Eye movements in face recognition in children with autism spectrum disorder, and their relationships with different sub-clusters of autistic traits

Abstract

22typically developing (TD) and 22 ASD children (aged 6-11) participated in the study. Parents rated children’s behaviors using AQ-Child Questionnaire and the items were classified into five sub-clusters of autistic traits. Both groups of participants completed a computer face recognition task and their eye movements were tracked using Red-N portable eye-tracker. Hidden Markov model (HMM) toolbox (EMHMM) was used to analyze eye movement data in face recognition task. Participants’ eye movement patterns are modelled by an HMM, and clustered into analytic and holistic eye movement patterns. Results indicate that the TD and ASD groups showed comparable face recognition performance and their eye movement patterns did not differ significantly. However, it was found that ASD children with lower AQ-Total, AQ-Social and AQ-Imagination scores (which indicate less autistic deficits in those areas) performed better in face recognition. ASD children with higher AQ-Attention score was more likely to be classified as having analytic eye movement pattern (more looking at the eyes) but they did not show better face recognition performance. Further research could be done to investigate how autistic traits contribute to face recognition and eye movement in other age groups or individuals with different cognitive abilities.