Abstract
Previous studies have demonstrated that faces increasingly capture infants' visual attention over the course of the first year of life. However, since most studies have used relatively simple stimuli, far less is known about the developmental changes in attention to faces in more naturalistic contexts and the degree to which infants prefer faces over competing visual information (e.g. low-level saliency). In the current study, remote eye-tracking techniques coupled with a free-viewing paradigm were adopted in order to investigate the development of more naturalistic visual orienting toward socially meaningful stimuli. Infants from two longitudinal cohorts (cohort one: 3.5, 5, 6.5 months; cohort two: 6, 8, 10, 12 months) and adults were presented with nine naturalistic videos which displayed three characters performing baby-friendly actions, such as waving a balloon or dancing with a toy. By establishing a face model, which assumed that attention would be directed toward locations containing faces, and a saliency model, which assumed that fixations would be driven by regions of high saliency, the predictive power of each model in determining fixation locations was evaluated. As predicted, the saliency model was a significantly better fit only for the youngest age group (3.5 months). Although neither model was a better predictor for fixations of infants aged 5 and 6.5 months (cohort one) or 8 months (cohort two), the 6-month group (cohort two) looked significantly more at faces than salient regions. This suggests that 5 to 8 months is a transition point at which individual differences in attention to faces may be high, possibly due to differences in the development of executive function and inhibitory control across individuals. The face model better accounted for fixations of 10- and 12-month-old infants and adults, suggesting that a clear face preference was established toward the end of the first year.
Meeting abstract presented at VSS 2016