Advances in neurotechnology have created new opportunities to measure, modulate and understand the neural activity underlying complex behaviors and diseases, including psychiatric disorders such as treatment resistant depression (TRD). Despite promising explorations of deep brain stimulation (DBS) as a therapy for TRD, one of the barriers to scaling this treatment is the lack of objective information supporting clinical decision making during recovery. I will describe our recent efforts collecting and analyzing longitudinal intracranial electrophysiology data from TRD patients undergoing subcallosal cingulate cortex (SCC) DBS, as well as our research using nonlinear dimensionality reduction to identify latent variables that can serve brain biomarkers of disease recovery. We will examine the potential use of this data-driven approach in clinical decision support through patient case studies, as well as highlight what we’re learning about microcircuit structures and complex behaviors that underly TRD pathology and recovery.