AI-Based Forecasting of Treatment Response in Major Depressive Disorder Using Combined Biomarkers

Abstract

Background: Major Depressive Disorder (MDD) is a leading cause of disability worldwide, characterized by heterogeneous treatment outcomes. Predicting antidepressant response remains a persistent clinical challenge, often resulting in prolonged trial-and-error approaches. Advances in artificial intelligence (AI) and biomarker research now offer opportunities to forecast treatment outcomes more accurately through the integration of multimodal data. Objective: To develop and evaluate an AI model capable of forecasting antidepressant treatment response in patients with MDD using combined clinical, neuroimaging, genetic, and digital biomarker datasets. Methods: A descriptive study was conducted over four months at a tertiary psychiatric center in Lahore, including 68 participants aged 18–55 years diagnosed with MDD according to DSM-5 criteria. Clinical severity was assessed using the Hamilton Depression Rating Scale (HDRS-17) and the Montgomery–Åsberg Depression Rating Scale (MADRS) at baseline and after eight weeks of antidepressant therapy. Neuroimaging, genetic, and digital biomarkers were collected and analyzed. Data integration and model training employed random forest and support vector machine algorithms, with 10-fold cross-validation to ensure reliability. Results: Significant reductions were observed in HDRS-17 and MADRS scores post-treatment (p < 0.001). Responders demonstrated higher cortical thickness, greater gray matter volume, and stronger functional connectivity. The random forest model achieved superior predictive accuracy (86.8%) compared with the support vector machine (83.2%), with an AUC-ROC of 0.91. These findings indicated that integrating multimodal biomarkers improved the precision of treatment response prediction in MDD. Conclusion: The study demonstrated that AI-based multimodal modeling can accurately forecast antidepressant response, supporting the potential of precision psychiatry in optimizing individualized treatment strategies for MDD.