Related Experiment Video
Updated: Aug 6, 2026
08:20
MRI-guided dmPFC-rTMS as a Treatment for Treatment-resistant Major Depressive Disorder
Published on: August 11, 2015
Resting-state fMRI-based machine learning for predicting SSRI treatment response in major depressive disorder
Yuxuan Hu1, Jinrui Gao1, Yaoze Liu1
1Department of Psychiatry, The Affiliated Xuzhou Oriental Hospital of Xuzhou Medical University, Xuzhou, 221004, China.
BMC Psychiatry
|August 5, 2026
Summary
Machine learning models predict Major Depressive Disorder (MDD) treatment response using resting-state fMRI. The models identified key brain regions and clinical factors, showing potential for personalized therapy.
Area of Science:
- Neuroscience
- Psychiatry
- Machine Learning
Background:
- Major Depressive Disorder (MDD) affects many, with debated antidepressant efficacy.
- Low treatment response rates necessitate better predictors for MDD pharmacotherapy.
- Identifying neurological predictors can improve treatment outcomes for MDD.
Purpose of the Study:
- Develop a machine learning model to predict MDD treatment response.
- Utilize resting-state functional magnetic resonance imaging (fMRI) metrics for prediction.
- Enhance personalized pharmacotherapy for MDD patients.
Main Methods:
- 116 MDD patients underwent resting-state fMRI and clinical assessments (HAMD-24).
- LASSO regression selected features (brain regions, symptoms, clinical variables) for model construction.
- Five machine learning models were trained and validated, with the logistic regression model selected for external validation.
Main Results:
- Early functional alterations observed in frontal and sensorimotor regions at 1 week, expanding by 4 weeks.
- The logistic regression model, using features like age, education, core depressive symptoms, and postcentral gyrus activity, achieved an AUC of 0.801 internally.
- External validation showed an AUC of 0.643, highlighting the predictive potential of specific brain regions (e.g., right postcentral gyrus).
Conclusions:
- MDD shows early functional brain changes in emotion, cognition, and sensorimotor areas.
- Integrating neural and clinical data with machine learning offers potential for predicting early SSRI treatment outcomes.
- The study suggests specific brain regions as potential biomarkers for antidepressant efficacy.
