This pilot study evaluated the feasibility and diagnostic utility of the Eye Gaze Shift Task (EGST), a brief standardized observational task designed to elicit gaze-related social communication responses to ambiguous social cues, in 58 children (29 ASD, 29 non-ASD) aged over 31 months who completed the ADOS-2 Module 1.
EGST performance was strongly associated with ASD diagnosis (χ²=68.95, p<0.001, Cramér's V=0.93); ROC analysis showed AUC=0.989 (95% CI 0.97–1.00). At a cutoff score of ≥2, sensitivity was 100%, specificity was 89.7%, and overall classification accuracy was 94.8%.
- Very small pilot sample (n=58) limits generalizability. - Sample likely lacks demographic diversity (single institution, Louisiana). - No independent replication; authors both administered and scored the task, raising potential for evaluator bias.
The EGST shows strong preliminary diagnostic discrimination for ASD in young children and may complement the ADOS-2 by specifically targeting eye gaze shift behavior, but replication in larger, diverse samples is needed before clinical adoption.
Explore related topics
