OBJECTIVES: Repetitive transcranial magnetic stimulation (rTMS) is a brain stimulation technique which has received increasing attention as an antidepressant treatment. However available studies are characterized by a substantial variability in response. We hypothesized that individual patients' characteristics might contribute to such heterogeneity. Therefore we investigated whether either alterations of regional cerebral blood flow (rCBF) or clinical characteristics may predict antidepressant response to rTMS. DESIGN: 24 patients with major depression and stable medication received high frequency (10 Hz) rTMS over the left dorsolateral prefrontal cortex (DLPFC) for two weeks as add-on treatment. ECD-Single photon emission computed tomographay (SPECT) imaging was performed 1 to 2 days before rTMS. SETTING: Tertial referral center RESULTS: After two weeks of rTMS a mean reduction of 30% of the initial Hamilton Depression Rating Score (HAMD) was observed. Using a multivariate regression model with simultaneous evaluation of the relative impact of a-priori chosen potential factors influencing treatment outcome, two variables, the pretreatment anterior cingulate rCBF and the former response to antidepressant agents proved significant. High pretreatment anterior cingulate activity and low treatment resistance to pharmacologic therapy were positive predictors for treatment response to rTMS. CONCLUSIONS: Pretreatment anterior cingulate activity seems to be a useful prognostic marker of rTMS treatment response, which is in line with other treatment strategies, like sleep deprivation, electroconvulsive therapy or antidepressant medication.