Mental Illnesses Detection Employing Features Extracted From EEG Signals Using Deep Learning

Mental illnesses have become a significant global health concern. Early detection of these conditions is challenging and expensive with current clinical approaches. Recently, the utilization of electroencephalography (EEG) for detecting mental illnesses has gained considerable attention. This study investigates the feasibility of the use of a dataset consisting of features extracted from EEG signals to train two deep learning models, CNN and biLSTM to detect mental health disorders, specifically focusing on mood disorders, schizophrenia, obsessive-compulsive disorder (OCD) and healthy subjects. Features, such as brain waves, alpha, beta, high beta, delta, gamma, and theta, are used. Results showed that the accuracies achieved by the CNN and biLSTM models were 76% and 74% respectively.