Personalized prediction of transcranial magnetic stimulation clinical response in patients with treatment-refractory depression using neuroimaging biomarkers and machine learning

•Functional anticorrelated connectivity between the subgenual anterior cingulate cortex and left dorsolateral prefrontal cortex cannot serve as biomarker of antidepressant response to repetitive transcranial magnetic stimulation.•Four alternative connectivity biomarkers of treatment response were identified.•Poorer pre-treatment connectivity with(in) the central executive network is associated with poorer antidepressant response to transcranial magnetic stimulation.•Pre-processing with global signal regression possibly removed some valuable neural signal.•Combining the biomarkers revealed outstanding results for personalized prediction of antidepressant response. Functional connectivity between the left dorsolateral prefrontal cortex (DLPFC) and subgenual cingulate (sgACC) may serve as a biomarker for transcranial magnetic stimulation (rTMS) treatment response. The first aim was to establish whether this finding is veridical or artifactually induced by the pre-processing method. Furthermore, alternative biomarkers were identified and the clinical utility for personalized medicine was examined. Resting-state fMRI data were collected in medication-refractory depressed patients (n = 70, 16 males) before undergoing neuronavigated left DLPFC rTMS. Seed-based analyses were performed with and without global signal regression pre-processing to identify biomarkers of short-term and long-term treatment response. Receiver Operating Characteristic curve and supervised machine learning analyses were applied to assess the clinical utility of these biomarkers for the classification of categorical rTMS response. Regardless of the pre-processing method, DLPFC-sgACC connectivity was not associated with treatment outcome. Instead, poorer connectivity between the sgACC and three clusters (peak locations: frontal pole, superior parietal lobule, occipital cortex) and DLPFC-central opercular cortex were observed in long-term nonresponders. The identified connections could serve as acceptable to excellent markers. Combining the features using supervised machine learning reached accuracy rates of 95.35% (CI=82.94–100.00) and 88.89% (CI=63.96–100.00) in the cross-validation and test dataset, respectively. The sample size was moderate, and features for machine learning were based on group differences. Long-term nonresponders showed greater disrupted connectivity in regions involving the central executive network. Our findings may aid the development of personalized medicine for medi