AUTHOR=Martinez Caroline , Chen Zhe Sage TITLE=Identification of atypical sleep microarchitecture biomarkers in children with autism spectrum disorder JOURNAL=Frontiers in Psychiatry VOLUME=Volume 14 - 2023 YEAR=2023 URL=https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2023.1115374 DOI=10.3389/fpsyt.2023.1115374 ISSN=1664-0640 ABSTRACT=Purpose Sleep disorders are one of the most frequent comorbidities in children with autism spectrum disorder (ASD). An improved understanding of etiology of sleep difficulties and identification of sleep-associated biomarkers for children with ASD can improve the accuracy of clinical diagnosis. Methods Sleep polysomnogram data was obtained from the Nationwide Children’ Health (NCH) Sleep DataBank. Children (age: 8-16 yr) with 149 autism and 197 age-matched controls without neurodevelopmental diagnosis were selected for analysis. An additional independent age-matched control group selected from the Childhood Adenotonsillectomy Trial (CHAT) and an independent smaller NCH cohort of younger infants and toddlers were used for additional validation. We computed periodic and nonperiodic characteristics from sleep EEG recordings. Machine learning models were trained using these features. We determined the autism class based on the prediction score of the classifier. The area under the receiver operating characteristics curve (AUC), accuracy, sensitivity, and specificity were used to evaluate the model performance. Results In the NCH study, the Random Forest Model outperformed two other models with a 10-fold cross-validated median AUC of 0.95 (interquartile range [IQR], [0.93,0.98]). Sleep spindle density, amplitude, spindle-slow oscillation (SSO) coupling, aperiodic signal’s spectral slope and intercept, as well as the percentage of REM sleep were found to be key discriminative features in the predictive models. Conclusions Our results suggest that integration of EEG feature engineering and machine learning can identify sleep-based biomarkers for ASD in children.