We sought to examine the effects of age, depression chronicity, and treatment responsiveness on glucose metabolism in a large well-characterized sample of depressed men and a psychiatrically unaffected control group. The subjects were unmedicated, symptomatic, right-handed males (n=66) who met DSM-IV criteria for a major depressive episode in the context of a major depressive disorder (MDD, n=66) and never depressed, right-handed, healthy control subjects (HC, n=24). Subjects in the MDD group were subsequently classified as responders, or non-responders to a six-week trial of paroxetine monotherapy (20-60 mg). Statistical parametric mapping (SPM) was used to analyze the relationship between age and cerebral glucose metabolism (18-fluorodeoxyglucose positron emission tomography) and the modulation by treatment responsivity and a history of prior depressive episodes. Metabolic activity in the rostral and dorsal anterior cingulate cortex showed a significant negative correlation with age in MDD, but not in HC. Non-response to treatment and previous depressive episodes were associated with a higher degree of age-dependent hypometabolism in the rostral and anterior cingulate cortex. The age-dependent changes documented herein may influence the distinct clinical presentation and treatment response described in older-age depression.