There is an increasing body of literature fuelled by advances in high-resolution structural MRI acquisition and image processing techniques which implicates subtle neuroanatomical abnormalities in the aetiopathogenesis of bipolar disorder. This account reviews the main findings from structural neuroimaging research into regional brain abnormalities, the impact of genetic liability and mood stabilizing medication on brain structure in bipolar disorder, and the overlapping structural deviations found in the allied disorders of schizophrenia and depression. The manifold challenges extant within neuroimaging research are highlighted with accompanying recommendations for future studies. The most consistent findings include preservation of total cerebral volume with regional grey and white matter structural changes in prefrontal, midline and anterior limbic networks, non-contingent ventriculomegaly and increased rates of white matter hyperintensities, with more pronounced deficits in juveniles suffering from the illness. There is increasing evidence that medication has observable effects on brain structure, whereby lithium status is associated with volumetric increase in the medial temporal lobe and anterior cingulate gyrus. However, research continues to be confounded by the use of highly heterogeneous methodology and clinical populations, in studies employing small scale, low-powered, cross-sectional designs. Future work should investigate larger, clinically homogenous groups of patients and unaffected relatives, combining both categorical and dimensional approaches to illness classification in cross-sectional and longitudinal designs in order to elucidate trait versus state mechanisms, genetic effects and medication/illness progression effects over time.