Administration of placebo can result in a clinical response indistinguishable from that seen with active antidepressant treatment. Functional brain correlates of this phenomenon have not been fully characterized. Changes in brain glucose metabolism were measured by using positron emission tomography in hospitalized men with unipolar depression who were administered placebo as part of an inpatient imaging study of fluoxetine. Common and unique response effects to administration of placebo or fluoxetine were assessed after a 6-week, double-blind trial. Placebo response was associated with regional metabolic increases involving the prefrontal, anterior cingulate, premotor, parietal, posterior insula, and posterior cingulate and metabolic decreases involving the subgenual cingulate, parahippocampus, and thalamus. Regions of change overlapped those seen in responders administered active fluoxetine. Fluoxetine response, however, was associated with additional subcortical and limbic changes in the brainstem, striatum, anterior insula, and hippocampus, sources of efferent input to the response-specific regions identified with both agents. The common pattern of cortical glucose metabolism increases and limbic-paralimbic metabolism decreases in placebo and fluoxetine responders suggests that facilitation of these changes may be necessary for depression remission, regardless of treatment modality. Clinical improvement in the group receiving placebo as part of an inpatient study is consistent with the well-recognized effect that altering the therapeutic environment may significantly contribute to reducing clinical symptoms. The additional subcortical and limbic metabolism decreases seen uniquely in fluoxetine responders may convey additional advantage in maintaining long-term clinical response and in relapse prevention.