Exploring Oscillatory Dysconnectivity Networks in Major Depression During Resting State Using Coupled Tensor...

Area of Science:

• Neuroscience
• Computational Psychiatry
• Network Science

Background:

• Major depression disorder (MDD) is associated with large-scale brain network dysconnectivity during resting state.
• The temporal dynamics of brain networks, influenced by oscillations, offer insights into MDD's neural underpinnings.

Purpose of the Study:

• To investigate brain network dysconnectivity in MDD using a novel coupled tensor decomposition model.
• To characterize spatio-temporal-spectral modes of covariation in resting-state electroencephalography (EEG) data from MDD patients.

Main Methods:

• Applied a novel double-coupled nonnegative tensor decomposition model with low-rank approximation.
• Utilized the phase lag index to compute functional connectivity within time-frequency windows.
• Employed a fast hierarchical alternative least squares algorithm for model optimization and clustering analysis.

Main Results:

• Identified four distinct oscillatory networks characterizing the healthy group.
• Identified four distinct oscillatory networks characterizing the major depression group.
• The model revealed specific spatio-temporal-spectral patterns of dysconnectivity in MDD.

Conclusions:

• The proposed tensor decomposition model offers novel insights into the pathoconnectomics of MDD during resting state.
• The findings highlight the role of oscillatory network dynamics in MDD.
• The model is adaptable for studying other psychiatric disorders.