Antidepressants are widely prescribed, but their efficacy relative to placebo is modest, in part because the clinical diagnosis of major depression encompasses biologically heterogeneous conditions. Here, we sought to identify a neurobiological signature of response to antidepressant treatment as compared to placebo. We designed a latent-space machine learning algorithm tailored for resting-state electroencephalography (rsEEG) and applied it to data from the largest imaging-coupled, placebo-controlled antidepressant study (n=309). Symptom improvement was robustly predicted in a manner both specific for the antidepressant sertraline (versus placebo) and generalizable across different study sites and EEG equipment. This sertraline-predictive EEG signature generalized to two depression samples, wherein it reflected general antidepressant medication responsivity, and related differentially to repetitive transcranial magnetic stimulation (rTMS) treatment outcome. Furthermore, we found that the sertraline rsEEG signature indexed prefrontal neural responsivity, as measured by concurrent TMS/EEG. Our findings advance the neurobiological understanding of antidepressant treatment through an EEG-tailored computational model and provide a clinical avenue for personalized treatment of depression.

Repetitive transcranial magnetic stimulation (rTMS) is a commonly-used treatment for major depressive disorder (MDD). However, our understanding of the mechanism by which TMS exerts its antidepressant effect is minimal. Furthermore, we lack brain signals that can be used to predict and track clinical outcome. Such signals would allow for treatment stratification and optimization. Here, we performed a randomized, sham-controlled clinical trial and measured electrophysiological, neuroimaging, and clinical changes before and after rTMS. Patients (N = 36) were randomized to receive either active or sham rTMS to the left dorsolateral prefrontal cortex (dlPFC) for 20 consecutive weekdays. To capture the rTMS-driven changes in connectivity and causal excitability, resting fMRI and TMS/EEG were performed before and after the treatment. Baseline causal connectivity differences between depressed patients and healthy controls were also evaluated with concurrent TMS/fMRI. We found that active, but not sham rTMS elicited (1) an increase in dlPFC global connectivity, (2) induction of negative dlPFC-amygdala connectivity, and (3) local and distributed changes in TMS/EEG potentials. Global connectivity changes predicted clinical outcome, while both global connectivity and TMS/EEG changes tracked clinical outcome. In patients but not healthy participants, we observed a perturbed inhibitory effect of the dlPFC on the amygdala. Taken together, rTMS induced lasting connectivity and excitability changes from the site of stimulation, such that after active treatment, the dlPFC appeared better able to engage in top-down control of the amygdala. These measures of network functioning both predicted and tracked clinical outcome, potentially opening the door to treatment optimization.