Mental disorders arise from brain circuit dysfunctions, but most of our treatments target the whole brain rather than defined circuits. Deep brain stimulation (DBS) is a more circuit-directed approach that has done well in movement disorders but has very mixed results in randomized clinical trials for mental illness. Part of the difficulty is that psychiatric DBS is delivered in a trial-and-error fashion, without clear evidence that it engages the target circuits. I will discuss new strategies for developing biomarkers to guide that target engagement, centered around the idea of understanding how brain stimulation changes cognition and decision-making. We have identified ways in which DBS can augment top-down executive function and have linked those changes to cortical electrophysiology. In animals, we have developed new approaches to understand how those changes occur and how we can leverage them for clinical benefit. Taken together, these offer the prospect of a new generation of rationally designed brain stimulation therapies.