The MR findings reviewed in this article suggest structural, chemical, and functional abnormalities in specific brain regions participating in mood and cognitive regulation, such as the DLPFC, anterior cingulate, amygdala,STG, and corpus callosum in subjects with bipolar disorder. These abnormalities would represent an altered anterior-limbic network disrupting inter- and intrahemispheric communication and underlying the expression of bipolar disorder. Available studies are limited by several confounding variables, such as small and heterogeneous patient samples, differences in clinical and medication status, and cross-sectional design. It is still unclear whether abnormalities in neurodevelopment or neurodegeneration play a major role in the pathophysiology of bipolar disorder. These processes could act together in a unitary model of the disease, with excessive neuronal pruning/apoptosis during childhood and adolescence being responsible for the onset of the disorder and subsequent neurotoxic mechanisms and impaired neuroplasticity and cellular resilience being responsible for further disease progression. Future MR studies should investigate larger samples of first-episode drug-free patients, pediatric patients, subjects at high risk for bipolar disorder, and unaffected family members longitudinally. Such a study population is crucial to examine systematically whether brain changes are present before the appearance of symptoms (eg, maldevelopment) or whether they develop afterwards, as a result of illness course (eg, neurodegeneration). These studies will also be instrumental in minimizing potentially confounding factors commonly found in adult samples, such as the effects of long-term medication, chronicity, and hospitalizations. Juvenile bipolar patients often have a strong family history of bipolar disorder. Future studies could help elucidate the relevance of brain abnormalities as reflections of genetic susceptibility to the disorder. MR studies associated with genetic, post-mortem, and neuropsychologic studies will be valuable in separating state from trait brain abnormalities and in further characterizing the genetic determinants, the neuropathologic underpinnings, and the cognitive disturbances of bipolar disorder.