignic-antlydierent.Themoststablereproductionwiththeleastamountoftestingwasfoundfortakingtheaverageoverthreedays.AlsoFrederiksenandGuil-ford(1934),whosestudyhasbeencitedoftenwithregardtothequestionofreproducibility,reportedhighcorrelationsbetweennumberofreversalsoftheNeckercubeforsubsequentdays.GuilfordandHunt(1931)furthermorefoundthatthevariationsofthemeannumberofreversalswithinoneindi-vidualovertimeweremuchsmallerthanthevariationofmeannumberofreversalswithinagroupofobservers.Thispropertyledtotheclassicationofobserversintofastandslowreversers(asinBorsellinoetal.(1972)).ThereseemnottobemorerecentstudiesthanthoseofGuilfordandco-workersexaminingthereproducibilityofthenumberofreversalsfortheNeckercube.Hence,apilotstudydescribedinSec.4.2wasconductedthatconrmedtheabovendings,alsoshowingnosignicantdierencesinthedwelltimedistributionsonconsecutivedays.Itmustbetakenintoaccountthatonlyasmallnumberofparticipantsweretestedinthatpilotstudy,sothatfurthercorroborationoftheseresultswithalargersamplewouldbe24 tovoluntarilycontrolperception.Furthermore,whilethexationpositiondidnotdeterminethepercept,theauthorsfoundthatobserverspreferredtolookatdierentpositionswheninstructedtoholddierentperceptsofthestimulus.Also,vanDamandvanEe(2005)examinedtheroleofeyemovementsandblinksforslantrivalry,anambiguousgurewithtemporaldynamicscomparabletotheNeckercube.Theauthorsfoundthattherewasnopositivecorrelationbetweenreversalsandeithersaccadesorblinksoccur-ringbeforethesereversals.Inconclusion,itcanbestatedthatforperceptionoftheNeckercube,blink-ingandsaccadesdonotdetermineperceptualreversals.vanDamandvanEe(2006)alsoshowed,thatthiswasdierentforbinocularrivalry.Forthatstimulustype,thereisinfactacorrelationbetweensaccadesandreversalsataboutthemomentofreversal.1.4.2Neuro-ImagingApartfromthepsychophysicalandouterphysiologicaldescriptions,itisde-sirabletogainaneuro-physiologicalunderstandingofperceptualbistability.Withtheadvanceddevelopmentofneuralimagingtechniquesitispossibletostudythebrainregionsandneuralstructuresinvolvedinbistableperception,andinparticularthoseinvolvedinperceptualreversals.EEGElectroencephalography(EEG)hasbeensuccessfullyappliedinordertostudyperceptualreversalsofambiguousgures.AgoodreviewhasbeengivenbyKornmeierandBach(2012).Comparedtofunctionalmagneticres-onanceimaging(fMRI),EEGprovidestheadvantageofaveryhightemporalresolutionintherangeofafewmiliseconds.Thisisofcourse,averyusefulfeatureinordertostudyatemporalprocesslikeperceptualreversals.AsthevoltagesmeasuredwithEEGacrosstheskullareusuallyverysmall,manytrialshavetobeaveragedinordertoachieveanacceptablesignal-to-noiseratio.Forthis,acommontemporalreferenceinallthesetrialsisnecessary.TherearetwomainapproachesinusingEEGtoexploreperceptualreversals:(1)usingmanualresponseastheafore-mentionedtimereferenceor(2)usingstimulusonsetastimereference.Bothapproacheshavesomediculties.Intherst,thetemporalinformationisstronglyblurredduetothelargevari-ationintimingofmanualresponse(i.e.buttonpress).KornmeierandBach(2004)demonstratedthiseectbyaveragingEEGrecordingsofunambiguous26 rightinferiorfrontalcortexforendogenouslycomparedtoexogenouslyin-ducedperceptualreversals.SimilarresultshavebeenfoundbyShenetal.(2009)foraNeckerlatticestimulus.6Alsocomparingendogenousandexo-genousreversals,theauthorsidentieddestabilisingsignalsfromtherightdorsalfrontalcortexandfurthermorestabilisingsignalsfromtherightan-teriorportionofsuperiortemporalsulcus.Thelatterwasassociatedwithperceptualmemory,i.e.theaccumulatedinuencesofpreviousperceptions.AreviewoftheneurophysiologicalprocessesinvolvedinbistableperceptionwascomposedbySterzeretal.(2009).Theauthorspointedoutevidenceforinteractionsofbothlow-andhigh-levelbrainregionsandforearlyandlateprocessing.TMSIncontrasttomethodslikeEEGandfMRI,transcranialmagneticstimula-tion(TMS)isbettersuitedtodetectcausalinvolvementofbrainregions,asitcanproducetemporallylimitedvirtuallesions.Withaspinningwheelillusionstimulus,abistableapparentmotionstim-ulus,Geetal.(2007)appliedTMStotherightsuperiorparietallobule.BycomparingthisconditiontoonewithoutTMSstimulation,theauthorsfoundthatthisregionplaysacriticalroleinperceptualreversalofthisam-biguousgure.Kanaietal.(2011)usedTMSinordertostudytheneuralbasesofvisualbistabilityforbothbinocularrivalryandforanambiguousstructure-from-motionrotatingsphere.Forbothstimuli,theauthorsfoundthatstimulationoftheanteriorrightsuperiorparietallobuledecreaseddwelltimeswhilestimulationontheposteriorpartofthesamestructureincreasedthem.Thisshowsthefractionationofparietalcortexfunctionwithrespecttobistableperception.Alsoforastructure-from-motionspherestimulus,deGraafetal.(2011)demonstratedthatthedorsolateralprefrontalcortexiscausallyrelevantforvoluntarycontroloverperceptualswitches,whileitisnotforpassiveobservationofthesamestimulus.Thus,theseTMSstudiescomplementthendingsusingfMRIpresentedabove,inparticularwithre-specttotheroleofthefrontalcortex.Showinganovelbistablestimulusbasedonapparentmotion,Zaretskayaetal.(2013)usedbothfMRIandTMSonthesamegroupofparticipants.ThisstimulusalternatesbetweenadynamicglobalillusorycontourGestalt 6ThisisanarrayofNeckercubes,whichhastheeectofenhancingneurophysiologicalstimulusresponses(Kornmeieretal.,2004;Shenetal.,2009).28 sionsandacontrolgroup.Usingbothvisualbistablestimuli,theNeckercubeandvase/faces,andauditorystreaming,vonSteinbücheldiscoveredin-creaseddwelltimesonlyforpatientswithlesionsintherightfrontalcortex.ThelesionstudiespresentedhereareingoodagreementwiththefMRIandTMSstudiescitedaboveandfurthervalidatetheimportanceoftherightfrontalcortexforperceptualreversals.ResearchemployingTMSandfMRIhasprovidedadditionalinsightsonexactlocationsanddemonstratedotherinvolvedbrainregions,whilethetemporalaspectsofthereversalprocesshavebeenelucidatedwiththehelpofEEG.Hence,thedierentapproachescom-plementeachotherindeepeningourunderstandingoftheneurophysiologyofbistableperception.1.5GeneticsRecently,alsogeneticshavebeenrelatedtobistableperception.Shannonetal.(2011)exploredperceptionoftheNeckercubeandbinocularrivalryinbothmonocygoticanddicygotictwins.Theyfoundstrongcorrelationsinbothstimulustypesbetweenmonocygoticbutnotdicygotictwins(r0:55).Also,theyshowedthatbetweenmonocygotictwinsthenumberofreversalsforbinocularrivalrycorrelatedwiththenumberofreversalsfortheNeckercube(r=0:37).Therewasnosuchcorrelationbetweendicygotetwins.Theseresultsledtheauthorstotheconclusionthatthereisaheritablebasisforbistableperceptionandthatsimilargenesareinvolvedindeterminingthetemporaldynamicsfordierentformsofmultistableperception.Kondoetal.(2012)comparedthenumberofreversalsamonggenotypegroupsfortwogenesanddierentvisualandacousticmultistablestimuli.Poly-morphismsofcatechol-0-methyltransferase(COMT)Val158Metandserotonin2Areceptor(HTR2A)-1438G/Awereconsidered.Theauthorsfounddier-encesinthenumberofreversalsforacousticbistability(auditorystreamingandverbaltransformation)intheCOMTgenotypegroups,whiletheHTR2AgenotypegroupsdieredinperceptionoftheNeckercubeandavase/facesstimulus.Itwasconcludedthattheserotoninsystemisrelatedtopercep-tionofambiguousgures,inparticularcloselylinkedtotheso-calledshapefactorthattheauthorsidentiedwithafactoranalysis.Theauthorssug-gestedthatdevelopmentaldierencesduetothevariancesintheserotoninsystemmightcausethedierencesinreversalbehaviour.TheresultsofShannonetal.(2011)suggestthatabout30%oftheinter-30 tobistableperceptionofaplaidmotionrivalrystimulus7(HupéandRu-bin,2003).Theproportionoftimespentwiththeperceptofasinglemov-ingplaidcorrelatedpositivelywithmoodstate.Itwouldbeinterestingtotestwhetherthisndingcanbereproducedwithperceptualbiasesforotherbistablestimuliandalargersample,asSheppardandPettigrew(2006)hadonlytenparticipants.Fordierentcontentreversalstimuli,AllenandChambers(2011)comparedambiguousgureperceptionbetweenadolescentswithautismspectrumdisorderandlearningdisability.Byhavingthemcopythestimuliunderdierentconditions,theauthorsfoundthattheparticipantswithautismspectrumdisorderprocessedtheimagesconceptuallydierently,notbeinginuencedbycontextualinformation.Bilingualismseemstobeanothercharacteristicrelatedtobistablepercep-tion.BialystokandShapero(2005)foundthatbilingualchildren(ca.6yearsofage)discoveredthealternativesofthecontentreversalstimulifasterthanmonolingualchildren.Patientsofschizophreniaandbipolardisorderweretestedinpercep-tualreversalsoftheNeckercubebyHuntandGuilford(1933).Theauthorsfoundthatthegroupofschizophrenicpatientswasalmostidenticaltoacontrolgroupintermsofnumbersofreversalswhereasthebipolarpatientsreportedmuchlessreversalsthanthecontrols.Theauthorssuggestedthatbistableperceptionmightbedevelopedintoadiagnostictool.Krugetal.(2008)alsoreportedaloweramountofreversalsforbipolarpatientscom-paredtohealthyobservers,butthistimeastructure-from-motionrotatingcylinderwasusedasstimulus.Theauthorsconclude,though,thatbistableperceptionisnotsuitableasadiagnostictoolduetothelargerinter-personalvariationofdwelltimes.Forbinocularrivalry,similarndingswerestatedbyMilleretal.(2003).Theauthorsfoundthatpatientswithbipolardis-orderexperiencedlessreversalsthancontrols,whileschizophrenicpatientsandthosewithmajordepressiondidnotdierfromacontrolgroupinthatrespect. 7Inthisstimulustwogratingsareoverlaidandmovedwithrespecttoeachother.Theyareseenthroughacircularapertureandcanbeperceivedaseitherslidingovereachotherindependentlyorasasingleplaidmovinginonedirection.32 fortwoambiguousguresandtwobinocularrivalrystimuli,ndingahighdegreeofsimilaritybetweenallofthem.34 Evidenceforbottom-uptheories Initialadaptation Initialincreaseofthenumberofreversals,i.e.decreaseofdwelltimes.AsdemonstratedinSec.4.1,thisef-fectseemstobemainlyduetoaninitialconfusionintheexperimentalsituationandtheneedtofamiliar-isewiththereversalphenomenon,asashorttrainingsessionseemstoremovetheeecttoalargeextend. Localadaptation ToppinoandLong(1987)foundthatif,afterinitialadaptation,astimulusismovedtoadierentlocationinthevisualeld,reversalratesareagainonthesamebaselinelevelaspriortoadaptation. Multiple-gurepresentation Simultaneousobservationoftwoormorebistablestimuliischaracterisedbyindependentreversalsaswellasindependentadaptation(ToppinoandLong,1987). Reverse-bias(priming) Prolongedexpositiontoanunambiguousversionofabistablestimulusleadstoapreferenceoftherespect-iveotherperspectiveofthatstimulusinsubsequentobservationofit(Longetal.,1992).Thisseemstobeanadaptationeectasitdisappearswhentheam-biguousstimulusispresentedtoanotherretinalregionafterthepriming. (Dis-)Continuityofpresentation Presentingbistablestimulidiscontinuouslyinuencesthenumberofreversalssubstantially,evenleadingtocompleteabsenceofreversalsforsucientlylonginter-stimulusintervals(Leopoldetal.,2002). Viewingparamet-ers Dierentviewingandstimulusparameters,likesize,stimuluscompletenessetc.,haveaninuenceonthenumberofreversals(cf.alsoSec.1.3.2).Table2.1:Someclassesofevidencesupportingbottom-upexplanationsofbistableper-ception.FurtherdetailsandreferencescanbefoundinLongandToppino(2004).36 dwelltimescouldbeincorporatedandquantitativelyreproduced.TheresultssummarisedinTab.2.2,ontheotherhand,showinuencesofhigherorderandcognitiveprocesses.Theseprocessesaremoreofanactivenatureandclosertoconsciousperception.Asthereisampleevidenceforbothbottom-upandtop-downeectsinbistableperception,ithasbecomeclearthatbothtypesofeectshavetobeimplementedinausefulmodelofbistableperception(Longetal.,1992;ToppinoandLong,1987).Thus,LongandToppino(2004)proposedaqual-itativehybridmodelconsistingoffourlevels:feature-extraction,processing,representationandanonsensorycorticallevel,whichinteractinseveralways.KornmeierandBach(2012)suggestedanotherqualitativemodelbasedontwoattractorsinastatespacewhicharemodulatedintheirdepthbydestabil-isationanddisambiguationprocesses.Inthisapproachbothbottom-opandtop-downaspectscaninuencetheseprocesses.2.2OscillatorsorAttractors?Theapproachespresentedabovearemainlyqualitative.Inthelastdecade,increasinglymorequantitativemodelshavebeendevelopedthatarebasedonactualneuronalstructures.Mostofthemfallintoeitheroftwoclasses.Theoscillator-typemodelsfeatureanoisyoscillatorcircuit,withadapta-tionbeingthedrivingforcebehindreversals.Intheattractor-models,ontheotherhand,noiseisthedriverofperceptualreversals,withadaptationonlymodulatingthisprocess.Thatmeans,thatwithoutnoise,therewouldbenoreversalsinanoise-drivenmodel,whileoscillatormodelswouldbeperfectlyperiodicalwithoutnoise(Shpiroetal.,2009).Ontheotherhand,noise-drivenmodelswithoutadaptationpredictexponentialdistributionsofdwelltimes,notgammaorlognormaldistributions(BraunandMattia,2010).Thus,itseemsthatmodelsbeingbasedexclusivelyononeortheotherpro-cessarenotrealistic.Shpiroetal.(2007)comparedfouroscillator-typemodelsforbinocularrivalrywhichwerebasedoncross-inhibition.Theauthorsfoundregimesofdier-entdynamicalcharactersiticsinthespacespannedbythemodelparameters,whichcanaccompanyeectsofvaryingstimulusstrength,i.e.variationsintheinputforthemodels.AnattractormodelwithweakadaptationwaspresentedbyMoreno-Boteetal.(2007),whichisimplementedbothinringratemean-eldandin38 timesothatthenormalisedshapeofthedwelltimedistributionoffastrevers-ersisverysimilartothoseofslowreversers.Furthermore,inordertoexplainthedecreaseinreversalsobservedforintermittentpresentationofambiguousguresforinter-stimulusintervalslargerthanabout350ms(KornmeierandBach,2012),noisyoscillatormodelshavetoresorttopostulatinganaddi-tionaladaptationprocessatlongertimescales.Thiscanalternativelybeachievedbyahierarchyofattractorsoperatingatdierenttimescales.2.3FurtherApproachesWhiletheoscillatorandattractormodelsreviewedabovearebasedonreal-isticneuralstructures,therearealsoafewmoreabstractmodelsforpercep-tionofbistablestimulithatwillbementionedverybrieyhere.ModelsbasedonBayesiandecisionprocesseswerepresentedbyvanEeetal.(2003)forslantrivalry,anambiguousgure,andSundareswaraandSchrater(2008)fortheNeckercube.Gershmanetal.(2009)suggestedamodelofperceptualinference,mainlyforbinocularrivalry,basedonMarkovChainMonteCarlomethods.Themathematicalformalismofaquantum-mechanicaleectisusedintheNecker-Zenomodelofbistableperception(Atmanspacheretal.,2008,2004).InanalogytothequantumZenoeect,atwo-levelsystemwithaperiodicupdating,ormeasurementprocess,wasproposed.Decoherencewasin-troducedwithatimeevolutionoperator,leadingtoadecayofthewavefunctionandeventuallyaperceptualreversal.Undertheinclusionofanearlyadaptationphaseinperceptionofabistablestimulus,calculationsofthetransitionprobabilitiespredictgammadistributeddwelltimes.40 routinesmakinguseoffunctionsoftheStatisticsToolboxforstatisticalana-lysisanddistributiontting.BothstudieswereapprovedbytheethicscommitteeoftheETHZürich.Participantswerecompensatedmonetarilyfortheirparticipationinthestud-ies,receivingabout20CHFperhour.Dierentparticipantsattendedeitherstudy.TheywererecruitedviaanonlineplatformoftheUniversityofZürichandweremainlystudentsofeitherETHZürichorUniversityofZürich.3.1NC-dist:TemporalDynamicsandLow-levelFeaturesinBistablePerceptionoftheNeckerCubeTherststudywasmainlyfocusedonlow-levelaspectsofbistableperceptionoftheNeckerCube,whilealsoexploringsomehigh-levelfeatures.TheNeckercubewasusedasstimulusasitischaracterisedbylowsemanticcontentandahighgeometricsymmetry.Thus,forthisgurelessconfoundinginuenceswereexpectedcomparedtoastimulusliketheoldwoman/youngwomandrawing.3.1.1ResearchQuestionsoftheNC-distStudyThefollowingquestionsweresupposedtobeansweredinthisstudy:Whatisthemostreliabledescriptionofthedwelltimedistribution?IsthereaclearlydistinctinitialphaseofbistableperceptionoftheNeckercubewhenanobserverexperiencesperceptualreversalsforthersttime?HowdoesperceptionoftheNeckercubedependonstimulussizeforsmallvisualangles?Whichcommonfeaturesaresharedbetweentheverbaltransforma-tioneectandbistableperceptionoftheNeckercube?Toaddressthesequestions,theNC-diststudywith5sub-experimentsand2questionnaireswasconductedinarandomiseddesignwith23healthy,Ger-manspeaking,right-handedparticipants.Foreachparticipant,dwelltimes42 datasetsactuallyusedwaslowerthanthatduetooutliersandmissingdata.Thefollowingexperimentswereused.InavoluntarycontrolparadigmfortheNeckercube,afteraneutral,passivecondition,participantswereaskedindierentconditionstotrytoholdeitherofthetwoperceptsorspeedupthereversalprocess.Toassesstemporalpro-cessing,asimpleGoNogoreactiontimetask,anacousticmeasurementoftemporalorderthreshold(Ulbrichetal.,2009)andatemporalintegrationtask(Szelagetal.,1996)wereconducted.Twoothercognitivemeasures,namelyworkingmemorycapacity,conceptualisedwithareadingspantaskandabackwarddigitspantask(Oberaueretal.,2000),andattentionprobedwiththed2attentiontask(Brickenkamp,2002)wereexamined.AparadigmbyHocketal.(1993)wasadaptedfortheNeckercubeinordertotestwhetherinthetransformationprocessfromoneunambiguousperspectiveoftheNeckercubetotheotherperceptualreversalsarelaggingbehindthetransformation.Thepresenceofsuchalagwouldbecalledhysteresiseect.Concerningpersonality,twoshortquestionnaireswereusedtoassesssensa-tionseekingbehaviourandambiguitytolerance,namelytheBriefSensationSeekingScale,BSSS,(Hoyleetal.,2002)andtheUngewissheitstoleranz-skala,UGT,Dalbert(1999).Furthermore,aGermanversionoftheNEO-Five-FactorInventory(NEO-FFI/BIG-5)wasusedforacoarsepersonalitycharacterisation(Körneretal.,2008).MindfulnesswasassessedusingtheFMI(Walachetal.,2006)andCHIMEquestionnaires(Bergomietal.,2012),bothofwhicharenotbasedonspecialisedvocabularydescribingmeditationorreligiousaspectsofmindfulness.TheSTAItraitandstateinventory(Lauxetal.,1981)wasusedtoassessanxiety,whileaction-controlwasexploredwiththeHAKEMPquestionnaire(Kuhl,1994)andself-leadershipwiththeRSLQ-D(AndreÿenandKonradt,2007).Alsoforthisstudy,moredetailsontheindividualexperimentswillbegiveninthefollowingchapters.3.3MeasuringBistablePerceptionThissectiondescribesthegeneralprocedurefortheexperimentsonbistableperceptionoftheNeckercubeasemployedinthecurrentwork.Inallex-periments,theNeckercubewaspresentedasablackonwhitelinedrawingonacomputerscreenusingPTB-3.AnillustrationofthecubeisgiveninFig.3.1.IntheNC-diststudytheNeckercubewasshownin6dierentsize44 B(theinverseofholdA)andspeedup,whereparticipantswereinstructedtoswitchbetweenperceptsasquicklyaspossible.Inbothstudiesparticipantsrestedtheirheadinachinrestinordertominim-izemovementsoftheheadandinordertokeeptheviewingdistanceconstant.Furthermore,theywerealsoinstructedtominimizeeyeandheadmovementswhilexatingalittlecrossthatwasshowninthemiddleoftheNeckercubethroughoutpresentation.Intherststudyandintheneutralconditionofthesecondone,theywereaskedtoobservereversalsinapassivefashion,withouttryingtoexerciseanyvoluntarycontroloverreversals.3.4AnalysisofDwellTimeDataTherawdwelltimedatarecordedforallNeckercubeexperimentsconsistedofthemachinetimeofeachbuttonpressandacodeforwhichbuttonwaspressed.Beforefurtheranalysis,relativedwelltimeswerecalculatedbysub-tractingsubsequentdwelltimesfromeachother.Therstdwelltimeofeachdatasetwasdeleted,asthecontextoftherstbuttonpressdiersfromalltheothers,notbeingprecededbyanotherreversal.Additionally,dwelltimedatahadtobecorrectedforconsecutivepressesofthesamebutton.Inthoseinstancesitwasunclearwhattheparticipantac-tuallyperceived.Fourpossibilitiesexist.First,theparticipantcouldhavepressedthebuttoninordertoreconrmhiscurrentpercept.Aftertheexper-iment,someparticipantsspontaneouslyreportedhavingusedthisbehaviourwhentheirperceptbecameambiguousforashortinstancebutthenreturnedtothesameperceptasbefore,i.e.noreversaloccurred.Tocorrectthedatainthiscase,itwouldbeindicatedtoaddupthetwoinvolveddwelltimesinordertogetthetimesbetweenperceptualreversals.Secondly,theparti-cipantcouldhavefailedtoindicateareversalinbetweenthetwoconsecutivebuttonpresses.I.e.asequenceAAwouldhaveactuallybeenABA.Here,therstofthetwodwelltimesshouldberemovedasitwouldsplitintotwoseparatedwelltimes,thesizeofwhichisnotknown.Thirdly,itispossiblethattheparticipantmixedupthetwobuttons,pressingthewrongone.Suchamix-up,ifnotcorrectedbyanotherquickbuttonpressoftheparticipant,wouldresultinthreeconsecutiveoccurrencesofthesamebut-ton.Thenthebuttoncodemiddleoneshouldbealtered.Fourthly,theparticipantmighthaveaccidentallypressedabutton.Thereisnowaytocorrectforthat,asonedoesnotknowwhethertherstortheseconddwell46 4.TemporalDynamicsThischapterwilldetailthetemporaldynamicsofbistableperceptionoftheNeckercube,withastrongfocusonthedistributionofdwelltimes.4.1StationarityBeforeexploringthedetailsofthedwelltimedistributionandndingad-equatetfunctionsforit,theissueofstationarityofthedwelltimedistri-butionshouldbeaddressed.SeveralarticlesonbistableperceptionoftheNeckercubementionaninitialincreaseofthenumberofreversalsoverroughlytherstminute.Brascampetal.(2005)reportedthatadriftindwelltimedatawasrestrictedtotherst30sofobservationwithoutshowingdataforthisnding.Cohen(1959b)recordedthenumberofreversalsinevery15s.Theauthorsfoundthismeas-uretoincreaseandthentoleveloafteroneminute.BabichandStanding(1981)presentedsimilarresultswithlevelingafter75s.Studyingtheactualdatagiven,though,thenumberofreversalsinfactdoesnotstayconstantbutvariesstill.Acompleteconstanceofreversalswould,ofcourse,contradictthestochasticnatureofbistableperception.SadlerandMeerd(1970),ontheotherhanddidnotndsuchanincreaseinreversalswithinoneminute.Furthermore,withapileofcubesstimulus,Price(1967)foundaninitialdecreaseofdwelltimesonlyforthepreferredperceptandaconstantdwelltimelevelforthenon-preferredpercept.ThisstimulusisdierentfromtheNeckercubeandmightvaryinitstemporaldynamics.ItshouldbenotethatthestudiesbyCohen,Babich,Sadlerandtheirrespect-iveco-workershadparticipantsindicatereversalseitherverballyorwitha48 atleasttwoandahalfminutesofdwelltimedatawasused.Thisresultedintheexclusionof3datasets,outof23.TheWilcoxontestsshowedthattherewerenosignicantdierencesbetweentherstandthesecondintervalforbothmeasuresandintervallenghts(cf.Tab.4.1).Measuresdescribingthedwelltimedistributionwerenotconsidered,astherearetoofewdwelltimesintherespectiveintervalsforareliableestimationofthedwelltimedistribution.Inordertoexplorestationarityofthedwelltimedistribution,awaytoincreasetheamountofavailabledatapointswithinashortintervaloftimewouldhavetobefound.Onepossibilitythatcouldbeconsideredforthiswouldbenormalisingdwelltimesandmergingdatafordierentparticipants(e.g.PressnitzerandHupé(2006)).Theassumptionsmadeinthisapproachwouldhavetobetestedthoroughly,though.Onecanconcludethatafterashorttrainingandwaitingphase,asimplemen-tedhere,thereisnoevidenceforasignicantsystematicdecreaseindwelltimeswithintherst30or60s.ThisndingisinagreementwithastudybyNakataniandvanLeeuwen(2006)whodidnotndaninitialdecreaseindwelltimeswhilestillreproducingtheusualtemporaldynamicsintermsofthedwelltimedistribution.Oneprerequisitetoobservingstablemeandwelltimesis,ofcourse,thatparticipantsareawareofthereversalphe-nomenonandhaveexperienceditclearly.InstructionsdoplayanimportantroleforthisasRockandMitchener(1992)alreadyshowed.Theauthorsdemonstratedthatignoranceofthereversalprocesswouldpreventreversalsforasmuchas70%ofparticipantsintherst30s.Thus,theexistenceornon-existenceofaninitialdecreaseindwelltimesindierentarticlesintheliteratureonbistableperceptionoftheNeckercubemightberelatedtothewayinstructionsaregiven.Forthecurrentpurposes,itissucienttoobservethatafterfamiliarisationwiththereversalphenomenonmeandwelltimesdonotchangesignicantlyovertheinitialphaseofobservationsothatalldwelltimedatacanbeusedforfurtheranalysis.4.2ReproducibilityAsdetailedinSec.1.3.3,GuilfordandHunt(1931)andFrederiksenandGuilford(1934)showedthatthemeannumberofreversalspertimeintervalwhentheNeckercubeisfairlystablewithinoneperson.Inasmallpilotstudy,fourparticipantsviewedthedrawingofaNeckercubeonconsecutive50 tifyingthereversalprocessforonegivenobserver,itisalsopossibletoonlyconsidercertaincharacteristicmeasuresofthedwelltimedistribution,likethemean,themedianorthemodaldwelltime.Themeandwelltime,inthefollowingindicatedwitht,canbeestimatedeitherbycalculatingthearithmeticmeanoftherecordeddwelltimesorbyttingaparametricPDFandcalculatingitsmean.Analogously,themedian,~t,canbeestimatedintwoways:eithercalculatingitdirectlyfromaparametricPDForbyndingthatvalueinthedatathatseparatesthelowerfromthehigherhalfofthesample.2Themode,whichwillbelabelledt0inthefollowing,isthatvalueofarandomvariablewhichismostlikelytooccur,i.e.itisthepointwherethePDFhasitsmaximum.ItcanthusbedirectlycalculatedfromaparametricPDF.Itcanalsobedeterminedfromanon-parametricPDFbyndingitsmaximum.Ontheotherhand,themodecannotbeestimateddirectlyfrommeasuredvaluesofacontinuousrandomvariable,inthiscasedwelltimes,withoutfurtherassumptions.3Asmean,medianandmodeare,ineachcase,measuresdescribingonlyonecharacteristicofthefullPDF,agoodestimationofthePDFprovidesamoredetaileddescriptionofthetemporaldynamicsofbistableperception.Ontheotherhand,measuresdescribingcentraltendencyofthedwelltimedistri-bution,inparticularmeanandmedian,seemtobemorerobustandhencethesemeasuresaremorelikelytobethebetterchoiceforsmalldatasets.Nevertheless,itisimportanttotrytogainasthoroughanunderstandingoftemporaldynamicsaspossible.ThisindicatestheuseofthefullPDFandnotonlyoneofitsmeasures,e.g.themean.Inparticular,thisapproachallowsforthestudyofrelationsbetweentheoverallshapeofthePDFandmeasuresdescribingotheraspectsofcognitionorpersonality,forexample.Thus,rstanoverviewoverpossiblettingprocedureswillbegiven,followedbyadescriptionofseveralparametricPDF'sthatseemtoyieldgoodtsfordwelltimedata.Finally,theissuesoftqualityandthestationarityofthedistributionofdwelltimeswillbeapproachedwiththehelpofdwelltimedatafromtheNC-persstudy. 2Ifthenumberofdatapointsiseven,suchadatapointdoesnotnecessarilyexist.Inthatcase,themedianisusuallyestimatedasarithmeticmeanofthesmallestvalueinthehigherhalfofthesampleandthelargestvalueinthelowerhalf.3Toestimatethemodefromalargeenoughsetofmeasuredvaluesofacontinuousrandomvariable,onecandiscretisethedatabyintroducingabinning.Themodewillbethemid-pointofthebinwiththehighestnumberofdatapointsinit.52 leastsquares,themaximumlikelihoodmethodyieldstheparametersoftheconsidereddistributionfromwhichCDF,PDFandmeasuresofthedistribu-tioncanbecalculated.Anadvantageofthismethodisthatnoassumptions,likendingabandwidthordeterminingsamplepoints,areneeded.UsingtheFisherinformationmatrix,estimatesofthestandarddeviationsoftheparametersofthetteddistributioncanbedetermined.Becauseofthesetwoadvantagesovertheothertwomethods,here,themax-imumlikelihoodestimationwillbepreferredoverKDEandleastsquarests.4.5ProbabilityDensityFunctions4.5.1TheGammaDistributionSeveralauthors(DeMarcoetal.(1977);Borsellinoetal.(1972);Brascampetal.(2005))foundthatdwelltimedistributionscanbemodeledusingthegammadistribution.Thefollowingparametrisationofthegammadistribu-tionisusedthroughoutthistreatise:f (t)=ba+1 �(a+1)tae�bt(4.1)wherea;b2R,witha￿�1;b￿0and�(x)isthegammafunction.Thegammadistributionisaright-skewed,unimodaldistribution.Itsmode,i.e.thet-valueofitsmaximum,caneasilybecalculatedanalytically.DeningC:=ba+1 �(a+1),thersttwoderivativesoff (t)are:f0 (t)=Cta�1e�bt(a�bt)f00 (t)=Cta�2e�bt(a(a�1)�2abt+b2t2)(4.2)Inordertodeterminethemaximumoff onesetsf0 (t0; )=0yieldingt0; =a b(4.3)astheonlynon-vanishingsolution.Thesecondderivativeoff att0; isf00 (t0; )=C(a b)a�2e�2(�a)0(4.4)fora;b&#x]TJ/;༕ ;.9;Ւ ;&#xTf 1;.42; 0 ;&#xTd [;0.Hencethegammadistributionhasalocalmaximumatt0; =a bfora￿0.54 4.5.3OtherPDF'sApartfromthegammadistributionandthelognormaldistribution,forttingdwelltimedataboththeWeibulldistributionfWeibull(t)=ba�btb�1e�(t a)ba;b￿0(4.9)andtheRayleighdistributionfRayleigh(t)=t 2e�t2 22￿0(4.10)canbeconsidered.TheWeibulldistributionisunimodalandright-skewedforb￿1.Infact,itisageneralisationoftheRayleighdistributionandforb=2theWeibulldistributionequalstheRayleighdistributionwitha2=22.Zhouetal.(2004)testedthetqualityoftheWeibulldistributionfordwelltimedataoftheNeckercubebutfoundthatitisinferiortothegammaandlognormaldistributions.TheRayleighdistributionhasnotbeenevaluatedyetforatofdwelltimedata.4.6FitQualityThetqualityofthegamma,lognormal,WeibullandRayleighdistributionsfordwelltimedataoftheNeckercubewereevaluated.Dwelltimedataof58participants(29:09:5years,27male)fromtheneutralconditionoftheNC-persstudywasusedfortheanalysis.Dataof7participantsoftheavail-able65wasnotusedbecauseoftoomanymultiplesubsequentpressesofthesamebuttonoccurred.AfterfamiliarisationwiththereversalprocesseachparticipantviewedtheNeckercubefor3minutesandindicatedreversalsviabuttonpresses.ParticipantswereinstructedtopassivelyviewtheNeckercube,i.e.nottotrytoinuencethereversalprocess.Furtherdetailsonex-perimentalprocedureandrstanalysisaregiveninSecs.3.2,3.3and3.4.Eachdatasetwasttedwitheachofthedistributionsdescribedabove.Ad-ditionally,inversedwelltimeswerettedwiththegammadistributioninordertoreproduceareportbyBrascampetal.(2005).Theauthorstteddwelltimeswithgammaandbetaprimedistributionsaswellasinversedwelltimeswithgammadistributionsandfoundthatthegoodnessoftwasbestfortheinversedwelltimegammat,whichtheycalledgammaratet.56 experimentalsamplewithareferencedistribution.TheK-Steststatisticisdenedas:Dn=suptjFn(t)�F(t)j(4.11)whereFn(t)istheempiricalCDFforasampleofnobservations,i.e.ndwelltimes.F(t)istheCDFofthettedprobabilitydensityfunction.I.e.geo-metrically,theteststatisticDnisthemaximalverticaldistancebetweentheempiricalandthettedCDF.DnisthencomparedtocriticalvaluesoftheKolmogorovdistribution.Thus,itcanbedecidedwhethertheNull-hypothesisshouldberejectedornot.Also,p-valuescorrespondingtoeachDnaretabled.Ifthereferencedistributioncontainsparametersthatareestimatedfromthesample,thenthecriticalvaluesdeterminedfromtheKolmogorovdistribu-tionaretooconservative,though(Woodruetal.(1984),Massey(1951)).I.e.theactualcriticalvalueswouldbelower.Thisisbecausebyestimat-ingtheparametersofthePDFfromthegivendatawillchangetherelationbetweenFn(t)andF(t)andhencethedistributionoftheteststatisticDn(Keutelian,1991).TheK-Stestcanstillbeconducted,butthepropercriticalvaluesoftheteststatisticDnhavetobedetermined.ThiscanbedonebyestimatingthedistributionofDnusingMonteCarlomethods(Keutelian,1991;Lilliefors,1967;Woodruetal.,1984).KeutelianusedthisapproachforaGaussiandistributionforwhichtheparametershadbeenestimatedfromthedatathatwastested.HeintroducedthenotationDNptolabeltheteststatisticinthecaseofestimatedparameters.UsingMonteCarlosimulationshedemon-stratedthatthedistributionofDNpwasindeedshiftedtotheleftcomparedtothatofDN.Inotherwords,themodiedK-Stestyieldedlowercriticalvalues.TheMonteCarloapproachhadtobeusedhereaswellbecauseallconsidereddistributionswerettedtotheempiricaldataandwerenotgivenapriori.Note,thatthisapproachwasnotusedinthecitedpublications(Brascampetal.,2005;Zhouetal.,2004),i.e.thecriticalvaluesoftheK-Stestusedinthesearticleswerenotmodiedtoaccommodatethefactthatparamet-erswereestimatedfromthetesteddata.Thus,adierenceinp-valuesandrejectionratestotheresultspresentedbelowisreasonable.Inthepresentanalysis,rst,foreachdatasettheparametersofthedistribu-tionsweredeterminedwiththemaximumlikelihoodmethod.ThentheNulldistributionoftheteststatisticwasestimatedwithMonteCarlomethods,58 Figure4.1:BoxplotsofSSEandpKSfor58observersoftheNeckercubeasmeasuresofgoodnessoftforallconsidereddistributions.Forthesumofsquarederror(SSE,toppanel),asmallnumbermeanagoodt.ForpKSavaluecloseto1indicatesagoodt.60 ratedistributions.Inspectionoftheresultsofthemaximumlikelihoodshowsthatthettedparametervaluesofthelognormalandthelognormalratetsareinfactidentical.Thus,nonewinformationisgeneratedbyttingtheinversedwelltimeswiththelognormaldistributioncomparedtotheoriginaldwelltimes.ButasSSEandpKSshow,neitherdwelltimesnorratesareperfectlylognormallydistributed.Theempiricaldwelltimedistributiondif-fersfromthelognormaldistributioninacertainway.Thisdierenceisingeneralnotinvariantunderinversion.ThatiswhybothSSEandpKSaredierentbetweenthelognormalandthelognormalratetandwhyadier-enttqualityisfoundfortimesandrates.ThisisinfactconrmedbyananalysisoftresidualsinSec.4.6.3,thatshowsasomewhatdierentshapeoftheresidualcurvesofthelognormalandlognormalratets.AsthetwolognormaltsareequivalentandasthelognormaldistributionconstitutesthesecondbesttatleastintermsofpKSvaluesitisap-propriatetousethelognormaldistributioninordertodescribethedwelltimedistributionforbistableperceptionoftheNeckercube.Thisprovidestheadvantagethatthedescriptionisintermsoftimesandnotinversetimes.WhiletheformerhaveaclearcorrespondencetotheperceptionoftheNeckercube,namelyasanestimationofthetimebetweenperceivedreversalsofper-spective,theinversetimesdonothavesuchacorrespondence,i.e.theyarenotdirectlyrelatabletotheobservationoftheambiguousgure.Thisisalsoanadvantageofthelognormaltsoverthegammaratets.Here,thisadvantageaswellastheclearlyhigherpKS-valuesofthelognormaltscom-paredtothegammaratetsarejudgedmoreimportantthanthebetterSSEvalueofthegammaratets.Hence,thelognormaldistributionwillbeusedthroughoutthisworkinordertodescribetheperceptionoftheNeckercubeasitprovidesthebestcombinationoftqualityanddescriptiveness.AfewmoreremarksconcerningthepKS-valuesareappropriate.ThenumberofrejectedtsforWeibullandRayleighdistributionsindicatethatthesetwofunctionsareclearlynotappropriatettingfunctions.Alsoforthegammadistribution,asignicantportionofthets,namelyonethird,havetoberejected.OneshouldfurthernotethatnorejectionoftheNullhypothesisofequality,i.e.pKS￿0:05,doesnotimplythatthetisnecessarilyagoodone.Itonlymeansthatintermsoftheteststatistic,thetestedtfallswithintheregionthatcovers95%ofthesampleddata.Tospeakofagoodt,pKSvaluessignicantlylargerthan0.05wouldbedesirable.Evenforthegammarateandthelognormaldistributions,abitlessthan20%ofdatasetsarenotttedadequately.Thismayofcoursebeduetothelownumberof62 Neckercube.Hence,theparametersandmeasuresofthelognormaltswillbeusedinthefollowingchapterstoexplorebistableperceptionoftheNeckercubeanditspotentialrelationstootherdomains,likecognition,temporalprocessingandpersonality.Beforethoseanalyses,tresidualswillbediscussedinthefollowingsection.4.6.3FitresidualsAnotherwaytostudythequalityofthetsistoexaminethetresiduals.Fitresidualsarethedierencesbetweenempiricalandtteddwelltimedis-tribution.ResidualswerestudiedhereinanalogytotheanalysispresentedinBrascampetal.(2005).Forthat,thedierencesbetweenempiricalandttedCDFwerecalculatedforlognormal,lognormalrate,gammaandgammaratets.Thesedierenceswerebinnedwithrespecttoadetrendedordinate:thetimevalueineachtime-residualdatapairwassubstitutedbythecorrespond-ingprobabilityvalueoftheempiricalCDF,sothatresidualplotsofdierentparticipantscouldbeaveragedinspiteofdierencesinthevarianceofthedistributions.Thesamewasdonefortherate-residualdatapairsincaseofthelognormalrateandthegammaratets.Afterthisdetrendingstep,everyparticipanthasanunevenlyanddierentlyspacedordinate.Hence,residualsofallparticipantswerecollectedin20equidistantbinscoveringtheordinateinterval[0;1].Eachbinthencontainsavaryingnumberofresiduals.Foreachsuchbin,andeachtmethod,meanresidualswerecalculatedandplottedvs.binordinate.TheresultsareplottedinFig.4.3.Thelognormalratedistributionshowsthelowestresiduals,withamoreorlessatcurve.Theresidualsofthelognormaltaresimilar,butabitlargerformediumprobabilitiesandsmal-lerforlowandhighprobabilities.Thisdemonstratesagainthedierenceintqualitybetweenlognormalandlognormalratetsmentionedintheprevioussection.Thegammaratedistributionshowsapeakjustbelow60%probability,whichisalsopresentforthegammats,butmuchbroaderandalsohigher.Thus,thisanalysisconrmsthatthebesttqualityisachievedbyttingtheinversedwelltimeswithalognormaldistribution,whilethelognormaltstothedwelltimesdonotperformmuchworse.Itisnoteworthythatallresidualcurvesarepositive,whichmeansthatthettedCDFliesbelowtheempirical.I.e.alltssystematicallyunderestimatetheactualdistribution.Thisndingcouldbeonestartingpointtofurtherimprovetquality.64 Figure4.3:Fitresidualsascalculatedfromthedierencebetweenempiricalandttedcumulativedistributionfunctionforlognormalrate,lognormal,gammarateandgammats.Ordinatesweretransformedfromtimeandrate,respectively,toprobabilities,whichisreferredtoasdetrendingandallowsthecomparisonbetweenparticipantswithdierenttemporaldynamicsinbistableperception.Standarddeviationswerenotplottedinordertoretainreadabilityoftheplot.66 5.StimulusPropertiesAftertheexplorationofdwelltimedynamicsinthelastchapter,inthisone,lowlevelstimuluspropertieswillbeconsidered.First,thequestionoftheinuenceofcubesizeonperceptionoftheNeckercubewillbeaddressed.Subsequently,apotentialhysteresiseectwillbeexplored.5.1SizeoftheNeckerCube5.1.1ReportsontheEectofCubeSizeThereareseveralstudiesontheNeckercubethatexploredtheinuenceofstimulussizeonthenumberofreversals.AlreadyWashburnetal.(1931)testedthreecubesizesofvisualangles0.7,7and64withasmallsampleoffemaleparticipantswhoindicatedreversalsverbally.Theauthorsfoundthatthelargecubewouldhavelessreversalsthanthesmallestone.DuggerandCourson(1968)consideredthreevisualangles,namly3,8and13,andfoundasignicantdecreaseinnumberofreversalsfrom8to13,againusingverbalreports.Testingsizesof2.6,12.8and25:1,BergumandFlamm(1975)alsodiscoveredasignicantdecreaseofthenumberofreversalswithincreasingcubesize.Inthisstudy,dwelltimeswererecordedusingbuttonpresses,thusprovidingahigheraccuracythanthetwostudiesmentionedbefore.Borsellinoetal.(1982)coveredabroadrangeofsizes(0.9,1.7,8.6,17,33.4,61:9)andreportedaplateauwithrespecttomeandwelltimesbetween5and20�30.Theseresultsallindicatethatdwelltimesincreasewithincreasingvisualangle.Thegoaloftheexperimentpresentedinthenextsectionwastoexplorethelowerrangeofcubesizesinordertogainanerdescriptionofsizedependence.67 Figure5.1:Sizedependenceofmean(left)andmediandwelltimes(right)aswellasthemode(bottomleft)andvariance(bottomright)ofthelognormaltfordierentvisualangles.Foreachmeasure,meanvaluesareshownwithstandarddeviations.69 anotheroneandbackagain,thepointinthemorphingseriesatwhichtheperceptionofanobserverchangesfromtherstimagetothesecondisdif-ferentfromthechangingpointintheotherdirection.Perceptualhysteresishasbeendescribedforabistable,apparentmotionstimulusbyHocketal.(1993).Theyformulatedthedenitionofhysteresisasfollows:Hyster-esisisindicatedwhenthetransitionfromPatternAtoPatternB,whichisobservedwhentheparameterisgraduallyincreased,occursatahigherparametervaluethanthetransitionfromBtoA,whichisobservedwhentheparameterisgraduallydecreased.ThisisbasicallyadicretisedversionofEwing'sdenitionabove,i.e.oneadaptedtoadiscretevariableinsteadofacontinuousonesuchasmagnetisation.ThedesignofHockandco-workerswasadoptedforbistableperceptionoftheNeckercubeinordertotestforpotentiallaggingeects.5.2.2ExploringHysteresisoftheNeckerCubePerceptualhysteresisfortheNeckercubewasdeterminedanaloguelytotheexperimentsofHocketal.(1993)whostudiedhysteresisforapparentmo-tion.TransitionimagesbetweenthetwounambiguousvariantsoftheNeckercubewerecreatedin10steps.Thiswasachievedbygraduallychangingtheopacityofthesixrelevantinnerlinesofthecube.Let0denotetheunam-biguousvariantA,10thevariantB,analoguelyforpointsinbetween.Ineachtrial11transitionimagesweredisplayed,eachfor200ms.End-pointsofthetransitionsvariedbetweenimage0and10.Therstimagewasdisplayedformorethanonestep,iftheendpointofthetrialwasnotthere-spectiveotherunambiguousimage.Thus,serieslike0-0-0-1-2-3-4-5-6-7werecreated,eachserieslastingfor2sintotal.TheA-to-Bserieshad10dierentdegreesoftransformationsrangingfromalmostnotransformationwith10%ofperceptBattheendoftheseries(imageseries:0-0-0-0-0-0-0-0-0-0-1)tofulltransformationwith100%ofperceptBattheend(0-1-2-3-4-5-6-7-8-9-10).TheB-to-AseriesrangedfromfulltransformationtoAwith0%ofperceptBattheendoftheseries(10-9-8-7-6-5-4-3-2-1-0)toalmostnotrans-formationwith90%ofperceptBattheend(10-10-10-10-10-10-10-10-10-10-9).Thedisplaytimeofallserieswaskeptthesamesothatnoconfoundinginuencewouldbeexpectedfromvaryingdisplaytime.I.e.thatswitcheswouldbemoreprobableinoneseriesbecauseitwaslonger.10trialsforeachofthe10possibleendpointsoftheseriesweredisplayed,formorphingseriesfromAtoBandfromBtoA.Thisresultedin200trials71 Figure5.2:TheratiosofperceptionofBintheA-to-BtransformationandperceptionofAintheB-to-AtransformationareplottedagainstthepercentageoftransformationfromAtoB.Thebluecurve(A-to-B)readsfromlefttoright,thegreenone(B-to-A)fromrighttoleft.73 Figure5.3:Fig.28fromEwing(1885)illustratingthehysteresiseectformagnetisationBofannealedsteelasafunctionofmagnetisingforceH.lastperceptisunambiguouslyarepresentationofperceptB,participantsdidnotseeBinabout40%ofthecases.Thisproducestheproblemthattheendpointofthetransformationinonedirectionisnotidenticalwiththestartingpointofthetransformationintheother.Thus,theexplanatorypowerofthisexperimentisverylimitedintermsofhysteresis,asitsdescriptionpresup-posesaclosedloopinparameterspacespannedbythetwovariablesinvolved.HerethisisthedegreeoftransformationtoperceptBandtheratiooftrialswithBaslastpercept.Intheoriginalcontextofmagnetisationthevariablesweretheeldstrengthoftheappliedmagneticeldandthemagnetisationofthetestedmaterial.Toillustratethedierencetotheoriginalreportofhysteresis,aplotfromEwing(1885)isreproducedinFig.5.3.Itcanbeclearlyseenthatthemag-netisationBontheverticalaxislagsbehindwithrespecttotheappliedmagneticforceH,withanitemagnetisationatzeroexternalforce.Incontrarytothephysicalexample,perceptionofNeckercubeasstudiedhereprovidesthedicultyofamixtureoftwoeects.Apotentialhysteresiseectisblendedwiththespontaneousreversalbehaviourcharacteristicofbistableperception.IncaseoftheapparentmotionstimulususedbyHock74 6.BiasEectEventhough,theNeckercubeisaverysymmetricgure,thereisapreferenceforoneofthetwopercepts.Thisndinghasbeennotedanecdotallybyseveralresearchers.Adetailedquantitativedescriptionofthisbiaseectwillbegiveninthischapter.Itisindeedaveryinterestingeectasitprovidesanotherpieceofevidenceforhigh-levelinuencesinbistableperception,andingwhichwillbediscussedbelow.6.1QualitativeReportsThereareseveralarticlesthatmentionapreferenceintheperceptionoftheNeckercube,namelythatitispreferablyseenintheperspectivefromabove(perceptA).TrojeandMcAdam(2010)callthistheviewing-from-abovebias.Intheirworktheyndthattherotationalbiasinanambiguouslyro-tatingshilouetteofadancerisduetoanelevatedviewpoint.Kornmeieretal.(2009)mentionthatitiswellknownthatthecube-front-sidebottomisthepreferredinitialperceptofmostobservers.TheyrefertotwostudiesauthoredbyJohnR.Pricethatdemonstratethatmeandwelltimesoverobserversarelongerforoneperceptthantheother.Inonecase(Price,1967)thestimulusisapileofcubesreproducedfromWarren(1919),andintheotheritisarotatingwirecube(Price,1969),whichisambigu-ouswithrespecttotheperceiveddirectionofrotation.Bothstudiesdonotquantifythestrengthofthisbias.Furthermore,andmoreimportantly,theyusestimulidierentfromtheNeckercube,sonodirectconclusioncanbedrawntotheNeckercube.NakataniandvanLeeuwen(2006)showttingresultsofboththegammaandthelognormaldistributionforbothperceptsseparately.However,theydonotdiscussthedierenceindwelltimesortestwhetherthedierencesingammaandlognormalparametersbetweenthetwoperceptsaresignicant.76 t ~t t0;logn   varlogn 0:85s 0:64s 0:20 0:18 0:05 8:22 22:8% 19:8% 8:4% 16:2% 9:6% 134:6%Table6.1:Absolute(top)andrelative(bottom)dierencesbetweenmeasuresforperceptAandperceptB(lefttoright:meanandmediandwelltimes,modeofthelognormalratedistribution,itsparametersandvariance).PositivevaluessignifylargervaluesforperceptA.Starsindicatethesizeofuncorrectedp-values::p0:05,:p0:01and:p0:001.cedure(BenjaminiandHochberg,1995).Meanvaluesofallmeasureswerecalculatedforbothpercepts.ValuesforperceptBweresubtractedfromthoseofperceptAtodeterminemeanabsolutedierencesbetweenpercepts.Bydividingbythesmallerofthetwovalues,relativedierencesbetweenperceptswerecalculated.6.3ResultsAbsoluteandrelativedierencesofthetestedmeasuresaregiveninTab.6.1.ThepositivevaluesindicatethatameasureisgreaterforperceptA.Infact,allsixtestedmeasuresaresignicantlylargerforperceptAcomparedtoperceptB.Aftercorrectionformultipletestingwiththefalsediscoveryratemethod,thevecorrelationsareallstillsignicantatthe0.01-level.TheeectsizesofthedierenceslistedinTab.6.1showthatforperceptAdwelltimesaresignicantlylargerthanforperceptB(20%formeanandmediandwelltimes,8%forthemode).Thevarianceofthelognormaltisparticularlysensitivetothedierences,showingaverylargeeectsize.6.4SeeingtheCubeFromAboveTheresultspresentedherequantitativelydemonstrateaperceptualbiasef-fectintermsofthedistributionofdwelltimesforbistableperceptionoftheNeckercube.Itisshownthattheperspectivefromabove(A)isperceptuallyfavouredovertheoneseenfrombelow(B).I.e.dwelltimesarelongerforperceptAthanforperceptB.Furthermore,itwasshownthattheshapeofthedistributionofdwelltimesisdierentbetweenpercepts:bothparametersand,aswellasthevarianceofthedistributionaresignicantlylargerforperceptA.Theeectofthesedierencesinthedistributionshapecanalso78 correspondingdwelltimes.Thistentativeexplanationmightbetestedbytryingtousecube-likeobjectsasprimesfortheNeckercubeinacontrolledexperiment.Inatwo-groupdesign,inonegroupparticipantswouldbeprimedmanytimeswithabird'seyeperspectiveonacuboidforshortperiodsoftimesandatdierentretinallocationstoavoidfatigueeects.Thesecondgroupwouldreceivesimilarprimesbutfromaworm's-eyeview,i.e.seeingtheobjectfrombelow.DwelltimedistributionsofasubsequentNeckercubeexperimentcouldthebecom-pared.IfthedwelltimesforperceptAwouldbelongerfortherstgroupthanforthesecond,thepresentedhypothesiswouldbesupported.81 conditionswasperformedbyKornmeieretal.(2009).Theauthorsusedthreedierentholdconditions,thepreviousonesplusanunspecicone,aswellasaspeedupconditionandexaminedthecombinedinuenceofvoluntarycontrolanddiscontinuouspresentationonbistableperception.Theauthorsfoundthattheeectsofbothinuenceswerefullyadditive.Thesereportsonvoluntarycontrolofbistableperceptionalsoshowedthatwhilereversalscouldbesloweddownorspedupbyparticipants,itwasnotpossibleforthemtopreventreversalsaltogether.1Twostraight-forwardstrategiestoinuenceperceptualreversalsareoftenmentionedbyobserversoftheNeckercube.Therstistouseblinkstopro-duceaperceptualreversal.Thesecondistheemploymentofgazepositiontoholdagivenperceptorswitchtotheotherone.Intermsofabottom-upvs.top-downcategorisation,bothofthesestrategiesconstitutebottom-upaspectsofbistableperception,astheyemployearly,ratherlow-levelfeaturesofvision.Butvoluntarycontroloverreversalratesisparticularinterestingasatop-downfeaturei.e.whenmentaleortisusedtoinuenceperceptionandnotlow-levelphenomenalikeblinking.Letvoluntarycontrolbeunderstoodastop-downfeatureinthefollowing,excludingpotentialeectsofgazepositionorblinking.Voluntarycontrolisthenaquantiableeectofconsciouscontrolofperception.AsdetailedinSec.1.1,bistableperceptionisaveryimportantandinterestingmodelinthequesttounderstandconsciousnessbecausetwoconsciousmentalstatesareassociatedwithaconstantexternalstimulus.Theessentialpointisthattheprocessofchangingbetweenthesetwomentalstatescanatleastapprox-imatelybequantied.Onepromisingapproachtounderstandinghowtwomentalstatesexistduringunchangingstimulationistotryandexploremod-ulationofthereversalprocessbetweenthosetwostates.Thatwasexactlythegoaloftheexperimentsdescribedinthischapter.Forthatitwasimportanttoexcludebottom-upinuencesonreversalsasmuchaspossible.HeadandeyemovementscanbereducedbyusingachinrestandaxationcrossinthemiddleoftheNeckercube.Thus,theheadoftheparticipantisstabilisedandthepersoncanresttheirgazeonthecross.Moreimportantforthejudgementofbottom-upinuencesareseveralstudiesthatreportthatneithersaccades(i.e.fastmovementsofthegaze)norblinks 1AnoteworthyoutlierinthisrespectisthestudyofCarteretal.(2005),whichdemon-stratedtheextremeprolongationofdwelltimesbyTibetanBuddhistmonksinbinocularrivalry.ThesendingswillbediscussedindetailinChapter9.83 theotherhand,itsupportsthehypothesisthatpersonalityislikelytoberelatedtovoluntarycontroloverbistableperception.Similarly,HaronianandSugerman(1966)foundthatthenumberofreversalsinaholdcondi-tionfortheNeckercubecorrelatedwithtwomeasuresofeld-independence.Field-independenceisaconceptofcognitivestylethatindicateshowmuchapersonreliesontheirowninnerknowledgeandanalysiscomparedtoexternalinformation.InastudybyAydinetal.(2013)age-relateddierencesintheabilitytoinuencebistableperceptionofavase/facesstimuluswerereported.Theauthorssuggestedabnormalattentionalmechanismsasanexplanation,categorisingthendingmoreasacognitiveeectthanoneofpersonality.Asanothernding,Saueretal.(2012)showedthatexperiencedmeditators,whoshowedhighscoresinmindfulness,wheresignicantlybetteratprolongingdwelltimesoftheNeckercubethanacontrolgroup.ThisresultanditsrelationtoownndingswillbediscussedinmoredetailinChapter9.Tofurtherexplorevoluntarycontroloverbistableperceptionmorerigor-ouslywasoneoftheobjectivesoftheNC-persstudy.InthisChapter,adetailedanalysisoftheexperimentonvoluntarywillbegiven.Relationstopersonalitywereexploredwithquestionnairesonself-leadership,action-controlandtheBigFivepersonalitytraits.Self-leadershipandaction-controlwerechosenastwopsychologicalconceptsthatdescribevoluntaryhumanbehaviour.Strategiesaimingattheimprovementofmotivationandself-directioninordertoperformwellwereoperationalisedwiththeconceptofself-leadership(AndreÿenandKonradt,2007).Theconceptofvolition,ormorespecicallyaction-control,wasaddedasameasureclosertoactualac-tionperformance(Kuhl,1994).Finally,theBIG-5inventory(Körneretal.,2008)wasusedtogainabroadclassicationofpersonality.TheBigFivepersonalitytraitsareopenness,conscientiousness,extraversion,agreeable-nessandneuroticism.Aclassicationintermsofthesecategoriescouldserveasabasisforfurtherexploration,relatingfacetsofeachofthesedi-mensionstobistableperception.TheexactmethodsofthestudyofvoluntarycontroloverperceptionoftheNeckercubewillbegiveninthefollowingsection.7.2MeasuringVolition65healthyparticipants(age28.89.1years,31male)completedthisexper-iment.Itconsistedoffour3-minutetrialsofviewingtheNeckercube.The85 TodescribetheabilitytoholdperceptAthefollowingmeasuresweredened:
hA=~thA�~tneut(7.1)
hApA=~thApA�~tneut;pA(7.2)
hApB=~thApB�~tneut;pB(7.3)InanalogyfortheholdB...
hB=~thB�~tneut(7.4)
hBpA=~thBpA�~tneut;pA(7.5)
hBpB=~thBpB�~tneut;pB(7.6)...andthespeedupconditions:
sp=~tneut�~tsp(7.7)
sppA=~tneut;pa�~tsppA(7.8)
sppB=~tneut;pB�~tsppB(7.9)Thesemeasureswerecalculatedforeachparticipant.Medianvalueswereusedinallcasesasforsomeparticipantstherewasonlyasmallnumberofdatapointsperconditionduetoslowreversalbehaviour.Themeanandstandarddeviationweredeterminedoverall58participants.Foranyoftheabovemeasures,anitevalueforanyofthesemeasureswouldindicateadeviationfromtheneutralcondition.Totestforsignicanceofsuchde-viations,non-parametricone-sidedWilcoxontestswereperformedbetweenmediandwelltimesfortheneutralandthevoluntarycontrolconditions.ThesetestswerecalculatedseparatelyfordwelltimesforperceptAandBaswellasfordwelltimesofbothpercepts.Anon-parametrictestmethodwasused,asmeanormediandwelltimesusuallyarenotnormallydistributedoverparticipants.Aone-sidedtestwasusedbecauseforeachcasethehypo-thesesweredirected.All
'swereexpectedtobepositive,exceptfor
hApBand
hBpA.p-valuesoftheWilcoxontestswerecorrectedwiththefalsedis-coveryratemethod(FDR,BenjaminiandHochberg(1995);BenjaminiandYekutieli(2001))Additionally,Spearmancorrelationcoecientsbetweenmediandwelltimeintheneutralcondition,i.e.passiveperceptionoftheNeckercube,andvolun-tarycontrol,namely
hApA,
hBpBand
sp,werecalculated.Thiswasdoneinordertobetterunderstandtherelationbetweenneutralandvoluntarily87 Figure7.1:ChangesinducedbyvoluntarycontrolcomparedtoneutralobservationoftheNeckercube.Starsindicatesignicanceoftheone-sidedWilcoxontestsbetweenvoluntarycontrolandneutralconditionaftercorrectionwiththeFDRmethod::p0:05,:p0:01and:p0:001. hAhApAhApBhBhBpAhBpBspsppAsppB p 0.321
10�60.0050.210.044
10�66
10�83
10�79
10�8padj 0.323
10�60.0080.240.067
10�64
10�79
10�74
10�7Table7.1:p-valuesoftheone-sidedWilcoxontestsfordierencesinmedianvaluesforallexperimentalconditions.Thesecondrowdisplaysthep-valuesadjustedformultipletestingwiththeFDRmethod.89 ~t
hApA �0:30=
hBpB �0:32=
sp 0:74= ~t
hApA;rel �0:45=
hBpB;rel �0:49=
sp;rel 0:61=Table7.2:CorrelationcoecientsrforWilcoxontestsbetweenmeasuresdescribingneut-ralandvoluntarilycontrolledobservationoftheNeckercube.Thelefttabledescribesabsolutemeasures,whiletherightgivesrelativemeasuresgivesthecorrelationforrelativeone.Asterisksindicatep-values::p0:05,:p0:01and:p0:001;asterisksaftertheslashgivep-valuesaftercorrectionformultipletestingusingthefalsediscoveryratemethod(FDR).FortheinstructiontoholdperceptAandtoavoidB,themedianoveralldwelltimesisnotdierentfromtotheneutralcondition.ButdwelltimesforAaresignicantlylargerthancorrespondingdwelltimesintheneutralcondition.Additionally,dwelltimesforperceptBaresignicantlysmaller.IntheholdBconditionasimilareectisobserved:overalldwelltimesdonotchange,butdwelltimesforperceptBaresignicantlylargerthanthoseintheneutralcondition.DwelltimesforAareslightlysmallerthanintheneutralconditionbutnotsignicantlyso(p=0:06).TheeectofthespeedupinstructionisseeninadecreaseofdwelltimesforperspectiveAandB,andhencealsoinalldwelltimestakentogether.Eectsizesareverysimilar.TheresultsofthecorrelationtestsbetweenpassiveandvoluntarycontrolconditionsaregiveninTab.7.2.Itshowsthat
hApAand
hBpBcorrelatenegativelywithmediandwelltimesintheneutralcondition,while
spcor-relatespositivelywiththemedian.Thecorrelationsarestillsignicantaftercorrectionformultipletesting.Thesamedirectionsofcorrelationsarefoundfortherelativemeasuresofvoluntarycontrol
hApA;rel,
hBpB;reland
sp;rel,withlargercorrelationcoecientsfortheholdconditionsandasmallerforthespeedupcondition.InFig.7.2,theabsolutedecreaseindwelltimesforthespeedupconditionisplottedagainstneutralmediandwelltimesinordertoillustratetherelationmoreprecisely.Aclearincreaseof
spwiththemedianofneutraldwelltimesisseen.Ofthecorrelationcoecientsbetween
hApA,
hBpBand
sponlytheonebetween
hApAand
hBpBissignicant,withr=0:49,p0:001.The
'softheholdconditionsarenotcorrelatedtotheoneofthespeedupcondi-tion.Again,thecorrelationremainssignicantafterapplicationoftheFDRmethod.90 ditions.SimilarresultsweredescribedbyKornmeieretal.(2009).Thus,theeectsofvoluntarycontrolasdemonstratedinthecurrentexperimentareingoodagreementwiththeliterature.ThecorrelationcoecientspresentedinTab.7.2showthatvoluntarycon-trol,bothinabsoluteandinrelativeterms,isnotindependentofdwelltimesintheneutralcondition,andingwhichhasnotbeenreportedsofar.ThecorrelationsindicatethatanincreasedmediandwelltimegoeshandinhandwithdecreasedabilitytoholdperceptsAandBandanincreaseintheabilitytospeedupthereversalprocess.Thisrelationisparticularlystrongforthespeedupcondition.Here,thecoecientofdeterminationisashighasroughly55%(estimatingthecoecientofdeterminationasr2=0:742=0:55).Thus,agreatpartofthevariabilityofthesuccesstospeedupreversalscanbeexplainedbyneutraldwelltimes.Fortheholdconditions,onlyasmallpart,about10%ofthevariabilityisexplainedbytheneutraldwelltimes.Theseresultsindicatethatitseemstobecomeincreasinglydiculttopro-longdwelltimesthehigherdwelltimesareunderpassiveperception.Ontheotherhand,aspeedupofreversalistheeasierthehigherthedwelltimesare.Botheectsdonotseemtobeonlyalinearscalingeect,astheyarealsopresentwhenoneconsidersrelativechangesofthevoluntaryconditionscomparedtotheneutralcondition.Moredatapoints,especiallyforslowreversers,wouldbedesirableinordertoquantifytheexactrelationbetween~tand
sp(cf.Fig.7.2).Thecorrelationbetween
hApAand
hBpBshowsthatbothperceptsoftheNeckercubearenotonlyperceivedfordierentamountsoftimebutthatsuccessofvoluntarilycontrollingthemisdierent.Thisisreectedinthecorrelationcoecientbetween
hApAand
hBpBbeingmuchlowerthan1.IfitwasaseasytoholdperceptBasitistoholdperceptA,r1wouldbeexpected.Infact,Fig.7.1indicatesthattheabilitytoholdBisslightlylowerthantheonetoholdA(n.s.).Hence,thebiaseectpresentedinChapter6isfurthersubstantiatedbythisnding,asitnotonlymanifestsinanasymmetryinneutraldwelltimesbutalsoinanasymmetryintheabilitytoinuencepercepts.Additionalevidenceforitisaddedbythefactthat
hApAand
hBpBcorrelatewithdierentpersonalitymeasures(cf.thenextparagraphandChapter9).Duetotheexploratorynatureofthisstudy,uncorrectedp-valuesweregivenforthetestsbetweenmeasuresquantifyingvoluntarycontroloverperceptionoftheNeckercubeandpersonalityquestionnaires.Asthesecorrelationsdo93 TheresultsontheBIG-5scalesontheotherhandsuggestthattheabilitytoholdperceptAisrelatedtothepersonalitydimensionofagreeableness,indicatedbyacorrespondingcorrelation.Thescaleexpresseshowcompas-sionateandcooperativeapersonbelievesthemselvestobe.Again,andingintermsofmindfulness(Chapter9),namelyacorrelationbetween
hApAwiththeFFAsubscaleforacceptance,seemstosupportthecurrentresult.Thatconceptofacceptanceexpresseshowwellapersoncanpositivelyacceptadverseexperiencesandhaveacompassionateattitudetowardsownshort-comingsandthoseofothers.Thisiscertainlyaverygoodbasisforbeingagreeabletowardsotherpeople.Hence,arelationbetweentheabilitytoholdperceptAandagreeablenessseemsplausible.Thus,itispromisingtofurtherexploreitusingcorrespondingpersonalityconcepts.Furthermore,theabilitytospeedupperception,
spiscorrelatedpositivelywithneuroticismandnegativelywithconscientiousness.Conscientiousnessindicatesecient,organisedbehaviourandself-discipline,i.e.thedesiretodoataskwell.Neuroticismontheotherhanddenotesemotionallability,thetendencytoexperienceunpleasantemotionsandalowimpulsecontrol.Asthecarefulconductionofataskusuallytakesacertainamountoftimeandcannotbespeduparbitrarily,itmakessensethatpeoplewithhighscoresinconscientiousnesscanonlyspeedupperceptualreversalstoacertaindegree.Lowimpulsecontrolandemotionalinstabilitymightberelatedtopercep-tualinstabilityandhenceananitytospeedupreversals.Furthermore,aprincipalcomponentanalysis(PCA)ofthe5scoresoftheBIG-5ques-tionnaireforallparticipantsshowedthatconscientiousnessandneuroticismpointinalmostoppositedirectionsinthespacespannedbythetwoprinciplecomponents.Hence,theoppositecorrelationsoftheconscientiousnessandneuroticismscoreswiththeabilitytospeedupreversalsarereasonable.Finally,thelackofcorrelationsbetweenthescoresoftheHAKEMP-90andtheabilitytoinuencebistableperceptionshowsthataction-controlasahigh-levelpsychologicalconceptisnotrelatedtocontroloverperceptionoftheNeckercube.Inmoredetail,thismeansthatnoneofthethreesubscalesofthequestionnaire,namelyorientationforactionafterfailure,orientationforactioninactionplanningandorientationforactionduringaction,over-lapsignicantlywiththeabilitytoholdorspeedupthereversalprocessoftheNeckercube.ThismightbeexplainedbythefactthattheHAKEMP-90operationalisesmoregeneralschemesofaction,i.e.howimmersedapersongenerallyisinataskorhowtheycopewithrepeatedfailure.Thesescalesmightbetooabstractandgeneralastocapturetheabilitytoinuence95 8.Perception&PersonalityInthelastchapterresultsontherelationbetweenvoluntarycontroloverbistableperceptionoftheNeckercubeandpersonalitywerereported.Thischapter,ontheotherhand,isfocusedonrelationsbetweenpersonalityanddwelltimesforpassiveperceptionoftheambiguousgure.8.1StudiesLinkingBistabilityandPersonalityIncontrasttovoluntarycontroloverbistableperception,therearequiteafewreportsrelatingneutralperceptionoftheNeckercubewithpersonalitytraits.Thesevaryintheirmethodologicalapproachandquality.Inmanystudies,especiallytheolderones,itwasnotdwelltimesthatweremeasuredbutonlythenumberofreversals.Hence,inthefollowingshortoverviewoftheliterature,mostlynumberofreversalswillbecitedasmeasureofbistableperception.Mostlyquestionnaireswereusedinordertoassesspersonalitytraits.AccordingtoBeer(1989)thenumberofreversalsfortheNeckercubeisnotrelatedtotheconceptofambiguitytolerance,butcorrelatesnegativelywithrigidity.KiddandCherymisin(1965)onlyconsideredtheveryrstreversaltimefordierentambiguousguresandfoundalsothatlongtimesgohandinhandwithhighvaluesinrigidityandfurthermoreinanxietyandeld-dependence.Field-dependenceisaconceptofcognitivestyle,indicat-inghowmuchapersonreliesonexternalinformationincontrasttotheirinnerknowledgeandanalysis.Itisimportanttonotethattestingfortherstreversalofanambiguousgureisstronglydependentoninstruction,asbeinguninformedaboutreversibilityleadstoanabsenceofreversalsinmostcases(RockandMitchener,1992).AsKiddandCherymisin(1965)donotgivedetailsofeitherinstructionsortheusedambiguousgures,theirndingsshouldbeusedcarefully.FrederiksenandGuilford(1934)found97 usedhereagaininordertoprobeforpotentialrelationstoneutralperceptionoftheNeckercube.Thefollowingsectionwilldetailthemethodsappliedtothatend.8.2OperationalisationofPersonalityTraitsPotentiallinksbetweenpersonalityandperceptionwereexploredbytestingforcorrelationsbetweenquestionnairescoresandmeasuresdescribingpass-ive,bistableperceptionoftheNeckercube.Forthat,questionnairedataof28participants(meanage30:610:9years,12male)fortheHAKEMP-90,theSTAI,theBSSSandtheambiguitytolerancequestionnaireswasused.FortheRSLQ-DandtheBIG-5questionnaires,datafrom58participants(29:09:5years,27male)wasused.ThedetailsoftheexperimentonbistableperceptionoftheNeckercubearegiveninSec.3.3.Note,thatfail-uretounderstandtheinstructionsforthisexperimentresultedinthelowernumberofusabledatasets(28and58,respectively)comparedtotheoverallavailabledata(33and65,respectively).FortheHAKEMP-90,threescoreswerecalculated,namelythescorefororientationforactionafterfailure(HandlungsorientierungnachMisser-folg,HOM),thescorefororientationforactioninactionplanning(Hand-lungsorientierungbeiderHandlungsplanung,HOP)andtheonefororient-ationforactionduringaction(HandlungsorientierungbeiderTätigkeit-sausführung,HOT).TheSTAIhastwosubscores,namelythescoreforstateanxietyandtheonefortraitanxiety.BSSSandambiguitytolerance(Un-gewissheitstoleranz,UGT)onlyhaveonescoreeach.Inordertodescribebistableperception,meanandmediandwelltimeswerecalculatedforeachparticipant.Furthermore,theparametersofthelognor-maltsaswellasitsmodeandvariancewereusedasmeasurestooperation-alisebistableperception.Forallcombinationsbetweenthequestionnairescoresandthemeasuresofbistableperception,Spearmancorrelationcoecientsandp-valueswerecal-culated.8.3ResultsNoneofthescoresforaction-control,anxiety,sensationseekingandambi-guitytolerancecorrelatessignicantlywithmeasuresdescribingthecentral99 Cherymisin(1965)wasnotpossible.Secondly,heretheSTAIquestionnairewasusedwhiletheauthorsofthecitedstudyusedtheTaylorManifestAnxi-etyScale.Asastudydirectlycomparingthesequestionnairesseemsnottobeavailable,itcannotbejudgedhowthetwoscalesdier.Therearenostudiesthattestedforpotentialrelationsbetweentheconceptofsensationseekingandbistableperception.Thus,thecurrentresultsdemon-strateforthersttimethatsensationseekingbehaviourisnotrelatedtoperceptionoftheNeckercube.Becauseofthebrevityoftheemployedscale,ananalysisintermsofthefoursub-traitsofsensationseekingwasnotper-formed.Fromthecurrentresults,itcannotbeexcludedthatoneofthesemightberelatedtobistableperception.Thelackofacorrelationbetweenambiguitytoleranceandbistablepercep-tionfoundbyBeer(1989)wasconrmedinthecurrentstudy.Thus,itcanbeconcludedthatalsowithadierentquestionnaireandstate-of-the-artmeas-urementandanalysismethodsforbistableperceptionthereisnorelationbetweenambiguitytoleranceandpassivebistableperception.ThelackofcorrelationsbetweenmeasuresdescribingbistableperceptionoftheNeckercubeandanxiety,sensationseekingandambiguitytoleranceprovideusefulindicationsinthesearchforpersonalityaspectsortraitsthatarerelatedtobistableperception.Thesendingsnarrowdownthepotentialcandidatesforrelatedconceptsandhenceindirectlyadvancethesearchforthem.Thus,theresultspresentedabovecanbeofuseforfurtherinvestiga-tionsoftheperception-personalityconnection.Regardingthecorrelationsthatwereindeedfound,similarlyasinChapter7,uncorrectedp-valueswerereportedintheprevioussection,withresultsnotbeingsignicantanymoreafterrigorouscorrectionusingtheFDRmethod.Hence,specialcareshouldbetakenintheinterpretationofthefollowingresultswhichcanserveasabasisforfurtherresearch.Thethreecorrelationstotheself-leadershipscaleforself-rewardaredepend-enttoahighdegree.Boththeparameterandthevariancevarlognarestronglycorrelatedwiththemeandwelltimes:r=0:97and0.88,respect-ively(p0:001inbothcases).Thus,thecorrelationsofmeandwelltimesandarealmostequivalent,showingthatastrongtendencytowardsself-rewardandtheapplicationofself-rewardingbehaviourgohandinhandwithshortdwelltimes.Thecorrelationtovarlognindicatesfurthermore,thatinthiscasedwelltimesdonotvarysomuch.Thatthreecorrelationspointal-mostinthesamedirectionisagoodindicatorofthereliabilityoftheresult.Comparedtopersonalitytraits,self-leadershipanditssubscalesisamore101 thermore,conscientiousnessseemstoberelatedtothemodeofthelognormaldistribution.Theseresultsfurthervalidateandadvancetheundertakingofexplaininginter-individualvariationsofdwelltimesintermsofpersonalitytraits.103 Bothstudiesindicatethatthereisindeedarelationbetweenmindfulnessandvisualperception.Goalofthepresentstudywastoexplorewhetherthesedierencescouldalsobedetectedinanunbiasedgroupofparticipants,i.e.agroupthatwasnotscreenedformeditationexperience.AsindicatedbytheresultsofCarteretal.(2005)thereareseveralaspectsnotonlyofmeditationbutalsoofmindfulness.Bishopetal.(2004)gaveanoperationaldenitionformindfulnessinwhichtheydistinguishedtwocomponents:(1)self-regulationofattentionsothatitismaintainedonim-mediateexperienceand(2)anorientationtowardsone'sownexperiencescharacterisedbycuriosity,opennessandacceptance.Bergomietal.(2012)evendiscernednineaspectsofmindfulnessbasedontheirsummaryofexistingmindfulnessscalesandincorporatedthemintotheCHIME-questionnaire.BuildingontheseresultstheauthorscreatedtheComprehensiveInventoryofMindfulnessExperience,CHIME,incorporatingthefollowingeightas-pectsofmindfulnesswithoutrelyingontechnicalexpressionsofmeditationorBhuddism:(1)awarenesstowardsinternalexperiences(innerawareness),(2)awarenesstowardsexternalexperiences(outerawareness),(3)actingwithawareness(actingwithawareness),(4)acceptingandnon-judgementalori-entation(acceptance),(5)decenteringandnonreactivity(decentering),(6)opennesstoexperiences(openness),(7)relativityofthoughts(relativity)and(8)insightfulunderstanding(insight).Anothermindfulnessquestionnaire,theFreiburgMindfulnessInventory(FMI),hasbeendesignedbyWalachetal.(2006),which,inashortversiondiscernsacceptanceandpresencealsowithoutrequiringparticipants'knowledgeoftheBhuddistbackgroundofmindfulness.BoththeCHIMEandtheFMIquestionnairewereusedinthecurrentstudytoexploremindfulnessinanunbiasedgroupofparticipants.Forthat,thepropertyofbothquestionnairesthattheydonotutilisetechnicaltermsofmeditationorBhuddismwasveryimportant.Itwashypothesisedthatneut-raldwelltimeswouldnotbecorrelatedtomindfulnessscores,asSaueretal.(2012)didnotndacorrespondingdierencebetweenmeditationexpertsandnon-meditators.ThesamestudyandtheoneofCarteretal.(2005),ontheotherhand,suggestedthattheremightberelationsbetweenmind-fulnessandvoluntarycontroloverperceptionoftheNeckercube.Thesetwohypothesesweretestedinthefollowingway.105 ingwascorrectedusingtheFalseDiscoveryRatemethod(FDR).Duetotheexploratorynatureofthestudy,mostlyuncorrectedp-valueswillbegiveninthenextsection,though.9.3ResultsBetweenthemeasuresdescribingneutralbistableperceptionandthemindful-nesssubscales,thereisonlyonesignicantcorrelation.Namely,theCHIMEsubscaleofrelativityofthoughtscorrelatesnegativelywiththeparameterofthelognormaldistribution(r=�0:27;p=0:04).Eachofthethreemeasuresofvoluntarycontroloverbistableperceptioncorrelatessignicantlywithatleastonemindfulnesssubscale.Uncorrectedp-valuesforthesecorrelationsaregivenhere.TheabilitytoholdperceptA,operationalisedby
hApA,isrelatedtothesubscaleforacceptanceoftheFMI(r=0:33;p=0:013).TheabilitytoholdperceptB,
hBpB,iscorrelatedtotheCHIMEsubscalesforawarenesstowardsexternalexperiences(r=0:27;p=0:042)andrelativityofthoughts(r=0:42;p=0:001).Theabilitytospeedupperceptualreversals,
sp,showsacorrelationtotheCHIMEsubscaleforawarenesstowardsexternalexperiences(r=0:32;p=0:014).1ThesecorrelationsdonotremainsignicantaftercorrectionusingtheFDRmethod.TheFMIcorrelateswithsevenoftheeightsubscalesofCHIME.AtablewithcorrelationcoecientsisgiveninTab.9.1.FheoverallFMIscorehasthehighestcorrelationswithacceptance,decenteringandinsight.Theaccept-ancesubscaleofFMIhasthehighestcorrelationsforthesamefourCHIMEsubscales.TheFMIpresencesubscalecorrelateshighestwithdecentering,outerawareness,acceptance,innerawarenessandinsight.TheopennesssubscaleofCHIMEshowsnosignicantcorrelationstotheFMI.Allcorrel-ationsremainsignicantaftercorrectionformultipletestingusingtheFDRmethod. 1Notethatverysimilarresultswereobtainedwhenusingrelativeinsteadofabsolutechangesofdwelltimes:
hApA;relandacceptance:r=0:32;p=0:015,
hBpB;relandrelativityofthoughts:r=0:35;p=0:008and
sp;relandawarenesstowardsexternalex-periences:r=0:33;p=0:011.Awarenesstowardsexternalexperiencesdoesnotcorrelatesignicantlyto
hBpB;rel(r=0:20;p=0:13).107 encesandrelativityofthoughtsarerelevant.Theawarnesstowardsexternalexperiencesisfurthermorerelatedtotheabilitytospeedupperception.Theseresultsshow(1)thatrelationsbetweenmindfulnessandperceptioncannotonlybedetectedbetweenexpertsandlaypersonsbutalsoviatheinter-individualdierencesinanunbiasedsample,(2)thattheasymmetryintheabilitytoholdperceptAcomparedtoholdperceptBasdetailedinChapter7isalsoreectedindierentaspectsofmindfulnessand(3)thatalsotheabilitytospeedupperceptioniscorrelatedtooneaspectofmind-fulness.Thesendingsshallbediscussedinmoredetailinthefollowing.Thecor-relationbetweenacceptanceasoperationalisedbytheFMIand
hApAisingoodagreementwiththecorrelationbetweenagreeablenessand
hApAdescribedinChapter8.IntheFMI,acceptanceisunderstoodasanon-judgemental,compassionatestancetowardsoneselfandone'ssurrounding.ExamplesofstatementsusedwhichareratedwithaLikert-typescaleareIchkanndarüberlächeln,wennichsehe,wieichmirmanchmaldasLebenschwermache.(IcansmilewhenIrealisehowIsometimesmakemylifedicult.)orIchnehmeunangenehmeErfahrungenan.(Iacceptuncom-fortableexperiences.).AgreeablenessintheBIG-5questionnaireincludesstatementslikeIchbekommehäugerStreitmitmeinerFamilieundmeinenKollegen.(Ioftenarguewithmyfamilyandcolleagues.,negativelycodedstatement)andIchversuchestetsrücksichtsvollundsensibelzuhandeln.(Ialwaystrytoactconsideratelyandsensitively.).Itisreasonabletoas-sumethatanacceptingstateofmindisconducivetobeingagreeablewhileprobablyanagreeablepersonwillalsomanifestthecharacteristicsofaccept-ance.Alsofor
hBpBwendsupportingcorrelationsintermsofmindfulnessandpersonality.
hBpBcorrelateswithbothrelativityofthoughtsandtheeval-uatingbeliefsandassumptionsscaleofRSLQ-D(cf.Chaper7).TheformerscalepostsitemslikeEsistmirimAlltagbewusst,dasssicheigeneMein-ungen,dieichzurZeitsehrernstnehme,deutlichverändernkönnen.(InmyeverydaylifeIamawarethatmyownbeliefswhichItakeveryseri-ouslyatthemomentcanchangeconsiderably.)andthelatteritemslikeInSituationen,indenenichaufProblemetree,prüfeich,obmeineÜberzeu-gungenangemessensind.(InsituationsinwhichIencounterproblemsIcheckwhethermyopinionsareappropriate.).Bothscalesqueryhowcriticalandexibleapersonoperateswiththeirbeliefs.Thus,thecorrelationsof
hApAand
hBpBwithmindfulnessandpersonality109 andfollowfastreversalsbetweentheperceptsoftheNeckercube.Thus,theresultspresentedhereaddontothendingsbySaueretal.(2012)inthattheyshowsimilarresultsinanunbiasedgroupofparticipantsandfurthermoreareabletoreecttheasymmetryinperceptionandmindfulnessbetweenthetwoperceptsoftheNeckercube.Additionally,theydemonstratearelationofmindfulnesstotheabilitytospeedupreversalsandingwhichhasnotbeenreportedsofar.Finally,theresultsonthecorrelationsbetweenthetwomindfulnessquestion-nairesusedheredemonstratehowtheyarerelatedtoeachother.AsshowninTab.9.1theoverallscoreoftheFMIhasthehighestloadingsonaccept-ance,decentering,insightandactingwithawareness,indecreasingstrengthofcorrelation.Thus,theFMIisamixedconceptualisationofthesefouraspectsofmindfulness.TheacceptancesubscaleoftheFMIalsohashighloadingsonthesamefouraspectsandverylowonesontheotheraspectsoftheCHIME.Buthere,thesubscaleforacceptanceoftheCHIMEclearlyhasthehighestloading(r=0:73).Thisisconrmsthatbothquestionnairesaimatthesameunderstandingofacceptance.ForthepresencesubscaleoftheFMI,therelationisnotasclear.Here,thestrongestcorrelationiswiththeCHIMEsubscalefordecentering(r=0:53).Buttheothercorrelationcoe-cientsarenotsmall,excepttoonetothescaleforaopenness.Nevertheless,asthegreatestgapisbetweenstrongestandsecondstrongestcorrelation,itcanbesaidthatpresenceintheFMIloadsstrongestondecenteringbuthassignicantloadingsonouterawareness,acceptance,innerawarenessandinsight.Thus,theFMIpresencescaleismoremixedintermsoftheaspectsdistinguishedwiththeCHIMEquestionnaire.Oneshouldnotethatusuallyahighernumberofparticipantsisdesirableforthevalidationorcompar-isonofanewquestionnaire,hencetheresultspresentedherearemoreofanapproximatenature.111 encompassesmultiplesecondsoverwhichcognitiveandemotionalprocessesaremaintained.Forbistableperceptionboththefunctionalmomentandtheexperiencedmomentmightbeimportant.Infact,thesetwotimescaleshavebeenlinkedtheoreticallyintheNecker-Zenomodel(Atmanspacheretal.,2008,2004).Repeatedorperiodicperceptualprocessesintherangeofthefunctionalmomentmightbeinvolvedintriggeringperceptualreversals.AsimpleGo/Nogoreactiontimeparadigmwasusedinordertoprobethistem-poralrange.Atemporalintegrationtask(Szelagetal.,1996)basedontheperceptionofmetronomebeatsaimedatthesecondtimerange,namelytheexperiencedmoment.Thisisalsotherange,intowhichNeckercubereversaltimesfall.Divertingattentionhasbeenshowntoslowdownperceptionofpercep-tual(ReisbergandO'Shaughnessy,1984)andbinocularrivalry(Alaisetal.,2010).ThelatterstudycomparedtheeectforbinocularwiththatfortheNeckercube.Itwasfoundthatattentionhadagreatereectforperceptualthanforbinocularrivalry.Thisindicatesahigh-level,top-downinuenceofattention.Thereareatleastthreedistinctdierencesbetweenattentionalprocessesandbistableperception,though,aspointedoutbyLeopoldandLogothetis(1999).First,voluntarycontroloverattentionislargerthanthatoverbistableperception.Secondly,attentioncanenhanceprocessingofavisualobjectwhileinbistableperceptiontheperceptioncanchangecom-pletely.Third,attentionalshiftscanbemuchfasterthantheshortestdwelltimes.Onthebasisofthesendings,itwasthegoaltondoutinwhichwayinter-individualdierencesinattentionwererelatedtodierencesinbistableperceptionoftheNeckercube.Acommonattentiontask,namelythed2task(Brickenkamp,2002),wasusedtooperationaliseattention.Thistaskcapturesseveralaspectsofattentioninavisualparadigmandisthuswellsuitedtocomparetobistablevisualperception.10.2ExploringLinksinTimeScalesAsdescribedabove,temporalprocessingwasoperationalisedviaanattentiontask,areactiontimetaskandatemporalintegrationtask.AllthreetasksaswellasthevoluntarycontrolexperimentforbistableperceptionoftheNeckercube(asdetailedinChapter7)werecompletedby65participants.Datasetsof58participants(29:09:5years,27male)wereusedforfurtheranalysis,with7beingexcludedbecauseofinsucientdatafortheNecker113 Figure10.1:StimulifortheGo/Nogotask.Participantswereaskedtoreactasquicklyaspossibletotherightstimulusandnotatalltotheleftone.so,theywereinstructedtoaccentuatementallyeverysecond,third,fourthetc.beat,thuscreatingasubjectiverhythmforthemselves,e.g.1-2-3-4,1-2-3-4,etc.Participantsreportedhowmanyclickstheycouldintegrateintooneperceptualunitbypressingthecorrespondingnumberkeyonthekeyboardofacomputeraftereachsequence.Eachofthe5dierentfrequencieswerepresented5timesinrandomisedorder.Thisresultedinatotalof25tri-als.Subsequenttrialswereinterruptedbybreaksof6stopreventcarry-overeects.Abreakoftwominuteswasgivenafterhalfofthetrials.Atthebeginningofthetask,participantsweregivenafewpracticetrialsinordertofamiliarisethemselveswithit.Forallmetronomefrequencies,medianintegrationtimeswerecalculatedforeachparticipant.Theintegrationtimeisthenumberofbeatsaparticipantintegratedintooneunitmultipliedbythetimebetweensuccessivebeats(whichistheinverseofthemetronomefrequency).Here,themedianwasused,andnotthemean,asthereisonlyasmallnumberofdatapointsperparticipantandfrequency,namely5,makingthemedianthemorerobustmeasure.Additionally,therangeofintegrationtimeswascalculatedperparticipantasthedierencebetweenthemedianintegrationtimefor1Hzand5Hz.Finally,meanandstandarddeviationformedianvaluesoverallparticipantswerecalculatedandplottedvs.frequency,inordertocomparethecurrentresultswiththoseofSzelagetal.(1996)whoshowedasimilarplot.Themeasuresfromthesethreetasksweretestedforcorrelationswithmeas-uresofneutralandvoluntarilycontrolledbistableperception,usingSpear-mancorrelationcoecients.Fortheneutralcondition,theseweremeanandmediandwelltimes,aswellasparameters,modeandvarianceofthelognor-115 Figure10.2:Meanintegrationtimeswithstandarddeviations,calculatedfrommedianintegrationtimesof58individualparticipants,asafunctionoffrequency.117 androbust.First,itshouldbenotedthatthesecorrelationsarehighlyde-pendent.Toillustratethat,Spearmancorrelationcoecientsbetweenmeandwelltime,t,andtheothermeasuresofneutralbistableperceptionwerecalculated:mediandwelltime,~t:r=0:97,modeoflognormalt,t0;logn:r=0:88,parameteroflognormalt:r=0:98andvarianceofthelognor-malt,varlogn:r=0:89.Allthesecorrelationsarehighlysignicant.Hence,thecorrelationsofthenumberofattendedtargetswiththelistedmeasuresallpointinthesamedirection.Theyshowthatshortdwelltimes,andthecor-respondinglysmallvariance,gohandinhandwithahighrateofprocessinglettersinthed2task.Thesimilarcorrelationsandtheverylowp-valuesindicateaveryrobustnding.Thespeedofprocessingvisualinformationcouldtobetheconceptconnect-ingbothprocesses.Itcanbeunderstoodwithinmodelsofbistabilitythatincorporateadaptation.Inthesemodels,withcontinuoussensualinput,cer-tainneuralcyclesareexciteduntiltheysatiateatwhichpointaperceptualreversaltakesplace.Formoredetailsonthesemodelscf.Sec.2.1.Whenprocessingspeedishighinanindividual,satiation,andhencetheperceptualinversion,willbeverylikelytooccurearlier.Note,thatatthispointnohypothesisisgivenofhowthisconceptofprocessingspeedwouldbeincor-poratedonaneurallevel.Ontheotherhand,ahighprocessingspeedwillalsoleadtoahighnumberoftargetsattendedinthed2task,irrespectiveoferrorrateinthattask.Thus,thisndingpointstowardsacommontem-poralstructureinvisualprocessing.Thisisanimportantndingasithelpstocreateaconceptualbridgebetweenbistableperceptionandtemporalpro-cessing.Thelackofacorrelationbetweenthemeasuresofbistableperceptionandthemeanreactiontimeprovidesanotherpieceofinformationforrelatingbistableperceptiontotemporalprocessingingeneral.ItindicatesthatperceptionoftheNeckercubedoesnotusethesametemporalprocessingstructuresasasimplereactiontimetaskatleastnottoalargeextend.Inotherwords,themechanismsthatenableapersontoreactquicklydonotmakethemswitchmoreoftenorlessforthatmatter.Thendingsontemporalintegration,ontheotherhand,suggestthattheremightbeanotherrelationtobistableperception.First,thetemporalinteg-rationvs.frequencyplot(Fig.10.2)demonstratesthatthetaskwascorrectlyreproducedafterSzelagetal.(1996).Theintegrationtimesforthedierentfrequenciesareverysimilartothoseinthecitedstudy.Standarddeviationsinthecurrentplotareconsiderablysmallerthanintheoriginalarticlewhich118 Therstquestionislikelytoputthemetronometaskintoaparameterregionwherethetimescalesmatchmuchbetterwiththoseofbistableperception.At1Hztherelevanttimescaleofthemetronometaskis1s,at2Hzitis0:5sandsoforth.I.e.withincreasingmetronomefrequencytheindividualtemporalunits,i.e.beats,getsmallerandsmallercomparedtothemeanre-versaltimes.Evenforveryfastreversers,theseareusuallynotsmallerthanroughly1:5s.Thus,itmightbethatforlowerfrequencies,i.e.largerdur-ationsbetweenbeats,thecorrelationstodwelltimeswillbecomestronger.Metronomefrequenciesof0.5,0.25and0:125Hzwouldbegoodchoicesforthisapproach,probingtimescalesof2,4and8s,respectively,thusaccessingtypicaldwelltimeranges.Ifthecorrelationsfoundinthecurrentstudywasconrmed,itwouldbeinterestingtogetahighernumberofdatapointsatonefrequencyforwhichintegrationtimescorrelatestronglywithdwelltimesandstudythedistributionoftheseintegrationtimes.Maybesimilaritiesorcleardierencestothedistributionofdwelltimescouldbefoundthatwouldallowforacomparisonofthetwoprocesses.Inconclusion,theresultspresentedinthischapteridentiedtwotemporalconceptsthatarerelatedtoandseemtoplayaroleinbistableperceptionoftheNeckercube:processingspeedandtemporalintegration.Whileahighprocessingspeedgoeshandinhandwithshortdwelltimes,ahighdegreeoftemporalintegrationco-occurswithlongdwelltimes.Reactiontime,ontheotherhand,seemstobebasedmainlyonotherprocessesthanthoseinvolvedinbistableperception.120 didnotndasignicantdierencebetweenacousticandvisualnormaliseddwelltimeswithaKolmogorov-Smirnovtest.Inthesamestudy,volun-tarycontroloverbistableperceptsinbothmodalitieswasquantied.Againnocorrelationwasfoundbetweenthetwocategories.Thesendingswereinterpretedasanindicationthatbothformsofbistableperceptionsharecommonprinciples,whichareimplementedatleastpartlyindependentlyforbothmodalities.Presentingauditorystreamingstimulisimultaneouslywithvisualplaidsorapparentmotionstimuli,Hupéetal.(2008)foundthatper-ceptualswitchesco-occurindependentlyinbothmodalities.Kondoetal.(2012)comparedthenumberofswitchesforbistableperceptionoftheNeckercubetobothauditorystreamingandtheverbaltransformationeect,butdidnotconsiderthedistributionofdwelltimes.Theyfoundsig-nicantcorrelationsbetweenvisualandacousticbistabilitywithcorrelationcoecientsofroughly0.30.Furthermore,signicantcorrelationsofreversalsforvisualplaidstoauditorystreamingwerefound(0:30r0:58).Prob-ably,theinteractionsweresignicantinthisstudybutnotinPressnitzerandHupé(2006)becauseofthehighernumberofparticipantsintheformer.Hereadetailedexaminationofthetemporaldynamicsofverbaltransform-ationforthesyllablepairauandgenutlilisedinRadilovaetal.(1990)willbepresented.Furthermore,comparisonstobistableperceptionoftheNeckercubewillbemade.11.1.2MethodsTheverbaltransformationeectwasstudiedusingaloopofthecomputer-generatedsyllablesauandgen(Radilovaetal.,1990).The23GermanspeakingparticipantsoftheNC-diststudyheardthisloopviatwocomputerloudspeakersfor3minutes.Adescriptionoftheacoustictransformationeectwasgivenbeforehandbutnotrainingsession.Inordertominimiseapotentialbiastowardsoneortheotherword,thesyllablesequencewasslowlyfadedin,i.e.thevolumeofthestimuluswasincreasedfrominaudibletoacomfortablelevel.ParticipantswereaskedtopassivelylistentothesyllablesequenceandindicateperceptualreversalsfromAugentogenauandvice-versawithtwobuttonsinasimilarwayasintheexperimentonvisualbistableperceptionoftheNeckercube.Computer-generatedstimuliwereusedinordertokeepthestimulusemotionallyneutral.Thersthalfminuteofrecordeddwelltimeswasdiscardedinordertoac-countforpossibleadaptationinthebeginningofthepresentation,asno122 Figure11.1:BoxplotsofSSEandpKSfor21observersoftheaugen/genauloopasmeasuresofgoodnessoftforallconsidereddistributions.Forthesumofsquarederror(SSE,toppanel),asmallnumbermeanagoodt.ForpKS(bottompanel),avaluecloseto1indicatesagoodt.124 currentstudy,though,thecorrelationcoecientsareessentiallyzero.ThereweresomedierencesintheexperimentalsetupontheNeckercubethatmightberesponsibleforthesedierences.Kondoetal.(2012)usedasinglebuttondesignforreporting,whichinprincipleshouldnotinuencetheres-ults.Furthermore,severalshorttrialswereusedleadingtoatotallengthofobservationofmorethandoublethetimeusedinthecurrentstudy.Thevisualanglewasthesameinbothstudies.EventhoughprobablynopracticesessionswereusedinthestudyofKondoetal.(2012),itisnotlikelythatshortadaptationeectsinthebeginningofthemeasurementwouldaccountforthedierentresultstothecurrentstudy,astheoverallmeasurementperiodwasverylongsothatapotentialinitialadaptationintheexperimentofKondoandco-workersshouldbeleveledoutbythelaterdata.Lastly,fortheacousticexperimetn,theverbaltransformationstimuliweredierentfromthecurrentone.Mostofthedierentverbalinterpretationsthesestim-ulicouldtakewerewordsfromtheJapaneselanguage.Thisisprobablythemostpronounceddierenceinexperimentaldesignbetweenbothstudiesandmightinitselfaccountforthedierencesinresults.Furthermore,thereisalsoamarkeddierenceinanalysisofdwelltimedata.Kondoetal.(2012)excludedthedataof8%oftheparticipantsbecausethesereportedalargenumberofreversalsinparticularfortheverbaltransforma-tioneect.Thismighthaveastrongeectonthecorrelationsbetweendwelltimesinbothmodalities.Nosuchcorrectionwasperformedinthecurrentstudy.Inconclusion,inthecurrentstudyitwasshownthatthelognormaldistri-butionproducesthebestcombinationoftqualityanddescriptivenessfortheverbaltransformationeectdwelltimedataoftheAugen/genaustimu-lus.Lognormalrateandgammaratedistributionsyieldgoodandacceptablets,respectively.ThisshowsaclearsimilaritytobistableperceptionoftheNeckercube,suggestingacertainoverlapintheprocessesinvolvedinbistableperceptionforthevisualandtheacousticmodality.Partsoftheprocessingleadingtoperceptualreversalsareverylikelytooccurindependentlyforbothmodalities,though,astherewasnocorrelationbetweendwelltimesforthevisualandacousticmodality.126 32participantscompletedtheexperimentonvoluntarycontroloverbistableperception,aswellastwotestsonworkingmemory:thebackwarddigitspantaskandthereadingspantask.AsdescribedindetailinChapter7,thebistableperceptionexperimentcomprisedoffourconditions,eachthreeminuteslong,inwhichparticipantsindicatedperceptualreversalsofaNeckercubeusingtwobuttons.Therstconditionwasalwaystheneutralone,whereparticipantswereaskedtoobservetheNeckercubepassively.Intheotherthreeconditions,participantswereinstructedtotryandholdoneper-spectiveoftheNeckercubeandavoidtheother(implementedinbothpossiblepermutations)ortospeedupperceptualreversalsasmuchaspossible.BackwarddigitspanandreadingspantaskswereimplementedasinOberaueretal.(2000).Thebackwarddigitspantaskcomprisedofdierentseriesofdigitspresentedonacomputerscreen.Participantswereaskedtorepeateachseriesinreversedorderbytypingitsdigitsonthekeyboardattheendoftheseries.Eachdigitofaserieswaspresentedfor1000ms.Twoserieswiththreeandfourdigitsservedasapracticesession.Afterthat,fteenserieswerepresented,startingwithfourdigitsandincreasingbyonedigiteverythreeseriesuptoeightdigits.Thetotalnumberofcorrectdigitsrecalledfromallfteenserieswascalculatedasanoverallscore.Inthereadingspantask,whichwasbasedonDanemanandCarpenter(1980),severalseriesofsentenceswerepresentedonthecomputerscreen.Eachsen-tencewasdisplayedfor3sandwasfollowedbya1sinter-stimulus-intervaltillthenextsentenceappeared.Participantswereaskedtorateeachsentenceastrueorfalseduringthecorrespondingfour-second-interval.Attheendofeachseries,participantswereaskedtorecallthelastwordofeachsentenceintheseriesintheorderofpresentation.Thereweretwopracticeseriesandfteentestseries.Theexperimentstartedwiththreesentencesperseriesandincreasedbyonesentenceaftereverythirdseries,upaserieslengthofsevensentences.Thesentencesusedabidedbythefollowingcriteria:shortandsyntacticallysimple,triviallytrueorfalseandthelastwordbeingafamiliarnounoflessthanfoursyllables.Ascorewascalculatedinthesamewayasforthebackwarddigitspantask.Forthecorrelationtests,meanandmediandwelltimeswerecalculatedfromdataoftheneutralconditionoftheNeckercubeexperiment,aswellasmode,distributionparametersandvariancesofthelognormaldwelltimet.Fur-thermore,thethreemeasuresquantifyingvoluntarycontroloverperceptionoftheNeckercubedescribedinChapter7wereused:
hApA,
hBpBand
sp.Afterdwelltimedatapreparation(cf.Sec.3.4),datasetsof28participants128 12.Bistabilitywithin3s?Pöppel(1997)proposedalow-frequencymechanismcreatingperceptualunitsofabout3s.Theauthorsuggestedthatthisbindingmechanismwasoperat-ive,amongothersituations,inbistableperceptionandthatthedwelltimeswereexamplesoftheresultantperceptualunits.Inthefollowing,itwillbearguedandsupportedwithempiricaldatathatthisdescriptionneglectstwoimportantcharacteristicsofdwelltimesinbistableperception,namely(1)theirintra-individualand(2)theirinter-individualvariation.Thisisthecaseforbothvisualandacousticbistability.Theproposedbindingmechanismishencenotwellsuitedtodescribebistableperceptionduetoitslacktocapturethestochasticnatureofbistableperception.InPöppel(1997)theauthorstated:Spontaneousalterationratesofam-biguousguressupportthenotionoftemporalintegration.Ifstimulicanbeperceivedwithtwoperspectives(forexample,theNeckercube[...]),thereisanautomaticshiftofperceptualcontentafter3s.[...]Suchaperceptualshiftalsooccurswheninterpretingambiguousauditorymaterial,suchasthephonemesequenceCU-BA-CU,whereonehearseitherCUBAorBACU.Pöppelpositedalow-frequencymechanism[that]bindssuccessiveeventsupto3sintoperceptualunits(Pöppel(1997)).Heproposedthatthetimebetweensuccessiveperceptualswitchesinbistableperceptionisanexampleofsuchaperceptualunit.Thisstatementneglectsthecharacteristicintra-individualandinter-individualvariationofdwelltimes.AsdescribedinChapters1and4,intra-individualvariationisanexpressionofthestochasticnatureofbistableperception:perceptualswitchesbetweenthetwoalternativesofabistablestimulusdonotoccuralwaysafterthesametimeinterval.Ratherdwelltimesshowanunimodalstatisticaldistributionwithanitewidth.Boththelognormalrate,thelognormalandthegammaratedistributionseemtotwelltoempiricaldata(Brascampetal.(2005),Zhouetal.(2004),aswellasChapter4).Fig.12.1showsanexampleof130 13.Summary&ConclusionTwoempiricalpsychophysicalstudieswerepresentedthataimedatimprov-ingthedescriptionofthetemporaldynamicsofbistableperceptionoftheNeckercubeanditsclassicationintermsofcognitiveprocessesandper-sonalitytraits.Astrongfocuswaslaidonbetterunderstandingthestronginter-individualdierencesinbistableperception.Temporaldynamicsandseverallow-levelorbottom-upaspectsofbistabiltyweredescribedforthersttimefortheNeckercube.Itwasdemonstratedthattheinitialphaseofadaptationwhichhasbeenreportedintheliterat-urewithsomewhatvaryingcharacteristicscanbeavoidedwithappropriateinstructionsandashorttrainingphase.Furthermore,tqualityofsev-eraldwelltimedistributionswascompared,amongstothers,withamodiedKolmogorov-Smirnovtestandfoundsuperiorforthelognormaldistributioncomparedtothegammadistribution.Theeectofcubesizewasshownnottobesignicantfortherangeof1to6ofvisualangle,inwhichmanystudiesontheNeckercubearesituated.MethodologicalchallengesoftestingforahysteresiseectfortheNeckercubewereindicated.Additionally,apercep-tualbiaseectwasanalysedquantitativelyforthersttime,demonstratingapreferencetoseetheNeckercubefromabove.Thisbiaswasshowntobereectedalsoinvoluntarycontroloverperceptionofthecubeandincorrel-ationstopersonalitytraitsandself-reportedmindfulness.IntermsoftheclassicationofbistableperceptionoftheNeckercube,volun-tarycontroloverperceptionwasstudiedasameasurewithconsiderableinter-individualdierences.Voluntarycontroloverperceptionwasreproducedasreportedinseveralstudies.Itwasshownthatneutraldwelltimespredicttheabilitytoslowdownreversalstoalowextendbuttheabilitytospeedupthemuptoahighextend.Thepsychologicalconceptofaction-controlwasfoundnottobedirectlyrelatedtovoluntarycontrolofreversals.Self-leadershipandthepersonalitytraitsofconscientiousnessandneuroticism,135 BibliographyAlais,D.,vanBoxtel,J.J.,Parker,A.,vanEe,R.,2010.Attendingtoaud-itorysignalsslowsvisualalternationsinbinocularrivalry.VisionResearch50(10),929935.Allen,E.,Mattarella-Micke,A.,Shevell,S.,Beilock,S.,2011.Workingmemorycapacitypredictsindividualdierencesinperceptionofabistablegure.JournalofVision11(11).Allen,M.L.,Chambers,A.,2011.Implicitandexplicitunderstandingofambiguousguresbyadolescentswithautismspectrumdisorder.Autism15(4),457472.Andersen,R.A.,Bradley,D.C.,1998.Perceptionofthree-dimensionalstruc-turefrommotion.TrendsinCognitiveScience2(6),222228.Andreÿen,P.,Konradt,U.,2007.MessungvonSelbstführung:Psycho-metrischeÜberprüfungderdeutschsprachigenVersiondesRevisedSelf-LeadershipQuestionnaire.ZeitschriftfürPersonalpsychologie6,117128.Atmanspacher,H.,Bach,M.,Filk,T.,Kornmeier,J.,Römer,H.,2008.Cog-nitivetimescalesinanecker-zenomodelforbistableperception.OpenCyberneticsandSystemicsJournal2,234251.Atmanspacher,H.,Filk,T.,Romer,H.,2004.QuantumZenofeaturesofbistableperception.BiologicalCybernetics90(1),3340.Aydin,S.,Strang,N.C.,Manahilov,V.,2013.Age-relateddecitsinatten-tionalcontrolofperceptualrivalry.VisionResearch77,3240.Babich,S.,Standing,L.,1981.Satiationeectswithreversiblegures.Per-ceptualandMotorSkills52(1),20310.137 Bonneh,Y.S.,Ceoperman,A.,Sagi,D.,2001.Motion-inducedblindnessinnormalobservers.Nature411(6839),798801.Boring,E.G.,1930.Anewambiguousgure.TheAmericanJournalofPsychology42,444445.Borsellino,A.,Carlini,F.,Riani,M.,Tuccio,M.T.,DeMarco,A.,Penengo,P.,Trabucco,A.,1982.Eectsofvisualangleonperspectivereversalforambiguouspatterns.Perception11(3),26373.Borsellino,A.,DeMarco,A.,Allazetta,A.,Rinesi,S.,Bartolini,B.,1972.Reversaltimedistributionintheperceptionofvisualambiguousstimuli.Kybernetik10(3),13944.Brainard,D.H.,1997.ThePsychophysicsToolbox.SpatialVision10(4),433436.Brascamp,J.W.,Pearson,J.,Blake,R.,vandenBerg,A.V.,2009.Inter-mittentambiguousstimuli:Implicitmemorycausesperiodicperceptualalternations.JournalofVision9(3),123.Brascamp,J.W.,vanEe,R.,Pestman,W.R.,vandenBerg,A.V.,2005.Distributionsofalternationratesinvariousformsofbistableperception.JournalofVision5(4),28798.Braun,J.,Mattia,M.,2010.Attractorsandnoise:Twindriversofdecisionandmultistability.NeuroImage52,740751.Brickenkamp,R.,2002.Aufmerksamkeits-Belastungstestd2,9thEdition.Hogrefe,Göttingen.Britz,J.,Landis,T.,Michel,C.M.,2009.Rightparietalbrainactivitypre-cedesperceptualalternationofbistablestimuli.CerebralCortex19(1),5565.Brouwer,G.J.,vanEe,R.,2006.Endogenousinuencesonperceptualbista-bilitydependonexogenousstimulicharacteristics.VisionResearch46,33933402.Carter,O.,Konkle,T.,Wang,Q.,Hayward,V.,Moore,C.,2008.Tactilerivalrydemonstratedwithanambiguousapparent-motionquartet.CurrentBiology18(14),10501054.139 Deutsch,D.,1975.Musicalillusions.ScienticAmericanOprints233(4),110.Dugger,J.G.,Courson,R.W.,1968.EectofangleofretinalvisionontherateofuctuationoftheNeckercube.PerceptualandMotorSkills26(3),12391242.Ehm,W.,Bach,M.,Kornmeier,J.u.,2011.Ambiguousguresandbind-ing:EEGfrequencymodulationsduringmultistableperception.Psycho-physiology48(4),547558.Einhäuser,W.,Martin,K.A.C.,König,P.,2004.Areswitchesinperceptionoftheneckercuberelatedtoeyeposition?EuropeanJournalofNeuros-cience20,28112818.Einhäuser,W.,Stout,J.,Koch,C.,Carter,O.,2008.Pupildilationreectsperceptualselectionandpredictssusequentstabilityinperceptualrivalry.PNAS105(5),17041709.Ewing,J.A.,1885.Experimentalresearchesinmagnetism.PhilosophicalTransactionsoftheRoyalSocietyofLondon176,523640.Frederiksen,N.O.,Guilford,J.P.,1934.Personalitytraitsanductuationsoftheoutlinecube.TheAmericanJournalofPsychology46(3),470474.Gao,J.B.,Billock,V.A.,Merk,I.,Tung,W.W.,White,K.D.,Harris,J.G.,Roychowdhury,V.P.,2006.Inertiaandmemoryinambiguousvisualperception.CognitiveProcessesing7(2),105112.Ge,S.,Ueno,S.,Iramina,K.,2007.Thertmseectonperceptualreversalofambiguousgures.ConfProcIEEEEngMedBiolSoc2007,47436.Gershman,S.J.,Vul,E.,Tenenbaum,J.B.,2009.Perceptualmultistabilityasmarkovchainmontecarloinference.AdvancesinNeuralInformationProcessingSystems22,611619.Gigante,G.,Mattia,M.,Braun,J.,DelGuidice,P.,2009.Bistableper-ceptionmodeledascompetingstochasticintegrationattwolevels.PLOSComputationalBiology5(7).141 Jones,M.B.,1955.Authoritarianismandintoleranceofuctuation.JournalofAbnormalandSocialPsychology50(1),125126.Kabat-Zinn,J.,1982.Anoutpatientprograminbahvioralmedicineforchronicpainpatientsbasedonthepracticeofmindfulnessmeditation:Theoreticalconsiderationsandpreliminaryresults.GeneralHospitalPsy-chiatry4,3347.Kanai,R.,Carmel,D.,Bahrami,B.,Rees,G.,2011.Structuralandfunc-tionalfractionationofrightsuperiorparietalcortexinbistableperception.CurrentBiology21(3),106107.Kang,M.-S.,Blake,R.,2011.Anintegratedframeworkofspatiotemporaldynamicsofbinocularrivalry.FrontiersinHumanNeuroscience5(88),19.Keutelian,H.,1991.TheKolmogorov-Smirnovtestwhenparametersarees-timatedfromdata.Tech.rep.,Fermilab,CDFnote1285,Version1.0(April30,1991).Kidd,A.H.,Cherymisin,D.G.,1965.Figurereversalasrelatedtospecicpersonalityvariables.PerceptualandMotorSkills20(3c),11751176.Kleiner,M.,Brainard,D.,Pelli,D.,2007.What'snewinPsychtoolbox-3?Perception36(ECVPAbstractSupplement).Klintman,H.,1984.Originalthinkingandambiguousgurereversalrates.BulletinofthePsychonomicSociety22(2),129131.Kohler,A.,Haddad,L.,Singer,W.,Muckli,L.,2008.Decidingwhattosee:Theroleofintentionandattentionintheperceptionofapparentmotion.VisionResearch48(8),10961106.Kondo,H.M.,Kitagawa,N.,Kitamura,M.S.,Koizumi,A.,Nomura,M.,Kashino,M.,2012.Separabilityandcommonalityofauditoryandvisualbistableperception.CerebralCortex22(8),19151922.Kornmeier,J.,Bach,M.,2006.Bistableperception-alongtheprocessingchainfromambiguousvisualinputtoastablepercept.InternationalJournalofPsychophysiology62,345349.143 Körner,A.,Geyer,M.,Roth,M.,Drapeau,M.,Schmutzer,G.,Albani,C.,Schumann,S.,Brähler,E.,2008.PersönlichkeitsdiagnostikmitdemNEO-Fünf-Faktoren-Inventar:Die30-Item-Kurzversion(NEO-FFI-30).Psycho-therapie,PsychosomatikundmedizinischePsychologie59(6),238245.Laux,L.,Glanzmann,P.,Schaner,P.,Spielberger,C.D.,1981.DasState-Trait-Angstinventar.TheoretischeGrundlagenundHandanweisung.BeltzTestgesellschaft,Weinheim.Leopold,D.A.,Logothetis,N.K.,1999.Multistablephenomena:changingviewsinperception.TrendsinCognitiveScience3(7),254264.Leopold,D.A.,Wilke,M.,Maier,A.,Logothetis,N.K.,2002.Stablepercep-tionofvisuallyambiguouspatterns.NatureNeuroscience5(6),605609.Liebert,R.M.,Burk,B.,1985.Voluntartycontrolofreversiblegures.Per-ceptualandMotorSkills61(3f),13071310.Lilliefors,H.W.,1967.OntheKolmogorov-Smirnovtestfornormalitywithmeanandvarianceunknown.JournaloftheAmericanStatisticalAssoci-ation62(318),399402.Long,G.M.,Toppino,T.C.,2004.Enduringinterestinperceptualambigu-ity:alternatingviewsofreversiblegures.PsychologicalBulletin130(5),74868.Long,G.M.,Toppino,T.C.,Mondin,G.W.,1992.Primetime:Fatigueandseteectsintheperceptionofreversiblegures.PerceptionandPsycho-physics52(6),609616.Mach,E.,1885/1902.DieAnalysederEmpndungenunddasVerhältnissdesPhysischenzumPsychischen.VerlagvonGustavFischer,Jena.Maier,A.,Panagiotaropoulos,T.I.,Tsuchiya,N.,Keliris,G.A.,2012.In-troductiontoresearchtopic-binocularrivalry:agatewaytostudyingconsciouness.FrontiersinHumanNeuroscience6(263),13.Massey,F.J.,J.,1951.TheKolmogorov-Smirnovtestforgoodnessoft.JournaloftheAmericanStatisticalAssociation46(253),6878.Meenan,J.P.,Miller,L.A.,1994.Perceptualexibilityafterfrontalortem-porallobectomy.Neuropsychologia32(9),11451149.145 Paunonen,S.V.,Ashton,M.C.,2001.Bigvefactorsandfacetsandthepredictionofbehavior.JournalofPersonalityandSocialPsychology81(3),524539.Pearson,J.,Brascamp,J.,2008.Sensorymemoryforambiguousvision.TrendsinCognitiveScience12(9),334341.Pelli,D.G.,1997.TheVideoToolboxsoftwareforvisualpsychophysics:transformingnumbersintomovies.SpatialVision10(4),437442.Pelton,L.H.,Solley,C.M.,1968.AccelerationofreversalsofaNeckercube.AmericanJournalofPsychology81(4),585588.Praissman,S.,2008.Mindfulness-basedstressreduction:Aliteraturereviewandclinician'sguide.JournaloftheAmericanAcademyofNursePracti-tioners20(4),212216.Pressnitzer,D.,Hupé,J.-M.,2006.Temporaldynamicsofauditoryandvisualbistabilityrevealcommonprinciplesofperceptualorganization.CurrentBiology16(13),13511357.Price,J.R.,1967.Perspectivedurationofaplanereversiblegure.Psycho-nomicScience9(12),623624.Price,J.R.,1969.Eectofextendedobservationonreversibleperspectiveduration.PsychonomicScience16(2),7576.Pöppel,E.,1997.Ahierarchicalmodeloftemporalperception.TrendsinCognitiveScience1(2),5661.Radilova,J.,Pöppel,E.,Ilmberger,J.,1990.Auditoryreversaltiming.Activ-itasnervosasuperior32(2),1378.Reisberg,D.,O'Shaughnessy,M.,1984.Divertingsubjects'concentrationslowsguralreversals.Perception13(4),461468.Riani,M.,Oliva,G.A.,Selis,G.,Ciurlo,G.,Rossi,P.,1984.Eectofluminanceonperceptualalternationofambiguouspatterns.PerceptualandMotorSkills58(1),267274.Ricci,C.,Blundo,C.,1990.Perceptionofambiguousguresafterfocalbrainlesions.Neuropsychologia28(11),11631173.147 Sterzer,P.,Kleinschmidt,A.,2007.Aneuralbasisforinferenceinperceptualambiguity.PNAS104(1),323328.Sterzer,P.,Kleinschmidt,A.,Rees,G.,2009.Theneuralbasesofmultistableperception.TrendsinCognitiveSciences13(7),310318.Strüber,D.,Ba³ar-Eroglu,C.,Miener,M.,Stadler,M.,2001.EEGgamma-bandresponseduringtheperceptionofNeckercubereversals.VisualCog-nition8(4/5/6),609621.Strüber,D.,Stadler,M.,1999.Dierencesintop-downinuencesonthereversalrateofdierentcategoriesofreversiblegures.Perception28(10),11851196.Sundareswara,R.,Schrater,P.R.,2008.PerceptualmultistabilitypredictedbysearchmodelforBayesiandecisions.JournalofVision8(5),119.Szelag,E.,vonSteinbüchel,N.,Reiser,M.,deLangen,E.G.,Pöppel,E.,1996.Temporalconstraintsinprocessingofnonverbalrhythmicpatterns.ActaNeurobiolgiaeExperimentalis56(1),215225.Toppino,T.C.,2003.Reversible-gureperception:Mechanismsofinten-tionalcontrol.PerceptionandPsychophysics65(8),12851295.Toppino,T.C.,Long,G.M.,1987.Selectiveadaptationwithreversibleg-ures:don'tchangethatchannel.PerceptionandPsychophysics42(1),3748.Troje,N.F.,McAdam,M.,2010.Theviewing-from-abovebiasandthesil-houtteillusion.i-Perception1(3),143148.Ulbrich,P.,Churan,J.,Fink,M.,Wittmann,M.,2009.Perceptionoftem-poralorder:Theeectsofage,sex,andcognitivefactors.Aging,Neuro-psychology,andCognition16(2),183202.vanDam,L.C.J.,vanEe,R.,2005.Theroleof(micro)saccadesandblinksinperceptualbi-stabilityfromslantrivalry.VisionResearch45(18),24172435.vanDam,L.C.J.,vanEe,R.,2006.Theroleofsaccadesinexertingvoluntarycontrolinperceptualandbinocularrivalry.VisionResearch46,787799.149 Woodru,B.W.,Viviano,P.J.,Moore,A.H.,Dunne,E.J.,1984.Modiedgoodness-of-ttestsforgammadistributionswithunknownlocationandscaleparameters.IEEETransactionsonReliability33(3),241245.Zaretskaya,N.,Anstis,S.,Bartels,A.,2013.Parietalcortexmediatescon-sciousperceptionofillusorygestalt.JournalofNeuroscience33(2),523531.Zhou,W.,Chen,D.,2009.Binaralrivalrybetweenthenostrilsandinthecortex.CurrentBiology19(18),15611565.Zhou,Y.H.,Gao,J.B.,White,K.D.,Merk,I.,Yao,K.,2004.Perceptualdominancetimedistributionsinmultistablevisualperception.BiologicalCybernetics90,256263.151 Summer2007ResearchinternshipatthenanophysicsgroupofProf.KlausEnsslinofETHZonatomicforcemicroscopyMarch-July2008DiplomathesisatProf.MarkSherwin'sgroupatUCSantaBarbara,California,usingterahertzFTIRspectroscopytostudyquantumposts,anewsemiconductornanostructureDecember2008-May2013PhDatCollegiumHelveticum,LaboratoryforTransdisciplinaryResearch,onbistablepercep-tionoftheNeckercubeResearchFieldsandInterestsVisualandbistableperception(psychology),timeperception,cognitivebasesofmathematics,philosophyofscienceLanguagesGerman(native),English,FrenchPublicationsWernery,J.,2011.FüreinenEmpirismusderWürde.In:Sigg,H.,Folkers,G.(Eds.),GüterabwägungbeiderBewilligungvonTierversuchen.Collegi-umsheft11,CollegiumHelveticum,Zürich,pp.119-121Wernery,J.,Kornmeier,J.,Candia,V.,Folkers,G.,Atmanspacher,H.,2011.DwelltimedistributionsforthebistableperceptionoftheNeckercube.Per-ception40(ECVPAbstractSupplement),172.153