Validation of Eye Gaze Response to Dynamic Social Stimuli As Biomarker Related to Social Communication for Clinical Trials Involving Children with ASD

Background: Eye-gaze tracking (EGT), a non-invasive and objective measure of an individual’s gaze response to visual stimuli, is easy to administer and has been shown to differentiate autism spectrum disorder (ASD) from other populations across the age span. It has been demonstrated that children with ASD show differential gaze patterns and attention when presented with socially salient stimuli, an important finding given that social communication deficits are a core feature of ASD. EGT has been proposed as a potential stratification or early efficacy biomarker in clinical trials assessing treatments for social communication impairments in children with ASD. However, its correlation with well-validated measures of social communication impairments has not yet been established.

Objectives: The study examined the degree to which eye-gaze patterns towards social stimuli are correlated with well-validated parent-report measures of social communication that are commonly used in clinical trials with children with ASD.

Methods: Participants were 25 children diagnosed with ASD based on the ADOS-2 and ADI-R (M age = 4.47 (+/- 1.10) years; M NVIQ = 64.3 (+/- 24.6)). EGT data were collected using the Tobii TX300 system at a sampling rate of 120 Hz. Participants were shown a three-minute video stimulus that included an actress and distractor toys (Chawarska et al. 2012). Two subjects were excluded due to noncompliance. Data analysis was constrained to time periods when the actress on the video directed speech toward the child. Dependent variables were proportion of time viewing the video presentation and proportion of time spent viewing actress. Analyses examined the correlations between the EGT variables and measures of social communication derived from the Vineland Adaptive Behavior Scales (Vineland-II), PDD Behavior Inventory (PDDB-I), and Behavior Assessment System for Children (BASC-2).

Results: The proportion of time viewing stimuli was highly correlated with the PDDB-I – Expressive Social Communication Abilities Composite score (r=0.72, p<0.001), PDDB-I – Expressive/Receptive Social Communication Abilities Composite score (r=0.68, p<0.001), and the Vineland-II – Expressive Communication Composite score (r=0.69, p<0.001). The proportion of time spent viewing the actress was strongly correlated with the BASC-2 – Functional Communication score (r=0.69, p<0.001), PDDB-I – Expressive Social Communication Abilities Composite score (r=0.73, p<0.001), PDDB-I – Expressive/Receptive Social Communication Abilities Composite score (r=0.68, p<0.001), and the Vineland-II – Expressive Communication Composite score (r=0.76, p<0.001). Notably, neither EGT measure was correlated with the PDDB-I Autism Composite score, a measure of broad ASD symptoms, (r=0.19 and 0.25, respectively).

Conclusions: Eye-gaze responses to dynamic social stimuli were found to be strongly and consistently correlated with well-validated parent-report measures of social communication. EGT variables did not correlate with broader autism symptom measures that include repetitive/restricted behaviors, suggesting that the EGT variables were specifically related to social communication impairments. Given the relative ease of using EGT with children with ASD, these findings support the ability of eye-gaze in response to dynamic social stimuli to serve as a non-invasive and objective biomarker of social communication in clinical trials. The present study is ongoing, and a final analysis will be conducted on a larger sample of children with ASD.