Brain Functional Networks Based on Resting-State EEG Data for Major Depressive Disorder Analysis and Classification

Area of Science:

• Neuroscience
• Computational Psychiatry
• Biomedical Engineering

Background:

• Major depressive disorder (MDD) diagnosis often relies on subjective assessments and traditional electroencephalography (EEG) analysis methods overlook crucial brain network correlations.
• Analyzing the brain as a complex system requires advanced methods to capture topological architecture alterations.

Purpose of the Study:

• To develop and validate a brain functional network framework for analyzing and classifying MDD using resting-state EEG.
• To identify potential neuroimaging biomarkers for MDD detection and severity assessment.

Main Methods:

• Constructed functional brain networks from 64-channel resting-state EEG using the phase lag index (PLI) to mitigate volume conductor effects.
• Applied network binarization based on small-world indices and performed statistical analyses across EEG frequency bands and brain regions.
• Correlated network metrics (average shortest path length, clustering coefficient, node betweenness centrality) with the PHQ-9 depression severity score.

Main Results:

• Significant alterations in brain synchronization were observed in specific regions of the left and right hemispheres.
• Key network metrics in the theta band (left central) and right parietal-occipital regions showed significant correlation with MDD severity (PHQ-9 scores).
• The proposed framework achieved a classification accuracy of up to 93.31% for distinguishing MDD patients from healthy controls.

Conclusions:

• The developed brain functional network framework provides a robust method for MDD analysis and classification using resting-state EEG.
• Specific network metrics demonstrate potential as biomarkers for MDD detection and severity assessment.
• MDD patients exhibit a trend towards a more random brain network structure with diminished small-world properties.