Diminished Salience of Social Stimuli, Not Enhanced Salience of Nonsocial Stimuli in Young Children with ASD

Background:   Whether in vivo or during eye-tracking, toddlers with ASD show an attentional bias toward objects in scenes containing both people and objects.  Recent studies have suggested that contrast and motion may be processed atypically in young children with ASD (McCleery et al., 2009; Shic et al., 2009) and thus might play a role in attentional differences.

Objectives:   To assess attentional preferences for basic perceptual features in young children with ASD. Are they biased toward high-contrast stimuli or certain kinds of motion?  Does the presence of perceptually salient distractors impede their attention to social stimuli? 

Methods: Visual attention was examined in young children with ASD (n=36), DD (n=12) and TD (n=42), (M=25mo). (1) Contrast Salience Task: A preferential looking paradigm using identical images that differed only in levels of contrast or color (high and low) assessed preferences for high/low contrast and high/low color. (2) Motion Salience Task: A preferential looking paradigm assessed preferences for three kinds of motion: rotational, translational, and scaling.  (3) Dynamic Distractibility Task: Patterns of visual attention during a two-minute video were compared in young children with ASD (n=22) and TD (n=24), (M=25mo). The video consisted of a woman reading nursery rhymes with two monitors in the background that were blank or that contained still images (high contrast, high color) or video (rotating, scaling, or translating objects; Dynamic Distractors). Looking time for the Regions of Interest (ROI): face, book, body, distractors, and background were recorded with an eye tracker.

Results: (1) Contrast Salience: Children in all three groups preferred to look at images containing high color vs. low color and high contrast vs. low contrast at above-chance levels (p< .001), with no group differences in the magnitude of the preference. (2) Motion Salience: Children in all groups overwhelmingly preferred to watch rotational motion over other kinds of motion (p<.001), with no main effect of diagnosis.  (3) Dynamic Distractibility: Although children in both groups allocated most of their attention to the face, children with ASD attended longer to Dynamic Distractors than TD children, especially the rotating distractors, p<.01, d=1.21. The increase in looking time to the distractors came at the expense of looking at the face.

Conclusions: When tested in a preferential looking paradigm outside of the social context (experiments 1 & 2) toddlers with ASD have similar attentional preferences as non-ASD controls. When the same preferred stimuli (e.g., rotating objects) were presented in competition with highly salient social stimuli (experiment 3), non-ASD toddlers were capable of suppressing their attention to moving objects. Toddlers with ASD had difficulty doing so and ended up spending a significant amount of time looking away from the person. Consistent with other studies (Chawarska et al., 2003; 2010; Shic et al., 2011), these results indicate a diminished capacity to maintain attention to dynamic faces in young children with ASD which is not explained by increased salience of perceptual features of nonsocial distractors. Rather, the difficulties attending to social partners exhibited by children with ASD appear to result from diminished salience of social stimuli