Predicting Autistic Traits Using Eye Movement during Visual Perspective Taking and Facial Emotion Identification.

Abstract

Autistic traits are broad in severity and difficult to measure quantitatively. Quantitative measurement would be helpful in determining the effectiveness of training and therapy for autistic traits. The development of eye-tracking technology has made it easier to understand autistic traits. Previous works showed that autistic traits can be predicted from eye movements in the facial emotion identification task. It may be possible to measure autistic traits more accurately than the only prediction from facial emotion identification ability. In this study, we used a visual perspective taking task. The results showed that the Social Responsiveness Scale-2 score, which is associated with autistic traits, was predicted at 0.414 in Spearman's correlation coefficient by using eye movements obtained from the two tasks.