A depression network of functionally connected regions discovered via multi-attribute canonical correlation graphs

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Psychiatry

Background:

• Major Depressive Disorder (MDD) is a prevalent mental health condition with complex neurobiological underpinnings.
• Resting-state functional magnetic resonance imaging (Rs-fMRI) is a key tool for investigating brain functional connectivity in psychiatric disorders.
• Current methods for analyzing Rs-fMRI data may not fully capture intricate brain network properties associated with MDD.

Purpose of the Study:

• To develop and validate a multi-attribute graph model for constructing high-resolution functional connectivity networks from Rs-fMRI data.
• To identify specific brain network alterations characteristic of major depressive disorder (MDD).
• To establish a network-based classifier for predicting MDD risk.

Main Methods:

• Utilized Rs-fMRI data from 20 MDD patients and 20 healthy controls from the PReDICT study.
• Developed a multi-attribute graph model to construct region-level functional connectivity networks, incorporating all voxel information.
• Defined connectivity strength using kernel canonical correlation coefficients and employed permutation testing for statistical significance.
• Constructed a network-based classifier for MDD risk prediction.

Main Results:

• The developed method successfully distinguished MDD patients from healthy controls.
• Significant differences in regional connectivity strength were identified between the two groups.
• Simulation studies confirmed the robustness and performance of the proposed analytical approach.

Conclusions:

• The multi-attribute graph model provides a sensitive method for detecting brain network alterations in MDD.
• Rs-fMRI analysis using this model can differentiate individuals with MDD from healthy controls.
• This approach holds potential for improving diagnostic accuracy and risk prediction in major depressive disorder.