DATA-DRIVEN CLUSTER SELECTION FOR SUBCORTICAL SHAPE AND CORTICAL THICKNESS PREDICTS RECOVERY FROM DEPRESSIVE SYMPTOMS

Area of Science:

• Neuroimaging
• Psychiatry
• Machine Learning

Background:

• Major depressive disorder (MDD) patients not fully recovering with antidepressants face relapse risk.
• Electroconvulsive therapy (ECT) offers rapid, significant response in severe depression but predicting ECT success is difficult.

Purpose of the Study:

• To develop a predictive model for ECT remission in MDD patients using structural MRI data.
• To identify neuroimaging biomarkers associated with ECT treatment response.

Main Methods:

• A random forest classifier was trained on structural MRI data from 42 MDD patients undergoing ECT.
• The model utilized data-driven shape cluster selection and cortical thickness features.
• Predictive features included right hemisphere hippocampal shape and right inferior temporal cortical thickness.

Main Results:

• The classifier achieved an average balanced accuracy of 73% in predicting remission.
• Thicker right hippocampus and right inferior temporal cortex prior to treatment correlated with a decreased probability of remission.
• Right hemisphere hippocampal shape and right inferior temporal cortical thickness were the most predictive features.

Conclusions:

• Structural MRI features can aid in identifying MDD patients likely to respond to ECT.
• This approach shows promise for developing personalized treatment strategies in psychiatry.
• Neuroimaging-based classification may improve clinical decision-making for ECT candidates.