Abstract
There is broad interest in discovering quantifable physiological biomarkers for psychiatric disorders to aid diagnostic assessment. However, fnding biomarkers for autism spectrum disorder (ASD) has
proven particularly difcult, partly due to high heterogeneity. Here, we recorded fve minutes eyesclosed rest electroencephalography (EEG) from 186 adults (51% with ASD an 49% without ASD) and investigated the potential of EEG biomarkers to classify ASD using three conventional machine learning models with two-layer cross-validation. Comprehensive characterization of spectral, temporal and spatial dimensions of source-modelled EEG resulted in 3443 biomarkers per recording. We found no signifcant group-mean or group-variance diferences for any of the EEG features.Interestingly, we obtained validation accuracies above 80%; however, the best machine learning model merely distinguished ASD from the non-autistic comparison group with a mean balanced test accuracy of 56% on the entirely unseen test set. The large drop in model performance between validation and testing, stress the importance of rigorous model evaluation, and further highlights the high heterogeneity in ASD. Overall, the lack of signifcant diferences and weak classifcation indicates that, at the group level, intellectually able adults with ASD show remarkably typical resting-state EEG.
proven particularly difcult, partly due to high heterogeneity. Here, we recorded fve minutes eyesclosed rest electroencephalography (EEG) from 186 adults (51% with ASD an 49% without ASD) and investigated the potential of EEG biomarkers to classify ASD using three conventional machine learning models with two-layer cross-validation. Comprehensive characterization of spectral, temporal and spatial dimensions of source-modelled EEG resulted in 3443 biomarkers per recording. We found no signifcant group-mean or group-variance diferences for any of the EEG features.Interestingly, we obtained validation accuracies above 80%; however, the best machine learning model merely distinguished ASD from the non-autistic comparison group with a mean balanced test accuracy of 56% on the entirely unseen test set. The large drop in model performance between validation and testing, stress the importance of rigorous model evaluation, and further highlights the high heterogeneity in ASD. Overall, the lack of signifcant diferences and weak classifcation indicates that, at the group level, intellectually able adults with ASD show remarkably typical resting-state EEG.
Original language | English |
---|---|
Article number | 19016 |
Journal | Scientific Reports |
Volume | 12 |
Number of pages | 14 |
ISSN | 2045-2322 |
DOIs | |
Publication status | Published - 2022 |