Name: Defining the Role Of Language in Infants Object Categorization with Eye-tracking Paradigms (Video) | JoVE Journal
Uploaded: 2019-02-08T15:00:00.000+00:00
Duration: 7 min 31 s
Description: Learn familiarization-test paradigms for infant categorization using eye-tracking and visual stimuli. This method reveals how infants form object categories and identifies conditions that enhance early learning.

ai_agent_IP

ai_agent

maintenance

instantTrialDurationMinutes

isBiopharma

enableNewMyJoveSectionDesign

educationFeatures

seoMetaApi

ai_copilot_institution_blacklist

ai_chat_institution_blacklist

freeAccessWindow

ssoRoleCapture

scheduledMaintenanceBanner

joveLabsFeatures

disable_autoplay_for_bots

extended_trial_duration

5 minutes

The research reported here was supported by the National Institute of Child Health and Human Development of the National Institutes of Health under award number R01HD083310 and a National Science Foundation Graduate Research Fellowship under grant no. DGE&#8208;1324585. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the National Science Foundation.

All methods described here have been approved by the Northwestern University Institutional Review Board.
1. Stimuli Creation
NOTE: The visual stimuli (see Figure 1) used in the representative design reported below were originally developed in Havy and Waxman (2016)18 and are available for download at https://osf.io/n6uy8/.
<ol>
	<li>To create a new continuous category, first design a pair of novel digital images. Next, morph the pair of images together, using software (see, e.g., Table of Materials) to form a continuum of exemplars between the two original images. Create at least two categories in this way so that one can serve as the category to be learned while the other provides the novel category exemplar for the test trial.</li>
	<li>Select the familiarization exemplars at evenly spaced intervals from across each learned category&#8217;s continuum (e.g., the 0%, 20%, 40%, 60%, 80%, and 100% exemplars). Select an appropriate number of exemplars (e.g., six). Commensurate with the difficulty of the category and age of the participants.</li>
	<li>To create the exemplars for the test phase, select the midpoints of the familiar category&#8217;s continuum and the novel category&#8217;s continuum (i.e., the 50% exemplar). Then match the color of the novel exemplar to that of the familiar exemplar using an image manipulation program (see, e.g., Table of Materials).</li>
	<li>Record auditory stimuli produced by a female native English speaker in a soundproof booth. If possible, use the same speaker for both labeling phrases (i.e., &#8220;Look at the modi&#8221;) and non-labeling phrases (i.e., &#8220;Look at that!&#8221;).
	<ol>
		<li>Instruct the speaker to produce all utterances in infant- or child-directed speech.</li>
		<li>Select utterances which are approximately the same length across conditions, likely around 1,500 ms per phrase.</li>
	</ol>
	</li>
</ol>
2. Apparatus
<ol>
	<li>Use an appropriate eye-tracker. To collect adequate eye-tracking data for a familiarization-test measure, most widely available eye-trackers will suffice: the objects occupy large portions of the screen, and the data analysis investigates performance over a long window, rather than individual, rapidly occurring eye movements such as saccades.</li>
	<li>Because this task requires eye-tracking infants, ensure that the system conforms to several requirements.
	<ol>
		<li>First, use an &#173;eye-tracker with a remote tracking mode, which does not require infants to place their heads on a chin-rest. Ensure that the eye-tracker can tolerate relatively large head movements or readjustments.</li>
		<li>Second, use a relatively large screen to display the images to infants, (e.g., 57 x 45 cm).</li>
		<li>Third, use an extendable arm mount for the eye-tracker to facilitate data collection by allowing the researcher to adjust the height of the eye-tracker to each infant.</li>
		<li>Fourth,&#160;make the eye-tracking equipment unobtrusive, focusing infants&#8217; attention solely on the display screen. For instance, some systems integrate the eye-tracking equipment with the display monitor or mount the equipment directly below the monitor.</li>
	</ol>
	</li>
	<li>Note that this task can also be completed by hand-coding high-quality video data of the infants&#8217; looking behavior. While hand-coding techniques may pose some challenges for using the more fine-grained time-course analyses, hand-coded data are entirely sufficient&#160;for the aggregate looking analyses.</li>
</ol>
3. Task Design
<ol>
	<li>In the eye-tracker&#8217;s associated software (see, e.g., Table of Materials), create four different conditions: Fully Supervised, Unsupervised, Semi-supervised, and Reversed Semi-supervised. Ensure these conditions are separate, so that each infant will see only one condition.</li>
	<li>Generate at least two pseudo-random orders of the learning exemplars, with the constraint that no more than two exemplars from the same side of the continuum (0-40% or 60-100%) can be shown consecutively.</li>
	<li>Create familiarization videos that pair the auditory stimuli with the visual stimuli as appropriate for each condition.
	<ol>
		<li>Combine the visual and auditory stimuli in video editing software (see, e.g., Table of Materials). Present all images on the same background. Set the onset of the auditory stimulus to an appropriate range, between 500 ms and 1,500 ms after the onset of the visual stimulus. Use this short delay to ease infants&#8217; processing load 19.</li>
		<li>For instance, in the Fully Supervised condition, pair each familiarization exemplar with a labeling phrase.</li>
		<li>In the Unsupervised condition, pair each familiarization exemplar with a non-labeling phrase.</li>
		<li>In the Semi-supervised condition, pair only the first two exemplars in each order with labeling phrases but the rest with non-labeling phrases.</li>
		<li>For the Reversed Semi-supervised condition, pair the final two exemplars with labeling phrases but the first four with non-labeling phrases (see Figure 1).</li>
		<li>Upload these videos into the eye-tracker software, ordering the familiarization videos as determined by the pseudo-randomized order.</li>
	</ol>
	</li>
	<li>Upload a short (10 s or less) attention-grabbing animation displayed in the center of the screen after familiarization: this will ensure that most infants are looking to the center of the screen when the test phase begins.</li>
	<li>Finally, for each learning category, design two test trials, each featuring two exemplars displayed side-by-side. Ensure that for both test trials, one exemplar will represent the midpoint of the now-familiar category while the other represents the midpoint of the novel category.
	<ol>
		<li>Counterbalance the trials so that the left/right positioning of the novel exemplar in the test trial is reversed across videos.</li>
		<li>Upload these test trials to the eye-tracker software, positioning them after the post-familiarization attention-getter. Counterbalance these trials&#8217; presentation so each infant has an equal chance of seeing a left-novel or right-novel test trial.</li>
		<li>Ensure that test trials last at least 5 s, and up to 20 s, in order for children initially looking away to accumulate sufficient looking.</li>
	</ol>
	</li>
</ol>
[Place Figure 1 here]
4. Study Procedure
<ol>
	<li>Before the infant arrives, set up the eye-tracker.
	<ol>
		<li>Randomly assign the infant to a condition and an order.</li>
		<li>Open the eye-tracker software and select the assigned condition/order pair.</li>
		<li>Now enter the participant number for this recording.</li>
	</ol>
	</li>
	<li>After performing the consent process, bring the infant and the caregiver to the eye-tracking room. Ensure the room is moderately lit without any distracting decorations on the walls.</li>
	<li>Place a chair in front of the eye-tracker at an appropriate distance for the model of eye-tracker being used. Seat the caregiver in this chair and the infant on the caregiver&#8217;s lap. If the infant does not wish to sit in the caregiver&#8217;s lap, they may sit on their own, or they may sit in a car seat.</li>
	<li>If the infant is sitting on the caregiver&#8217;s lap, instruct the caregiver not to bias infants&#8217; behavior in any way but to try to keep the infant centered on the caregiver&#8217;s lap. Provide caregivers with a pair of blacked-out sunglasses to wear so they cannot see the stimuli.</li>
	<li>Ask the infant to look at the eye-tracker screen; consider displaying an engaging image or video to attract their attention. Position the screen so that infants&#8217; eyes are within the calibration window.</li>
	<li>Perform the eye-tracker&#8217;s calibration procedure. Use a five-point calibration if possible, but less comprehensive calibrations are also likely to be adequate. Infants often respond better when the calibration image is an animation with auditory accompaniment.</li>
	<li>If the infant passes calibration, then begin the experiment. If not, recalibrate until they are successful. Any infants who cannot be calibrated are excluded.</li>
	<li>If multiple experiments are run consecutively, or if a single experiment is quite long, consider re-calibrating after each section.</li>
</ol>
5. Data Analysis
<ol>
	<li>Use data analysis software to perform this analysis (e.g., see Table of Materials).</li>
	<li>Create areas of interest (AOIs) around the exemplar positions on the left and right sides of the screen.</li>
	<li>For familiarization trials, use the appropriate AOI to assess the time infants spent looking to the exemplar displayed on each trial. Exclude any infant who does not show sustained looking for a majority of the exemplars (e.g., require that infants attend to 4 of a possible 6 familiarization exemplars for at least 25% of those trials).</li>
	<li>For the test trial, include only infants&#8217; first 5 s of accumulated looking. For younger infants, from 3 to 12 months of age, consider using a longer window such as 10 seconds of accumulated looking. Consider excluding infants who show insufficient sustained looking at test (e.g., accumulating less than 2.5 s of looking) or who fail to look to both of the exemplars.</li>
	<li>Now create a preference score for each infant&#8217;s test trial by dividing the amount of time spent looking to the novel exemplar by the total amount of time looking to both exemplars. To analyze these proportions, transform them first with an empirical logit or arc-sin square-root to make them suitable for analysis with linear models.</li>
	<li>For a time-course analysis of infants&#8217; looking behavior at the test, separate data into small bins (e.g., between 10 and 100 ms), and calculate a preference score within each bin for each infant.</li>
	<li>Perform an analysis of the time-course data, testing whether infants&#8217; pattern of looking throughout the test trial varies by condition. Note that multiple forms of analysis may answer this question, including a cluster-based permutation analysis20, as demonstrated here, and growth curve modeling.21
	<ol>
		<li>For a cluster-based permutations analysis, select a t-value threshold, corresponding to the desired alpha level (recommended alphas range from .01 to .20; note that this alpha value does not represent the overall test&#8217;s alpha level, merely the level required for individual time-bins to exceed the threshold). Sum the t-statistics for every consecutive time-bin that surpasses the chosen t-threshold; these cumulative t-statistics indicate the size of the divergences between conditions in the data.</li>
		<li>To determine if these divergences are greater than expected by chance, perform at least 1,000 simulations with the condition labels randomly shuffled. Evaluate the unshuffled data&#8217;s divergences against this chance-based distribution. 
		NOTE: It is this comparison of the original divergence against the chance-based distribution that determines the false-positive rate of the analysis, rather than the number of time-bins in which t-tests were conducted or even the t-value threshold selected for those initial t-tests. As a result, this analysis provides a conservative alternative to directly reporting the results from multiple t-tests across pre-specified time-bins (e.g., conducting tests every 500 ms).</li>
	</ol>
	</li>
</ol>

<ol><li>Eimas, P. D., Quinn, P. C. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Studies+on+the+Formation+of+Perceptually+Based+Basic-Level+Categories+in+Young+Infants%5BTitle%5D)+AND+%22Child+Development%22%5BJournal%5D)">Studies on the Formation of Perceptually Based Basic-Level Categories in Young Infants.</a> Child Development. 65 (3), 903-917 (1994).</li>
<li>Madole, K. L., Oakes, L. M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Making+sense+of+infant+categorization%3A+Stable+processes+and+changing+representations%5BTitle%5D)+AND+%22Developmental+Review%22%5BJournal%5D)">Making sense of infant categorization: Stable processes and changing representations.</a> Developmental Review. 19 (2), 263-296 (1999).</li>
<li>Gelman, S. A., Markman, E. M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Categories+and+induction+in+young+children%5BTitle%5D)+AND+%22Cognition%22%5BJournal%5D)">Categories and induction in young children.</a> Cognition. 23 (3), 183-209 (1986).</li>
<li>Ferry, A. L., Hespos, S. J., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Categorization+in+3-+and+4-month-old+infants%3A+An+advantage+of+words+over+tones%5BTitle%5D)+AND+%22Child+development%22%5BJournal%5D)">Categorization in 3- and 4-month-old infants: An advantage of words over tones.</a> Child development. 81 (2), 472-479 (2010).</li>
<li>Fulkerson, A. L., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Words+%28but+not+Tones%29+Facilitate+Object+Categorization%3A+Evidence+From+6-+and+12-Month-Olds%5BTitle%5D)+AND+%22Cognition%22%5BJournal%5D)">Words (but not Tones) Facilitate Object Categorization: Evidence From 6- and 12-Month-Olds.</a> Cognition. 105 (1), 218-228 (2007).</li>
<li>Balaban, M. T., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Do+words+facilitate+object+categorization+in+9-month-old+infants%3F%5BTitle%5D)+AND+%22Journal+of+Experimental+Child+Psychology%22%5BJournal%5D)">Do words facilitate object categorization in 9-month-old infants?</a> Journal of Experimental Child Psychology. 64 (1), 3-26 (1997).</li>
<li>Waxman, S. R., Braun, I. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Consistent+%28but+not+variable%29+names+as+invitations+to+form+object+categories%3A+New+evidence+from+12-month-old+infants%5BTitle%5D)+AND+%22Cognition%22%5BJournal%5D)">Consistent (but not variable) names as invitations to form object categories: New evidence from 12-month-old infants.</a> Cognition. 95 (3), B59-B68 (2005).</li>
<li>Balaban, M. T., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((An+examination+of+the+factors+underlying+the+facilitative+effect+of+word+phrases+on+object+categorization+in+9-month-old+infants%5BTitle%5D)+AND+%22Proceedings+of+the+20th+Boston+University+Conference+on+Language+Development%22%5BJournal%5D)">An examination of the factors underlying the facilitative effect of word phrases on object categorization in 9-month-old infants.</a> Proceedings of the 20th Boston University Conference on Language Development. 1, 483-493 (1996).</li>
<li>Waxman, S. R., Markow, D. B. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Words+as+invitations+to+form+categories%3A+evidence+from+12-+to+13-month-old+infants%5BTitle%5D)+AND+%22Cognitive+Psychology%22%5BJournal%5D)">Words as invitations to form categories: evidence from 12- to 13-month-old infants.</a> Cognitive Psychology. 29 (3), 257-302 (1995).</li>
<li>Zhu, X. <a target="_blank" href="http://scholar.google.com/scholar?hl=en&safe=off&q=author%3aZhu%2C+X+%22Semi-supervised+learning+literature+survey%22">Semi-supervised learning literature survey</a>. , (2005).</li>
<li>Chapelle, O., Scholkopf, B., Zien, A. <a target="_blank" href="http://scholar.google.com/scholar?hl=en&safe=off&q=author%3aChapelle%2C+O+%22Semi-supervised+learning%3A+Adaptive+computation+and+machine+learning%22">Semi-supervised learning: Adaptive computation and machine learning</a>. , MIT Press. Cambridge, Mass., USA. (2006).</li>
<li>Zhu, X., Goldberg, A. B. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Introduction+to+semi-supervised+learning%5BTitle%5D)+AND+%22Synthesis+lectures+on+artificial+intelligence+and+machine+learning%22%5BJournal%5D)">Introduction to semi-supervised learning.</a> Synthesis lectures on artificial intelligence and machine learning. 3 (1), 1-130 (2009).</li>
<li>Ferry, A. L., Hespos, S. J., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Nonhuman+primate+vocalizations+support+categorization+in+very+young+human+infants%5BTitle%5D)+AND+%22Proceedings+of+the+National+Academy+of+Sciences+of+the+United+States+of+America%22%5BJournal%5D)">Nonhuman primate vocalizations support categorization in very young human infants.</a> Proceedings of the National Academy of Sciences of the United States of America. 110 (38), 15231-15235 (2013).</li>
<li>Hunter, M. A., Ames, E. W. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((A+multifactor+model+of+infant+preferences+for+novel+and+familiar+stimuli%5BTitle%5D)+AND+%22Advances+in+infancy+research%22%5BJournal%5D)">A multifactor model of infant preferences for novel and familiar stimuli.</a> Advances in infancy research. , (1988).</li>
<li>Rose, S. A., Feldman, J. F., Jankowski, J. J. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Infant+visual+recognition+memory%5BTitle%5D)+AND+%22Developmental+Review%22%5BJournal%5D)">Infant visual recognition memory.</a> Developmental Review. 24 (1), 74-100 (2004).</li>
<li>Wetherford, M. J., Cohen, L. B. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Developmental+changes+in+infant+visual+preferences+for+novelty+and+familiarity%5BTitle%5D)+AND+%22Child+Development%22%5BJournal%5D)">Developmental changes in infant visual preferences for novelty and familiarity.</a> Child Development. , 416-424 (1973).</li>
<li>Perone, S., Spencer, J. P. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Autonomous+visual+exploration+creates+developmental+change+in+familiarity+and+novelty+seeking+behaviors%5BTitle%5D)+AND+%22Frontiers+in+psychology%22%5BJournal%5D)">Autonomous visual exploration creates developmental change in familiarity and novelty seeking behaviors.</a> Frontiers in psychology. 4, 648(2013).</li>
<li>Havy, M., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Naming+influences+9-month-olds%E2%80%99+identification+of+discrete+categories+along+a+perceptual+continuum%5BTitle%5D)+AND+%22Cognition%22%5BJournal%5D)">Naming influences 9-month-olds’ identification of discrete categories along a perceptual continuum.</a> Cognition. 156, 41-51 (2016).</li>
<li>Althaus, N., Plunkett, K. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Timing+matters%3A+The+impact+of+label+synchrony+on+infant+categorisation%5BTitle%5D)+AND+%22Cognition%22%5BJournal%5D)">Timing matters: The impact of label synchrony on infant categorisation.</a> Cognition. 139, 1-9 (2015).</li>
<li>Maris, E., Oostenveld, R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Nonparametric+statistical+testing+of+EEG-+and+MEG-data%5BTitle%5D)+AND+%22Journal+of+Neuroscience+Methods%22%5BJournal%5D)">Nonparametric statistical testing of EEG- and MEG-data.</a> Journal of Neuroscience Methods. 164 (1), 177-190 (2007).</li>
<li>Raudenbush, S. W., Bryk, A. S. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Hierarchical+Linear+Models%3A+Applications+and+Data+Analysis+Methods%5BTitle%5D)+AND+%22SAGE%22%5BJournal%5D)">Hierarchical Linear Models: Applications and Data Analysis Methods.</a> SAGE. , (2002).</li>
<li>LaTourrette, A., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((A+little+labeling+goes+a+long+way%3A+Semi-supervised+learning+in+infancy%5BTitle%5D)+AND+%22Developmental+Science%22%5BJournal%5D)">A little labeling goes a long way: Semi-supervised learning in infancy.</a> Developmental Science. , e12736(2018).</li>
<li>Dink, J., Ferguson, B. <a target="_blank" href="http://scholar.google.com/scholar?hl=en&safe=off&q=author%3aDink%2C+J+%22eyetrackingR%3A+An+R+library+for+eyetracking+data+analysis%22">eyetrackingR: An R library for eyetracking data analysis</a>. , (2015).</li>
<li>Kalish, C. W., Zhu, X., Rogers, T. T. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Drift+in+children%27s+categories%3A+When+experienced+distributions+conflict+with+prior+learning%5BTitle%5D)+AND+%22Developmental+Science%22%5BJournal%5D)">Drift in children's categories: When experienced distributions conflict with prior learning.</a> Developmental Science. 18 (6), 940-956 (2015).</li>
<li>Gibson, B. R., Rogers, T. T., Zhu, X. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Human+semi-supervised+learning%5BTitle%5D)+AND+%22Topics+in+Cognitive+Science%22%5BJournal%5D)">Human semi-supervised learning.</a> Topics in Cognitive Science. 5 (1), 132-172 (2013).</li>
<li>Keates, J., Graham, S. A. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Category+Markers+or+Attributes+Why+Do+Labels+Guide+Infants%27+Inductive+Inferences%3F%5BTitle%5D)+AND+%22Psychological+Science%22%5BJournal%5D)">Category Markers or Attributes Why Do Labels Guide Infants' Inductive Inferences?</a> Psychological Science. 19 (12), 1287-1293 (2008).</li>
<li>Booth, A. E., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((A+horse+of+a+different+color%3A+Specifying+with+precision+infants%E2%80%99+mappings+of+novel+nouns+and+adjectives%5BTitle%5D)+AND+%22Child+development%22%5BJournal%5D)">A horse of a different color: Specifying with precision infants’ mappings of novel nouns and adjectives.</a> Child development. 80 (1), 15-22 (2009).</li>
<li>Perszyk, D. R., Waxman, S. R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Listening+to+the+calls+of+the+wild%3A+The+role+of+experience+in+linking+language+and+cognition+in+young+infants%5BTitle%5D)+AND+%22Cognition%22%5BJournal%5D)">Listening to the calls of the wild: The role of experience in linking language and cognition in young infants.</a> Cognition. 153, 175-181 (2016).</li>
<li>Althaus, N., Mareschal, D. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((Labels+direct+infants%E2%80%99+attention+to+commonalities+during+novel+category+learning%5BTitle%5D)+AND+%22PLoS+ONE%22%5BJournal%5D)">Labels direct infants’ attention to commonalities during novel category learning.</a> PLoS ONE. 9 (7), e99670(2014).</li>
<li>Fulkerson, A. L., Haaf, R. A. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/pubmed?term=((The+influence+of+labels%2C+non-labeling+sounds%2C+and+source+of+auditory+input+on+9-+and+15-month-olds%E2%80%99+object+categorization%5BTitle%5D)+AND+%22Infancy%22%5BJournal%5D)">The influence of labels, non-labeling sounds, and source of auditory input on 9- and 15-month-olds’ object categorization.</a> Infancy. 4 (3), 349-369 (2003).</li>
</ol>

The authors have nothing to disclose.

Here, we present a procedure for evaluating the role of labeling in categorization. By presenting 2-year-olds with a realistic mix of labeled and unlabeled exemplars, we demonstrate that very young children are capable of learning in semi-supervised environments, extending work with adults and older children24,25. Thus, this method offers a resolution to the paradox posed above: if even a few labeled exemplars can spark category learning, then labels can be both ...

Categorization is a fundamental building block of human cognition: infants&#8217; categorization abilities emerge early in infancy and become increasingly sophisticated with age.1,2,3 Research has also revealed a powerful role for&#160;language in infant categorization: from 3 months of age, infants learn categories more successfully when category exemplars are paired with language.4,5,6 Moreover, by the end of the first year, infants are attuned to the role of count noun labels in categorization. Pairing category exemplars with a consistent labeling phrase (&#8220;This is a vep!&#8221;) facilitates infants&#8217; category learning relative to providing either a distinct label for each exemplar (&#8220;This is a vep,&#8221; &#8220;This is a dax,&#8221; etc.) or a non-labeling phrase (&#8220;Look at this.&#8221;).7,8,9
In infants&#8217; everyday experiences, however, the vast majority of objects they encounter will likely remain unlabeled. No caregiver could label every object an infant&#160;sees much less provide the labels which apply to every object (e.g., &#8220;malamute,&#8221; &#8220;dog,&#8221; &#8220;pet,&#8221; &#8220;animal&#8221;). This presents a paradox: how can we reconcile the power of labels in infant categorization with their relative scarcity in infants&#8217; daily lives?
To answer this question, we developed a protocol to assess how infants learn categories in a variety of different learning environments, including when they receive a mixture of labeled and unlabeled exemplars. Specifically, we propose that receiving even a few labeled exemplars at the beginning of learning can facilitate categorization&#8212;by enhancing infants&#8217; ability to learn from subsequent, unlabeled exemplars as well. This strategy of using a small number of labeled exemplars as a foundation for learning from a larger number of unlabeled exemplars has been widely implemented in the field of machine learning, spawning a family of semi-supervised learning (SSL) algorithms10,11,12. Of course, the learning strategies implemented are not identical across different kinds of learners: in machine learning, algorithms typically are exposed to many more exemplars, make explicit guesses about each exemplar, and learn multiple categories simultaneously. Nevertheless, both machine and infant learners may benefit from successfully integrating both labeled and unlabeled exemplars&#160;to learn new categories in sparse labeling environments.
Our design focuses on whether 2-year-old children, in the process of acquiring words for numerous new categories, are capable of this kind of semi-supervised learning. We employ a standard infant categorization measure: a familiarization-test task. In this paradigm, 2-year-olds were exposed to a series of exemplars from a novel category during a familiarization phase. Each exemplar was paired with a different auditory stimulus, depending on the condition (i.e., either a labeling or a non-labeling phrase). Then, at the test, all 2-year-olds saw two new objects presented in silence: one object from the now-familiar category and one from a novel category.
If the 2-year-olds successfully form the category during the familiarization phase, then they should distinguish between the two exemplars&#160;presented at the test. Importantly, because a systematic preference for either the novel or familiar test image reflects an ability to distinguish between them, both familiarity and novelty preferences are interpreted as evidence of successful categorization. Note that on a given task, the nature of this preference is a function of infants&#8217; processing efficiency for the stimulus materials, with familiarity preferences associated with less efficient stimulus processing 4,13,14,15,16,17. Presenting the test phase in silence makes it possible to directly assess infants&#8217; success in object categorization and how this success varies according to the information that accompanied the exemplars during familiarization. Thus, this paradigm provides a compelling test of how different types of linguistic environments affect category learning. If labeling enhances category learning in both semi-supervised and fully supervised environments, then 2-year-olds in these conditions should show stronger test preferences than infants in other environments.

<table><tbody><tr><td>Final Cut Pro X</td><td>Apple</td><td>N/A</td><td>Video editing, composition software</td></tr><tr><td>MorphX</td><td>Norrkross</td><td>N/A</td><td>Image-morphing software</td></tr><tr><td>PhotoShop</td><td>Adobe</td><td>N/A</td><td>Image-editing software</td></tr><tr><td>R</td><td>R Core Team</td><td>N/A</td><td>Statistical analysis software</td></tr><tr><td>T60XL Eyetracker</td><td>Tobii Pro</td><td>Discontinued</td><td>Large, arm-mounted eyetracker suitable for work with infants and children</td></tr><tr><td>Tobii Pro Studio</td><td>Tobii Pro</td><td>N/A</td><td>Software directing eyetracker display, data collection</td></tr></tbody></table>

defining the role of language in infants' object categorization with eye-tracking paradigms

Defining the Role Of Language in Infants' Object Categorization with Eye-tracking Paradigms

Assessing infant category learning is a challenging but vital aspect of studying infant cognition. By employing a familiarization-test paradigm, we straightforwardly measure infants&#8217; success in learning a novel category while relying only on their looking behavior. Moreover, the paradigm can directly measure the impact of different auditory signals on the infant categorization across a range of ages. For instance, we assessed how 2-year-olds learn categories in a variety of labeling environments: in our task, 2-year-olds successfully learned categories when all exemplars were labeled&#160;or the first two exemplars were labeled, but they failed to categorize when no exemplars were labeled or only the final two exemplars were labeled. To determine infants&#8217; success in such tasks, researchers can examine both the overall preference displayed by infants in each condition and infants&#8217; pattern of looking over the course of the test phase, using an eye-tracker to provide fine-grained time-course data. Thus, we present a powerful paradigm for identifying the role of language, or any auditory signal, in infants&#8217; object category learning.

Using the protocol above, we ran two experiments22. Analyses were conducted with the eyetrackingR package23, and the data and code are available at https://github.com/sandylat/ssl-in-infancy. In the first experiment, we contrasted a fully supervised condition (n = 24, Mage = 26.8 mo), featuring only labeled exemplars, with an unsupervised condition (n = 24, Mage = 26.9 mo), featuring only unlabeled ex...

Here we present a protocol for familiarization-test paradigms which provide a direct test of infant categorization and help to define the role of language in early category learning.

Assessing infant category learning is a challenging but vital aspect of studying infant cognition. By employing a familiarization-test paradigm, we ...

defining-role-language-infants-object-categorization-with-eye

Research

JoVE Journal

Behavior

A reliable home-based way to assess the language comprehension of very young typically developing children, as well as those with autism, is described.  The method analyzes children's eye gaze while viewing side-by-side images but hearing an audio that matches only one image.  Stimuli are designed with young participants in mind.

Portable Intermodal Preferential Looking (IPL): Investigating Language Comprehension in Typically Developing Toddlers and Young Children with Autism

At 3-4 months, listening to human and nonhuman primate vocalizations boosts infants' cognition; by 6 months, only human vocalizations exert this cognitive advantage. We describe an exposure manipulation that reveals the powerful shaping role of experience as infants specify which sounds to link to cognition and which to tune out.

Experience is Instrumental in Tuning a Link Between Language and Cognition: Evidence from 6-  to 7- Month-Old Infants' Object Categorization

The present article reviews an eye-tracking methodology for studies on language comprehension. To obtain reliable data, key steps of the protocol must be followed. Among these are the correct set-up of the eye tracker (e.g., ensuring good quality of the eye and head images) and accurate calibration.

Eye Tracking During Visually Situated Language Comprehension: Flexibility and Limitations in Uncovering Visual Context Effects

Infants and toddlers view the world in a fundamentally different way from their parents. Head-mounted cameras provide a tractable mechanism to understand the infant visual environment. This protocol provides guiding principles for experiments in the home or laboratory to capture the egocentric view of toddlers and infants.

A View of Their Own: Capturing the Egocentric View of Infants and Toddlers with Head-Mounted Cameras

We present a protocol to examine the use of morphological cues during real-time sentence comprehension by children with autism.

Using the Visual World Paradigm to Study Sentence Comprehension in Mandarin-Speaking Children with Autism

Young children do not passively observe the world, but rather actively explore and engage with their environment. This protocol provides guiding principles and practical recommendations for using head-mounted eye trackers to record infants' and toddlers' dynamic visual environments and visual attention in the context of natural behavior.

Gaze in Action: Head-mounted Eye Tracking of Children's Dynamic Visual Attention During Naturalistic Behavior

Eye tracking studies using a preferential looking paradigm can be used to study infants' emerging understanding of, and attention to, their external visual world.

Exploring Infant Sensitivity to Visual Language using Eye Tracking and the Preferential Looking Paradigm

The NICU Network Neurobehavioral Scale (NNNS) was developed as an assessment for the at-risk infant. The purpose of this article is to describe the NNNS, provide video examples of the NNNS procedures and discuss the ways in which the exam has been used.

Assessment and Evaluation of the High Risk Neonate: The NICU Network Neurobehavioral Scale

Methods to record auditory P50 sensory gating, a physiological marker of cerebral inhibition which reflects early stages of attention, are described.

P50 Sensory Gating in Infants

A method is described to quantify the quality of visual information processing based on reflexive eye movements in response to specific visual modalities. Reaction times and fixation output parameters are used to characterize visual performance in children with and without visual impairments from 6 months of age.

A Method to Quantify Visual Information Processing in Children Using Eye Tracking

Learn familiarization-test paradigms for infant categorization using eye-tracking and visual stimuli. This method reveals how infants form object categories and identifies conditions that enhance early learning.

Defining the Role Of Language in Infants' Object Categorization with Eye-tracking Paradigms

Explore More Videos