A Novel & Hybrid Algorithm for Real Time Depression Detection via EEG Analysis

Depression is among the most common mental health conditions, significantly affecting an individual’s wellbeing and productivity. Detecting depression effectively requires early and accurate identification to enable timely intervention and achieve positive outcomes. This research introduces a hybrid model that integrates multiple Feedforward Neural Networks (FNNs) with a Shallow Neural Network (SNN) for depression detection using electroencephalogram (EEG) data. EEG signals are analyzed across key frequency bands including gamma, high beta, beta, alpha, theta and delta. The hybrid architecture employs FNNs to extract complex, nonlinear features from the data, while the SNN leverages these outputs to efficiently identify simpler patterns and improve classification performance. The model’s effectiveness is thoroughly evaluated based on its accuracy, precision, recall, F1-score, F2-score demonstrating its effectiveness in detecting neural abnormalities associated with depression. The findings emphasize the proposed hybrid system’s potential as a reliable and non-invasive diagnostic tool for mental healthdiagnosis, offering new possibilities for advanced and proactive care strategies.