Bosl WJ, Tager-Flusberg H, Nelson CA. EEG Analytics for Early Detection of Autism Spectrum Disorder: A Data-driven Approach. Sci Rep. 2018;8(1):6828. DOI: 10.1038/s41598-018-24318-x.
Autism spectrum disorder (ASD) is defined by a broad array of behavioral symptoms with an increasingly diverse etiology and developmental course. Formal diagnosis before age 3 years is challenging because it is behaviorally, rather than biologically, diagnosed. In this study of almost 200 children in Boston followed longitudinally, EEG measurements as early as 3 months of age were strongly predictive of the diagnosis of ASD by age 3 years. This finding suggests that EEG “digital biomarkers” may be helpful in ASD diagnosis in early life.
Sibling studies have illustrated that behavioral signs of ASD often appear as early as the first year of life. In this study, 99 infant siblings of older children with a diagnosis of ASD ("high-risk" group) were enrolled beginning at 3 months of age; 89 "low-risk" infants with at least one typically developing older sibling and no first-degree relatives with ASD were also enrolled as a control group. Children were evaluated with established clinical tools including the Autism Diagnostic Observation Schedule (ADOS; which also quantifies severity of symptoms) and with EEG. The authors found that EEG measurements distinguished high-risk ASD-positive children (diagnosed by ADOS) from ASD-negative children in the first year of life and continuing to age 3 years, with positive predictive values of .88 or greater at multiple age cutoffs. EEG findings were also associated with severity of ASD symptoms. The greatest challenge in using early-life EEG measurements was in distinguishing ASD-positive versus ASD-negative children in the high-risk group. Future research will need to test these approaches in a larger and more diverse sample.