Discriminating between major depressive disorder and bipolar depression: Aberrant EEG microstate dynamics and machine...

Area of Science:

• Neuroscience
• Psychiatry
• Computational Neuroscience

Background:

• Major depressive disorder (MDD) and bipolar depression (BD) share clinical features, complicating diagnosis and treatment.
• Electroencephalography (EEG) microstates offer insights into large-scale neural network dynamics.
• EEG microstates may reveal neural abnormalities in mood disorders.

Purpose of the Study:

• Investigate EEG microstate differences between MDD, BD, and healthy controls (HCs).
• Assess the utility of EEG microstates in classifying MDD versus BD, MDD versus HCs, and BD versus HCs using machine learning.

Main Methods:

• Analyzed resting-state EEG microstate features from 210 participants (78 MDD, 45 BD, 87 HCs).
• Utilized machine learning models to classify participants based on EEG microstate data.
• Examined microstate metrics (e.g., duration, transition probabilities).

Main Results:

• MDD patients exhibited altered microstate C and D metrics and transition probabilities, suggesting specific network engagement.
• BD patients displayed prolonged microstate B duration, indicating excessive visual network activity.
• Machine learning models achieved moderate to good classification performance (AUCs: 83.4% MDD vs. BD, 86.0% MDD vs. HCs, 93.3% BD vs. HCs).

Conclusions:

• EEG microstates show potential as a diagnostic biomarker for differentiating MDD and BD.
• Findings provide preliminary insights into the neural underpinnings of MDD and BD.
• Future research can refine methods for enhanced classification and targeted interventions.