BACKGROUND: Preliminary research findings have shown that mindfulness-based cognitive therapy improves anxiety and depressive symptoms in bipolar disorder. In this study, we further investigated the effects of MBCT in bipolar disorder, in a controlled fMRI study. METHOD: Twenty three patients with bipolar disorder underwent neuropsychological testing and functional MRI. Sixteen of these patients were tested before and after an eight-week MBCT intervention, and seven were wait listed for training and tested at the same intervals. The results were compared with 10 healthy controls. RESULTS: Prior to MBCT, bipolar patients reported significantly higher levels of anxiety and symptoms of stress, scored significantly lower on a test of working memory, and showed significant BOLD signal decrease in the medial PFC during a mindfulness task, compared to healthy controls. Following MBCT, there were significant improvements in the bipolar treatment group, in measures of mindfulness, anxiety and emotion regulation, and in tests of working memory, spatial memory and verbal fluency compared to the bipolar wait list group. BOLD signal increases were noted in the medial PFC and posterior parietal lobe, in a repeat mindfulness task. A region of interest analysis revealed strong correlation between signal changes in medial PFC and increases in mindfulness. LIMITATIONS: The small control group is a limitation in the study. CONCLUSION: These data suggest that MBCT improves mindfulness and emotion regulation and reduces anxiety in bipolar disorder, corresponding to increased activations in the medial PFC, a region associated with cognitive flexibility and previously proposed as a key area of pathophysiology in the disorder.