Early Attentional Processing of Affective Faces in Toddlers with ASD

Background: Toddlers with ASD view scenes with human faces atypically, choosing instead to divert their visual attention to less social aspects of presented scenes (Chawarska & Shic, 2009; Jones, Carr, & Klin, 2008; Pierce, Conant, Hazin, Stoner, & Desmond, 2011; Shic, Bradshaw, Klin, Scassellati, & Chawarska, 2011). It is not clear whether these atypical gaze patterns are associated with early attentional biases to social information; similarly it is unknown if limited attention to neutral faces extends to emotional faces.

Objectives: To examine the effect of facial affect on attention capture and looking time in toddlers with ASD and typical development.

Methods: Toddlers with ASD (N=57; Age: M=21.1, SD=2.1 months) and typically developing (TYP) toddlers (N=66; Age=19.5, SD=2.1 months) were tested using a preferential looking eye-tracking paradigm examining attention capture by and preferences for: 1) upright vs. inverted faces (4 trials); 2) happy vs. neutral faces (2 trials); and 3) fearful vs. neutral faces (2 trials). The child’s own anxious apprehension in a context of play-based interactions was estimated based on the global summary score (E3) on the ADOS-G.  A linear mixed-model diagnosis (2) by condition (3) analysis was conducted for: a) likelihood of the first saccade to be directed towards the target (attention capture); b) looking time ratio at the target.

Results : The attention of TYP, but not ASD, toddlers was captured by upright (p<.05) and fearful faces (p<.01).  Interestingly, both ASD and TYP groups looked longer at the target in the Fearful condition compared to other conditions (p<.01); in the Fear condition the total time spent looking at fearful faces was greater than chance (.5) for both ASD and TYP groups (p<.001). However, toddlers with ASD looked less at fearful faces as compared to TYP controls (p<.05). Greater levels of anxiety as observed on the ADOS-G in the ASD group were associated with longer looking time at the fearful faces (r = .33, p<.05).

Conclusions: Consistent with results in older children and adults, faces in general and affective faces in particular capture the attention of TYP toddlers, but not toddlers with ASD, suggesting impairments in an elementary attentional bias toward these salient stimuli. Fearful expressions elicited greater attention in both groups; though the response in ASD was less pronounced. The level of the child’s own fearful responses to play probes during the ADOS-G was positively associated with the duration of looking at the fearful face.  Thus, while basic mechanisms related to sustained attention to fearful faces may be intact in autism, toddlers with ASD are likely to process information about fearful faces differently than age-matched controls.