Infant neural sensitivity to dynamic eye gaze is associated with later emerging autism

Abstract

Autism spectrum disorders (henceforth autism) are diagnosed in around 1% of the population [1]. Familial liability confers risk for a broad spectrum of difficulties including the broader autism phenotype (BAP) [2, 3]. There are currently no reliable predictors of autism in infancy, but characteristic behaviors emerge during the second year, enabling diagnosis after this age [4, 5]. Because indicators of brain functioning may be sensitive predictors, and atypical eye contact is characteristic of the syndrome [6-9] and the BAP [10, 11], we examined whether neural sensitivity to eye gaze during infancy is associated with later autism outcomes [12, 13]. We undertook a prospective longitudinal study of infants with and without familial risk for autism. At 6-10 months, we recorded infants' event-related potentials (ERPs) in response to viewing faces with eye gaze directed toward versus away from the infant [14]. Longitudinal analyses showed that characteristics of ERP components evoked in response to dynamic eye gaze shifts during infancy were associated with autism diagnosed at 36 months. ERP responses to eye gaze may help characterize developmental processes that lead to later emerging autism. Findings also elucidate the mechanisms driving the development of the social brain in infancy.

Figure 1

Association between Infant ERPs in Response to Eye Gaze and Autism Outcomes (A) Participating families first visited the lab when their infants were 6–10 months of age. Electrophysiological recording was done during this visit. Infants were prepared for the EEG session. (B) Electrophysiological response to gaze shifts over occipitotemporal channels. (C) Around 2 and 3 years of age, the same infants were tested by an independent team using several measures including the ADOS, a semistructured observational measure of autism-related characteristics. Based on information from all visits, combined with expert clinical judgment, infants in the at-risk group were classified as having ASD or not. (D) Controlling for age at the first visit, significant condition × risk-group interactions were observed for the amplitude of the P400 [F(1,92) = 6.7, p = 0.01]; planned post hoc tests focused on within-group difference between response to direct versus averted gaze controlling for age at baseline and developmental level at 36 months. Estimated mean differences between responses to gaze toward versus away are displayed for each group (standard error bars are displayed). Findings suggest that differentiation between gaze toward versus away was reliable in the both the control group (p < 0.001) and the at-risk without ASD group (p = 0.04). By contrast, the at-risk group that developed ASD showed no differentiation (p = 0.67) nor did the subgroup that developed early and persistent symptoms (p = 0.27). Findings from static face and face versus noise contrasts are presented in Figure S1 and Table S1.