Presentation on theme: "And a big thanks to Julie Markant"— Presentation transcript:
1And a big thanks to Julie Markant Categorical perception of speech: Task variations in infants and adultsBob McMurrayJessica MayeAndrea LathropandRichard N. AslinAnd a big thanks to Julie Markant
2Categorical Perception & Task Variations OverviewPrevious workCategorical perception and gradient sensitivity to subphonemic detail.Categorical perception in infantsReassessing this with HTPP & AEMInfants show gradient sensitivityA new methodologyAdult analogues
3Categorical Perception Is subphonemic detail retained [and used] during speech perception?For a long time…NO!Subphonemic variation is discarded infavor of a discrete label.Categorical perception & gradiency
4Non-categorical Perception A number of psychophysical-type results showed listeners’ sensitivity to within-category detail.4AIX TaskPisoni & Lazarus (1974)Speeded ResponseCarney, Widen & Viemeister (1977)TrainingSamuel (1977)Pisoni, Aslin, Henessey & Perey (1982)Rating TaskMiller (1997)Massaro & Cohen (1983)
5Word Recognition These results did not reveal: Whether on-line word recognition is sensitive to such detail.Whether such sensitivity is useful during recognition.
6Word Recognition Mounting evidence that word-recognition is sensitive: Lahiri & Marslen-Wilson (1991): vowel nasalizationAndruski, Blumstein & Burton (1994): VOTGow & Gordon (1995): word segmentationSalverda, Dahan & McQueen (in press): embedded words and vowel length.Dahan, Magnuson, Tanenhaus & Hogan (2001): coarticulatory cues in vowel.
7Gradient Sensitivity McMurray, Tanenhaus & Aslin (2002) Eye-movements to objects after hearing items from 9-step VOT continuum.McMurray, Tanenhaus & Aslin (2002)Systematic relationship between VOT and looks to the competitor.5101520253035400.020.030.040.050.060.070.08VOT (ms)CategoryBoundaryResponse=Looks toCompetitor FixationsBear
8Gradient Sensitivity ….BUT This systematic, gradient relationship between lexical activation and acoustic detail would allow the system take advantage of fine-grained regularities in the signal.Gow, McMurray & Tanenhaus (Sat., 6:00 poster session)Anticipate upcoming material.Resolve AmbiguityIf fine-grained detail is useful we might expect infants and children toShow gradient sensitivity to variationTune their sensitivity to learning environment….BUT
9Infant categorical perception Early findings of categorical perception for infants (e.g. Eimas, Siqueland, Jusczyk & Vigorito) have never been refuted.Most studies use:Habituation (many repetitions)Synthetic SpeechSingle continuumPerhaps a different method would be more sensitive?Categorical perception in infants
10Head-Turn Preference Procedure Jusczyk & Aslin (1995)Infants exposed to a chunk of language:Words in running speech.Stream of continuous speech (ala stat. learning)Word listAfter exposure, memory for exposed items (or abstractions) is assessed by comparing listening time to consistent items with inconsistent items.
11How do we measure listening time? After exposure…Center Light blinks.Brings infant’s attention to center.
12How do we measure listening time? When infant looks at center…One of the side-lights blinks.
13How do we measure listening time? Beach… Beach… Beach…When infant looks at side-light…she hears a word.
14How do we measure listening time? When infant looks at side-light…she hears a word.…as long as she keeps looking…
15Experiment 1: Gradiency in Infants 7.5 month old infants exposed to either 4 b-, or 4 p-words Bomb Bear Bail BeachPalm Pear Pail Peach80 repetitions totalForm a category of the exposed class of words.Infants show gradient sensitivityMeasure listening time onBear Pear (Original word)Pear Bear (opposite)Bear* Pear* (VOT closer to boundary).
16Experiment 1: StimuliStimuli constructed by cross-splicing natural, recorded tokens of each end point.B: M= 3.6 ms VOTP: M= 40.7 ms VOTB*: M=11.9 ms VOTP*: M=30.2 ms VOTBoth were judged /b/ or /p/ at least 90% consistently by adult listeners.B: 98.5% B*: 97%P: 99% P*: 96%
17Measuring gradient sensitivity Looking time is an indication of interest.After hearing all of those B-wordsP sounds pretty interesting.So: infants should look longer for pear than bear.What about in between?Listening TimeBearPearGradientBear*Categorical
18Individual Differences Novelty/Familiarity preference varies across infants and experiments.We’re only interested in the middle stimuli (b*, p*).Infants categorized as novelty or familiarity preferring by performance on the endpoints.NoveltyFamiliarityB2711P1910Within each group will we see evidence for gradiency?
19Novelty Results Novelty infants, Trained on B VOT: p=.001** Linear Trend: p=.001**40005000600070008000900010000BB*PListening Time (ms).14.004**
20Novelty Results Novelty infants, Trained on P VOT: p=.001** Linear Trend: p=.001**40005000600070008000900010000PP*BListening Time (ms).1.001**
21Familiarity Results Familiarity infants showed similar effects. B exposureTrend: p=.001B vs B*: p=.19B* vs P: p=.21P exposureTrend: p=.009P vs P*: p=.057P* vs. B: p=.096Trained on BTrained on P
22Experiment 1: Conclusions 7.5 month old infants show gradient sensitivity to subphonemic detail.Individual differences in familiarity/novelty preferences. Why?Length of exposure?Individual factors?Limitations of paradigm may hinder further study:More repeated measuresBetter understanding of “task”Wider age-range.
23Anticipatory Eye-Movements An ideal methodology wouldYield an arbitrary, identification response.Yield a response to a single stimuliYield many repeated measuresMuch like a forced-choice identificationA new methodologyAnticipatory Eye-Movements (AEM):Train Infants to look left or right in response to a single auditory stimulus
24Anticipatory Eye-Movements Visual stimulus moves under occluder.Reemergence serves as “reinforcer”Concurrent auditory stimulus predicts endpoint of occluded trajectory.Subjects make anticipatory eye-movements to the expected location—before the stimulus appears.TeakLamb
25Anticipatory Eye-Movements After training on original stimuli, infants are tested on a mixture oforiginal, trained stimuli (reinforced)Maintain interest in experiment.Provide objective criterion for inclusionnew, generalization stimuli (unreinforced)Examine category structure/similarity relative to trained stimuli.
26Experiment 2: Pitch and Duration Goals:Use AEM to assess auditory categorization.Assess infants’ abilities to “normalize” for variations in pitch and duration…or…Are infants’ sensitive to acoustic-detail during a lexical identification task...
27Experiment 2: Pitch and Duration Training:“Teak” -> rightward trajectory.“Lamb” -> leftward trajectory.“teak!”“lamb!”Test:Lamb & Teak with changes in:Duration: 33% and 66% longer.Pitch: 20% and 40% higherIf infants ignore irrelevant variation in pitch or duration, performance should be good for generalization stimuli.If infants’ lexical representations are sensitive to this variation, performance will degrade.
30Results Each trials is scored as correct: longer looking time to the correct side.incorrect: longer looking time to incorrect side.Binary DV—similar to 2AFC.On trained stimuli:11 of 29 infants performed better than chance–this is a tough tasks for infants. Perhaps more training.
32Experiment 2: Conclusions Infants’ developing lexical categories show graded sensitive to variation in duration.Possibly not to pitchMight be an effect of “task relevance”AEM yieldsmore repeated measurementsbetter understood task: 2AFCCould it yield a picture of the entire developmental time course? Is AEM applicable to a wider age range?
33Treating undergraduates like babies Extreme case: Adult perception.Adults generally won’tLook at blinking lights…Suck on pacifiers…Kick their feet at mobiles…Result: few infant methodologies allow direct analogues to adults.They do make eye-movements……could AEM be adapted?
34Treating undergraduates like babies Pilot study.5 adults exposed to AEM stimuli.Training:“Ba” left“Pa” rightTestBa – Pa (0-40 ms) VOT continuum.
35Results Second group of subjects run in an explicit 2AFC. Same category boundary.Steeper slope: less sensitivity to VOT.10.90.80.70.6% /p/2AFC0.50.4AEM0.30.20.1510152025303540VOT (ms)
36Adult AEM: Conclusions AEM paradigm can be used unchanged for adults.Should work with older children as well.Results show same category boundary as traditional 2AFC tasks, perhaps more sensitivity to fine-grained acoustic detail.Potentially useful for speech categorization when categories are not:nameablepictureableimmediately obvious
37ConclusionsLike adults,7.5-month-old infants show gradient sensitivity to subphonemic detail.VOTDurationPerhaps not pitch (w.r.t. lexical categories)
38Conclusions Task makes the difference: Moving to HTPP from habituation revealed subphonemic sensitivity.Taking into account individual differences crucial.Moving to AEM yieldsBetter ability to examine tuning over time.Ability to assess perception across lifespan with a single paradigm.
39And a big thanks to Julie Markant Categorical perception of speech: Task variations in infants and adultsBob McMurrayJessica MayeAndrea LathropandRichard N. AslinAnd a big thanks to Julie Markant
40Natural Stimuli Infants may show more sensitivity to natural speech Stimuli constructed from natural tokens ofactual words with progressive cross-splicing.PalmBomb
41Experiment 1: RepriseDifficult to examine how sensitivity might be tuned to environmental factors in head-turn-preference procedure.High variance/individual differences—can’t predict novelty/familiarity.Only a single point to look at.Between-subject comparison.Difficult interaction to obtainListening Time6 m/o10 m/o8 m/oBearBear*Pear
42Experiment 1: Reprise AEM presents a potential solution: 1) Looking at whole continuum would yield more power.6 m/o8 m/o10 m/oBearPear2) Is AEM applicable to a wider age range?
44Data analysisData coded by naive coders from video containing pupil & scene monitors.
45Data analysisLeftRight-outStartInCenterOffLeft-outRight-outLeft-InRight-InLeft-out, Right-out, center & start treated as “neither”.Left-in, Left treated as anticipation to left.Right-in, Right treated as anticipation to right.Eye-movements coded frommaximal size of stimulus to first appearance (or end of trial).