Accurate and Interpretable Prediction of Antidepressant Treatment Response from Receptor-informed Neuroimaging

Conventional antidepressants show moderate efficacy in treating major depressive disorder. Psychedelic-assisted therapy holds promise, yet individual responses vary, underscoring the need for predictive tools to guide treatment selection. Here, we present graphTRIP (graph-based Treatment Response Interpretability and Prediction) – a geometric deep learning architecture that enables three advances: 1) accurate prediction of post-treatment depression severity using only pretreatment clinical and neuroimaging data; 2) identification of robust, patient-specific biomarkers; and 3) causal analysis of treatment effects and underlying mechanisms. Trained on data from a clinical trial comparing psilocybin and escitalopram (NCT03429075), graphTRIP achieves strong predictive accuracy (r = 0.75, p < 10−8), and generalises both to an independent dataset and across brain atlases. The model links better outcomes to reduced functional coupling within serotonin systems, and broader serotonergic integration with sensory-motor networks. Finally, causal analysis reveals a group-level advantage of psilocybin over escitalopram, but also identifies individuals with specific stress-related neuromodulatory profiles who may benefit more from escitalopram. Overall, this work advances precision medicine and biomarker discovery in depression.