Riemannian Approach Based Depression classification using Transfer Learning for MEG signals

Major Depressive Disorder (MDD) is a mental disorder characterized by persistent feelings of sadness, loss of interest or pleasure, and a variety of physical and cognitive symptoms. This paper proposes a novel approach for the detection and classification of depressed subjects from healthy subjects using Riemannian geometry and transfer learning techniques. High dimensional 275-channel Magnetoencephalogram (MEG) provided by the BIOMAG 2022 and OpenNeuro datasets were used for the processing. Accurate and early detection of MDD is key to alleviating the stigma and encouraging individuals to seek help sooner, improving their chances of recovery, and our proposed methodology on 36 subjects' MEG data predicted them with 0.982 validation accuracy. This study discovered that by utilizing multidimensional MEG data, it is possible to distinguish between people with MDD and healthy people using Riemannian features and transfer learning techniques. Therefore, this technique may be a useful Machine Learning based automated tool for helping clinicians make the right identification well ahead of time.