Diagnosis of Major Depressive Disorder Based on Individualized Brain Functional and Structural Connectivity

Area of Science:

• Neuroimaging
• Machine Learning
• Psychiatry

Background:

• Traditional neuroimaging focuses on group-level analysis, overlooking individual brain connectivity differences.
• Understanding individual variations in brain connectivity is crucial for deciphering major depressive disorder (MDD) mechanisms.
• There is a growing interest in individual-specific brain connectivity for clinical applications.

Purpose of the Study:

• To integrate individualized functional connectivity (IFC) and structural connectivity (ISC) using machine learning.
• To develop a model capable of distinguishing individuals with MDD from healthy controls (HCs).
• To explore the utility of multimodal fusion for enhanced MDD classification.

Main Methods:

• Prospective study involving 182 MDD patients and 157 HCs, with a verification cohort.
• rs-fMRI and DTI data were used to construct functional and structural brain networks.
• Individualized connectivity features were extracted and fused using multimodal canonical correlation analysis with joint independent component analysis (mCCA+jICA), followed by Support Vector Machine (SVM) classification.

Main Results:

• Multisequence fusion of IFC and ISC significantly improved MDD classification performance from 72.2% to 90.3%.
• The developed prediction model demonstrated significant power in assessing depression severity in MDD patients (r=0.544).
• The individualized connectivity approach showed superior performance compared to traditional methods.

Conclusions:

• Integrating IFC and ISC via multisequence fusion enhances the identification of MDD.
• The individualized approach offers significant advantages in MDD research and clinical assessment.
• This study underscores the importance of personalized neuroimaging in understanding and diagnosing psychiatric disorders.