Abstract
Small single-site studies found that transcranial magnetic stimulation (TMS) targets with better antidepressant response were more negatively functionally connected to the subgenual cingulate cortex (SGC). These led to "anti-subgenual" TMS targeting in recent clinical trials. We conducted a larger prospective multi-site observational study to test the robustness of this observation in more diverse clinical populations. Sixty-six treatment-seeking individuals with major depressive disorder (MDD) received 3-8 weeks of daily rTMS to the left dorsolateral prefrontal cortex using scalp-based targeting as part of standard clinical care. Stimulation sites were recorded with MRI neuronavigation on multiple days. Our primary outcome was the correlation between change in Beck Depression Inventory (BDI-II) score and connectivity of each individual's TMS site to the SGC, computed using resting-state functional connectivity data from 1000 healthy individuals. Secondary (post hoc) analyses incorporated additional clinical covariates. No relationship was found between antidepressant response and normative connectivity of TMS site to SGC (r = 0.1, p = 0.39). This was not due to inconsistency in the location of the TMS sites, which showed smaller within- than between-individual variance (p < 0.0001). Post hoc analyses showed significant associations when adding clinical covariates (r = -0.27, p = 0.014). Baseline anxiety (p < 0.0001) and comorbid psychiatric conditions (p < 0.001) accounted for the most variance in response. Atlas-based connectivity of TMS site to the SGC accounted for minimal variance in antidepressant response in this diverse sample. The "anti-subgenual" target derived based on normative connectome may be suboptimal for MDD patients with high baseline anxiety or psychiatric comorbidities. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03276793.