Analysis of EEG in Unipolar and Bipolar Depression Based on Phase Synchronization

Resting-state EEG studies have shown significant differences in functional connectivity networks between patients with unipolar and bipolar depression as well as healthy controls (HC). However, the present study has paid little attention to identify the EEG resting-state functional connectivity patterns of bipolar depression patients (BD). This paper proposes a method for unipolar and bipolar disorder recognition based on phase-synchronized feature fusion. From the perspective of time-frequency domain, phase lay index, phase locking value and weighted phase lay index and their mutual fusion were extracted, respectively. Feature matrix were sent to different machine learning classifiers to classify in the different bands. It can be seen from the classification results that the fusion feature PLV_ PLI _WPLI more effect than the other feature and the best performance in the beta band. Otherwise, the visualization results showed that the patients main brain regions were distributed in the frontal lobe, temporal lobe and parietal lobe compared with normal person. It revealed that new light on the pathological mechanism of BD and suggests that EEG resting-state phase synchronization analysis may identify potentially effective biomarkers for its clinical diagnosis in our study.