Object size in view changed through infants’ developmental trajectory and growth of vocabulary

• In Event: Attention, learning and memory in infancy

Abstract

An ostensive definition of typical early language learning is associating speech sound and the referred object in the visual scene (Gogate, Bolzani & Betancourt, 2006). In recent years, studies using head-mounted eye-tracker systems recorded infants’ eye gaze behaviors during interactive play and documented various visual experiences that may support in generating the association during a natural activity. These include infants’ sustained attention to objects (Yu & Smith, 2012) parent-child’s joint attention to an object (Pereira, Smith & Yu, 2014); infants’ isolated view on a single object through parents’ holding (Yoshida & Smith, 2008), etc. These visual experiences have also been linked to the promotion of infants’ language learning (Smith, Yu & Pereira, 2011). While these studies typically consider the fixation on an object as looking at everything of that object, previous studies concerning early visual early capacity also suggest that infants are capable of focusing on parts of an object during such fixation (Rakison & Cohen, 1999; Althaus & Mareschal, 2014). Therefore, current interpretation of early fixations as “seeing” whole object may leave out the visual properties makeups of the associations that are critical to language learning. To address this potential gap, the current study borrows the notion of foveation from vision science and uses the size of the fixated object in the infants’ view to estimate (1) how much object information is taken and (2) how this amount of information change as a function of infant’s age and/or vocabulary size.
Thirty-seven 6-18-month-old parent-child dyads participated in a parent-child semi-naturalistic object play. Eight toys were given for free playing: bunny, bear, basket, car, carrot, cookie, cup, jar. Eight words were played as audio prompts every 40 seconds: bunny, eat, open, drink, bear, car, cookie, put. Gaze data were collected throughout the session.
Polynomial regression was used to test the effect of infant’s age and vocabulary size on fixated object in view (in 1k pixels). Results show that fixated object size reduces through ages (Myoung=38, SDyoung=2.3; Mold=31.4, SDold=1.9), critically reaching the lowest point during the 12-15-month interval. Though the size increases slowly after 15 months, it remains lower than which of younger age group (Mbefore15=35.1, SD=0.9; Mafter15=32.7, SD=1.7). The result is significant for all toys except the basket (pbasket=0.0586, t=-1.9). Supposedly, the word, basket, is the late learned word in infants’ vocabulary checklist among the given toys. However, across 8 toys, only bunny, bear, cookie, cup’s size significantly negatively correlated by vocabulary size. Interestingly, the size of all fixated objects decreases once the child learned that object’s name or an action verb associating with that object.
The preliminary study results revealed that size of fixated area within seen object dynamically changes, and the amount of information taken in at different developmental milestones is also influenced by child’s familiarity of the object. Size reduction might indicate a shift from local to holistic visual processing. We are currently conducting time-to-event analysis on fixation pattern specifically on objects to further explore if other factors such as object properties effect on visual experiences and learning.

Authors

• Giang Le, University of Houston
  Presenting Author

• Hanako Yoshida, University of Houston