Objective: Antidepressants with different mechanisms of action might have different effects on brain functions. The aim of the study was therefore to investigate effects of 2 antidepressants on brain activation and to identify predictors for therapy response. Method: Twenty-four untreated patients with major depressive disorder (according to Structured Clinical Interview for DSM-IV) were enrolled in a prospective, randomized, 4-week trial with mirtazapine and venlafaxine. Functional magnetic resonance imaging (fMRI) was performed at baseline and after 4 weeks in the patients and in 15 healthy controls. The primary outcome measure was fMRI blood-oxygen-level dependence (BOLD) activation. The patients were recruited in 2007 and 2008. Results: Comparison between patients and controls revealed that emotional face matching elicited enhanced activation in the anterior cingulate cortex (ACC), dorsomedial prefrontal cortex, dorsolateral prefrontal cortex, and basal ganglia in patients. During treatment, a significant decrease of BOLD responses was seen in the hippocampus, basal ganglia, thalamus, and cerebellum of venlafaxine- treated patients, and a significant increase in BOLD responses was seen in the middle cingulate gyrus and supplementary motor area of mirtazapine-treated patients (P < .05, family wise error [FWE] cluster-level corrected). Larger BOLD responses in the left fusiform gyrus at baseline predicted a better response to venlafaxine, and smaller BOLD responses in the right rolandic operculum at baseline predicted a better response to mirtazapine (P < .05, FWE cluster-level corrected). Conclusions: These fMRI results indicate that antidepressants with different mechanisms of action have different effects on brain function. It therefore seems that fMRI can be used for therapy evaluation and response prediction and can facilitate the development of new pharmaceuticals.