Research on Depression Recognition Based on EEG Microstate Functional Connectivity.

Area of Science:

• Neuroscience
• Computational Psychiatry
• Biomedical Engineering

Background:

• Major Depressive Disorder (MDD) diagnosis relies on subjective criteria.
• Objective biomarkers for MDD are needed to improve diagnostic accuracy.
• Electroencephalogram (EEG) offers a non-invasive window into brain activity.

Purpose of the Study:

• To investigate differences in dynamic functional connectivity using EEG microstate analysis between MDD patients and healthy controls (HC).
• To develop and validate a classification approach for identifying MDD using EEG-derived network features.
• To enhance the effectiveness of depression identification through objective physiological indicators.

Main Methods:

• Combined EEG microstate analysis with functional connectivity network construction.
• Analyzed resting-state EEG data from 19 MDD patients and 17 HC.
• Extracted topological characteristics (node degree, clustering coefficient, local/global efficiency) from phase locking value (PLV) networks derived from microstates A and C.
• Fused significant group-discriminative network features and assessed classification performance using SVM, BP, and KNN models.

Main Results:

• Network features from microstate C demonstrated superior discriminative ability.
• Node degree features showed the highest accuracy among individual topological attributes.
• The K-Nearest Neighbors (KNN) model achieved 96.48% accuracy using node degree.
• A fused feature set integrating comprehensive EEG information improved classification accuracy to 97.35% across all models.

Conclusions:

• Dynamic brain network analysis effectively distinguishes MDD patients from HC.
• The study provides a foundation for understanding brain region dynamics in depression.
• Objective physiological indicators derived from EEG show potential for MDD diagnosis.