Identifying Neural Biomarkers of Risk-Taking from Intracranial EEG Recordings

Abstract

Risk-taking behavior is associated with neuropsychiatric disorders such as addiction. In this study, we present a data-driven approach to identify neural biomarkers of risk-taking from intracranial stereo-electroencephalography (sEEG) recordings. Using time and frequency domain features, we train a classifier to distinguish between risk-taking and risk-avoidance behaviors. Based on data from a single patient, the model achieves an accuracy of 68.5%, which is significantly above chance. These results highlight the potential of identifying risk-taking states from invasive recordings. Clinical SignificanceThis work demonstrates the feasibility of identifying risk-related biomarkers from intracranial recordings, highlighting a promising direction for closed-loop neuromodulation for neuropsychiatric conditions.