Resting EEG biomarkers

Abstract

Autism spectrum disorder (ASD) is characterized by atypical brain activity, yet early and accurate diagnosis remains challenging. This study proposes a novel EEG-based method to detect ASD by analyzing hemispheric energy asymmetry during the resting state. EEG data were recorded from 44 children (22 ASD, 22 controls) at medical city, Baghdad. Signals were pre-processed using decimation, band-pass filtering, artifact subspace reconstruction, detrending, and normalization. A sliding window (8 s, 12.5% overlap) extracted maximum theta wave energy from each hemisphere. Results revealed greater right-hemispheric energy in ASD children, particularly enhanced in the Theta band. The method achieved 100% ASD detection accuracy for both genders, while control classification yielded 77% accuracy for males and 55.6% for females, resulting in an overall control accuracy of 68.1%. Total accuracy across groups was 84% (88.5% males, 77.8% females). These findings highlight right-hemispheric dominance in ASD and the need for gender-specific considerations in EEG-based diagnosis.