Structural imaging predictors of ketamine response in treatment-resistant depression: a machine learning approach.
Linda Bryant, Laith Alexander, Sergio Mena, Yael Jacob, Jenna Jubeir, Mu Li, Philipp T Neukam, Laurel S Morris, James W Murrough, Rebecca Price, Nikolaos Koutsouleris, Mitul A Mehta, Mario Juruena, Fiona Coutts, Paris Alexandros Lalousis
Translational psychiatry May 12, 2026 Peer reviewed DOI: 10.1038/s41398-026-04085-4 via PubMed
Summary
A machine-learning model was developed to predict response to ketamine in adults with treatment-resistant depression (TRD) using pre-treatment MRI data. Among 99 adults treated with a single intravenous dose of ketamine, 52.5% showed clinical response 24 hours later. The model achieved balanced accuracy of 72.2% in the initial group and 60.0% in external validation. Greater gray matter volume in frontal regions predicted positive responses, while greater cerebellar volume indicated non-response.
Study at a glance
| Design | observational cohort |
|---|---|
| Sample size | 99 |
| Population | adults with treatment-resistant depression |
| Key finding | The machine-learning model predicted antidepressant response to ketamine with a balanced accuracy of 72.2% in the discovery sample. |
Abstract
Ketamine has demonstrated rapid antidepressant efficacy in treatment-resistant depression (TRD), but clinical decision-making is challenging due to variability in individual response. Current trial-and-error prescribing practices may expose patients to ineffective treatment and avoidable adverse effects, underscoring the need for reliable predictive tools to optimize treatment selection and support personalized, evidence-based care. We developed a machine-learning model (support vector classifier) to predict antidepressant response to ketamine using pre-treatment structural MRI data. The model was trained on 99 adults with TRD given a single intravenous ketamine infusion (0.5 mg/kg). Clinical response was defined as a ≥50% reduction in MADRS scores 24 h post-infusion. Internal validation used repeated nested cross-validation, and generalizability was tested in two independent ketamine-treated cohorts (n = 51) and a saline-treated control group (n = 49). Among ketamine-treated participants, 52 (52.5%) responded to treatment. The model achieved a balanced accuracy of 72.2% (sensitivity = 72.3%, specificity = 73.1%, AUC = 0.72) in the discovery sample and 60.0% (p = 0.01, AUC = 0.65) in external validation. Greater gray matter volume in frontal regions predicted response, whereas greater cerebellar volume predicted non-response. Performance dropped to chance in the saline cohort (BAC = 41.1%, AUC = 0.45), supporting pharmacologic specificity. These findings present the first machine-learning model for the prediction of ketamine response in TRD using structural neuroimaging and highlight its potential utility for stratified treatment planning and biomarker-informed interventions while providing mechanistic insight into neuroanatomical predictors of antidepressant response.