Neuroimaging has contributed profoundly to our understanding of the pathophysiology of psychiatric disorders but has had little impact on treatment. An important goal in neuroscience research is identifying biological markers that predict subsequent response to given treatments. This approach may be especially valuable when considering high-risk and high-cost treatments such as psychiatric neurosurgery. Here, we review neuroimaging findings pertaining to treatment-refractory major depression and obsessive-compulsive disorder, and imaging markers that predict response to neurotherapeutic interventions. Thus far, studies conducted with neurotherapeutic interventions have found patterns of predictive brain activity that are similar to those conducted with pharmacological treatments. The predictive neural correlates are concordant with pathophysiological models of major depressive and obsessive-compulsive disorder. These promising findings should motivate additional work establishing the reliability and cost-effectiveness of neuroimaging to predict treatment response across psychiatric diagnoses and interventions.