Autism Spectrum Disorder (ASD) is a neurodevelopmental condition characterized by challenges in social interactions, emotion recognition, and repetitive behaviors, exhibiting substantial individual variations. The Autism Diagnostic Observation Schedule (ADOS) is extensively utilized in ASD diagnosis; however, implementing the ADOS protocol is time-consuming, and accurate interpretation relies on the clinician's expertise. This study aims to explore neuronal correlates using EEG in a VR-based ASD assessment, analogous to ADOS. Three interactive game modules were developed for probing the social skills of communications, joint attention and emotion recognition (Fig. 1A). Fourteen typically developing (TD) and 14 ASD children were recruited. During the VR-based assessment, EEG data from FC3, FCZ, PZ, and OZ were recorded and the power spectra were analyzed.The preliminary results show that the δ, Ө, and α power at Oz were significantly lower in ASD when compared to that of TD during a task related to social skills (Fig. 1B). Alterations in δ power might reflect differences in cognitive processing in ASD. The decrease of Ө power in children with ASD might indicate the deficits of eye gaze, facial processing, or social cognition. Abnormalities in α oscillations might relate to the lesser coordination of wider brain areas. This study demonstrates the feasibility of employing EEG during VR-based ASD assessments and establishes a connection between behaviors and brain activities. The findings could assist in shaping the design of treatment strategies. Considering the substantial heterogeneity within ASD, conducting further studies with sufficient data is essential to enhance clarity in the results.