Structural alterations as a predictor of depression - a 7-Tesla MRI-based multidimensional approach

Area of Science:

• Neuroimaging
• Psychiatry
• Machine Learning

Background:

• Major depressive disorder (MDD) is linked to alterations in the default-mode network (DMN), including gray matter volume (GMV) and gyrification.
• Ultra-high field strength MRI and machine learning offer potential for novel insights into depression pathophysiology.

Purpose of the Study:

• To investigate the predictive value of DMN structural features (GMV, cortical thickness, gyrification) for MDD using 7-T MRI and machine learning.
• To correlate structural changes with depression severity.

Main Methods:

• Acquired 7-T MRI data from 41 MDD patients and 41 controls.
• Parcellated DMN using Schaefer 600 Atlas.
• Employed mixed-model analysis, Support Vector Machine (SVM) with leave-one-out cross-validation, and permutation testing.

Main Results:

• SVM achieved 0.76 prediction accuracy for MDD using gyrification data.
• Cortical thickness was not a significant predictor.
• GMV did not predict MDD presence but correlated with disease severity in the left parahippocampal gyrus.

Conclusions:

• DMN structural data, particularly gyrification, holds potential for predicting MDD presence.
• Challenges remain in predicting disease course or treatment response due to GMV variance and gyrification's static nature.