Non-Invasive Classification of Mental Health Disorders Using Resting-State EEG with Dry Electrodes for Scalable Triage

Abstract The purpose of this study is to develop a model utilizing resting-state EEG data acquired through dry electrodes to classify individuals with and without mild psychopathological symptoms for triage purposes. The proposed methodology uses a Random Forest model. Coherence connectivity was identified as the most predictive feature, achieving an accuracy of 73% in distinguishing between epochs. The classifier is applied to sequences of epoch-level predictions for each file, demonstrating file-level classification and achieving an accuracy of 86% on the test files. These findings highlight the potential of non-invasive, dry electrode EEG and Machine Learning approaches as scalable tools for mental health disorder classification and clinical triage.