EEG-based MG-GCN model for Major Depressive Disorder detection. Uses multi-scale graph representations to capture spatial and frequency features of brain signals, modeling local and global neural connections. Developed in MATLAB, it outperforms traditional ML and deep learning approaches.