Quantization on EEG covariance matrix images during TOVA attention test for depression disorder classification

摘要

The major depression disorder (MDD) is the mental disorder that causes patients to lose their life ability. It is important to develop a computer-aid diagnosis method based on electro-encephalography (EEG) screening. Test of variables of attention (T.O.V.A.®) was applied in this research and the surface EEG was recorded simultaneously. The proposed method quantizes covariance matrix images of EEG channel interactions by combining both approximate entropy and 2D Fourier transform methods. The higher approximate entropy values were found at MDD group and lower high-frequency components were found in the control group. Overall, the research classified MDD patients and normal population with high accuracy. The covariance matrix images from TOVA target sessions offered better distinguishability on depressive disorder patients than TOVA non-target sessions.