Virtual Reality Music-Based Game as a Co-Tool to Diagnose Autism in Young Children

Autism Spectrum Disorder (ASD) is a neurodevelopmental condition that impacts various aspects of life, including motor skills, social interactions, and communication, with symptoms that differ widely among individuals. Early diagnosis is critical, especially in young children, as timely intervention can significantly improve outcomes. This paper presents a novel approach utilizing a music-based virtual reality (VR) game to assist therapists in differentiating children with ASD from their typically developing (TD) peers. A classification model was trained using data collected from 21 TD children and 9 children with ASD, focusing on their actions and movements within the game. The model achieved an accuracy of 93% and a sensitivity of 85%, showcasing the potential of VR games in advancing ASD/TD classification and autism screening.