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.