Development of a visual attention based decision support system for autism spectrum disorder screening

Area of Science:

• Developmental psychology
• Computational neuroscience
• Biomedical engineering

Background:

• Autism spectrum disorder (ASD) diagnosis in young children remains challenging.
• Visual attention patterns have been previously identified as potential indicators in ASD.
• Early identification of ASD is crucial for timely intervention and improved outcomes.

Purpose of the Study:

• To develop and evaluate a Decision Support System (DSS) utilizing machine learning (ML) for early ASD identification.
• To assess the efficacy of visual attention as a biomarker in distinguishing young children with ASD from typically developing (TD) peers.
• To analyze the performance of the ML-based DSS in early ASD assessment.

Main Methods:

• Development of a DSS employing machine learning algorithms.
• Recruitment of young children aged 26–36 months, including those with ASD (n=61) and TD children (n=72).
• Analysis of visual attention patterns as input features for the ML model.

Main Results:

• The proposed DSS demonstrated a high success rate of up to 87.5% in the early assessment of ASD.
• Machine learning techniques effectively differentiated between young children with ASD and TD children based on visual attention.
• Visual attention was confirmed as a significant and promising biomarker for early ASD detection.

Conclusions:

• The developed DSS shows significant potential for aiding in the early identification of ASD in young children.
• Visual attention serves as a valuable and unique biomarker for early ASD assessment.
• Further research is warranted to refine the DSS and explore its clinical applicability.