Objective outcome prediction in depression through functional MRI brain network dynamics.

Area of Science:

• Neuroimaging
• Psychiatry
• Computational Neuroscience

Background:

• Subjective clinical decisions in major depressive disorder (MDD) treatment often lead to suboptimal effectiveness.
• Objective predictors are needed to improve treatment outcomes in MDD.
• Resting-state functional MRI (fMRI) offers potential for identifying such predictors.

Purpose of the Study:

• To identify objective predictors of MDD treatment outcome using resting-state fMRI.
• To evaluate the predictive capability of static and dynamic fMRI features for treatment response.
• To explore the utility of group independent component analysis (GICA) for network-level feature extraction.

Main Methods:

• Acquired resting-state fMRI scans from 25 MDD patients at baseline.
• Assessed patients every 3 months for a year, categorizing outcomes as positive or negative.
• Extracted static and dynamic fMRI features from GICA-identified networks and subnetworks.
• Utilized binary classifiers to predict MDD outcome at each follow-up.

Main Results:

• Total coherence, a measure of network interactivity, showed the highest predictive performance (AUC 0.70).
• Increased total coherence between the default mode network and ventral salience network was observed in the positive outcome group.
• Classification using total coherence alone achieved a high AUC (0.76 ± 0.10), indicating its strong discriminative capability.
• Optimal classification performance was achieved using higher GICA orders, dividing major networks into subnetworks.

Conclusions:

• Dynamic fMRI measure, total coherence, demonstrated superior classification performance for MDD outcome.
• Enhanced switching capability between internal and external brain states is a potential predictor of symptom improvement in MDD.
• These findings support the development of total coherence as a prognostic biomarker for MDD.