A Machine Learning Algorithm to Assess Functional "Brain Age" from an In-Home EEG Sleepband

1 PROJECT SUMMARY2 Humans are living longer, resulting in our bodies outliving our brains. Although behavioral changes and better3 management of health conditions can help reverse or slow down premature decline in brain function, pre-4 symptomatic adults are not identified early enough when these changes would make the most significant positive5 impact. Effective intervention requires early detection before irreversible brain damage occurs, but there is a lack6 of objective, scalable tools to assess brain function in pre-symptomatic adults.78 Brain Age (BA) is a biomarker that can be used for early detection of deterioration in brain function. BA reflects9 an individual’s age-adjusted structural and/or functional brain characteristics and has been shown to detect 10 cognitive impairment. While effective, the cost and inconvenience of current assessment methods (e.g., MRI, 11 polysomnography) prevent widespread usage of BA as a biomarker. Therefore, there is a clear unmet need for 12 a new method for assessing BA. 13 14 NeuroGeneces’ BA machine learning (ML) model, using at-home sleep data, will provide an objective and 15 interpretable measure of age-adjusted brain function. In this Phase I STTR project, NeuroGeneces will expand 16 the dataset of sleep recordings by conducting a Human Subject Study and validate the feasibility of a ML model 17 that accurately predicts biological BA in cognitively healthy adults using an at-home sleep EEG headband.