Feasibility of a Novel Eye Gaze Shift Task in Autism Assessment: A Pilot Study

Eye gaze is a fundamental component of social communication and is frequently evaluated during Autism Spectrum Disorder (ASD) assessments. Although behavioral assessments of ASD frequently include observations of eye gaze, relatively few standardized tasks are designed to specifically elicit gaze-related social communication in response to an ambiguous social cue. This pilot study examined the feasibility and preliminary diagnostic utility of the Eye Gaze Shift Task (EGST). Data from 58 children (29 ASD; 29 non-ASD) who completed the ADOS-2 Module 1 were analyzed. The EGST was administered concurrently during comprehensive clinical evaluations and scored independently using a standardized 4-point observational hierarchy modeled after the ADOS-2 hierarchy of presses. Chi-square and receiver operating characteristic (ROC) analyses were conducted to examine the relationship between EGST performance and final clinical diagnosis. EGST performance was significantly associated with ASD diagnosis, χ²(1, N = 58) = 68.95, p < 0.001, with a very large effect size (V = 0.93). ROC analyses demonstrated excellent diagnostic discrimination (AUC = 0.989, 95% CI [0.97-1.00]). Using a cutoff score of ≥ 2, the EGST demonstrated 100.0% sensitivity, 89.7% specificity, and overall classification accuracy was 94.8%. Findings provide preliminary evidence that the EGST may be a feasible, brief, standardized observational task that may complement existing ASD assessment procedures by providing additional information regarding eye gaze behavior in children over 31 months of age. Replication of these findings with larger and more diverse samples is recommended.