Altered Cortical Morphological Brain Networks and Their Diagnostic Classification Utility in Major Depressive...

Area of Science:

• Neuroscience
• Psychiatry
• Medical Imaging

Background:

• Major depressive disorder (MDD) is increasingly viewed as a brain dysconnectivity syndrome.
• Previous studies on MDD brain networks were limited by small sample sizes or single-feature analyses.

Purpose of the Study:

• To investigate individual-level cortical morphological networks in a large cohort of MDD patients and controls.
• To explore alterations in multi-feature morphological connectivity (MC) and nodal properties in MDD.
• To assess the diagnostic potential of these network alterations using machine learning.

Main Methods:

• Utilized structural MRI data from 1442 MDD patients and 1277 controls.
• Constructed individual-level cortical networks based on cortical thickness (CT), cortical volume (CV), surface area (SA), and sulcal depth (SD).
• Performed group comparisons of MC and graph-theoretical nodal properties, and applied Support Vector Machine (SVM) for classification.

Main Results:

• MDD patients showed widespread MC alterations with distinct patterns across the four morphological networks.
• CT networks exhibited reduced MC in higher-order networks; CV networks showed increased MC within the default mode network.
• SA and SD networks displayed enhanced MC in lower-order networks; increased betweenness centrality was observed in multiple cortical regions.
• SVM models achieved moderate-to-good classification performance in distinguishing MDD patients from controls.

Conclusions:

• Individual-level morphological network alterations support the dysconnectivity hypothesis of MDD.
• Distinct patterns of altered brain morphology across different features offer potential for improved MDD diagnosis.
• Findings highlight the complexity of brain connectome alterations in major depressive disorder.