Multiple lines of evidence, such as impaired energy metabolism in the brain detected by magnetic resonance spectroscopy, a possible role of maternal inheritance, co-morbidity with mitochondrial diseases, the effects of mood stabilisers on mitochondria, increased mitochondrial DNA (mtDNA) deletion in the brain, and association with mtDNA mutations/polymorphisms or nuclear-encoded mitochondrial genes, suggest that mitochondrial dysfunction is an important component of bipolar disorder. Global reduction of mitochondria-related gene expression in the postmortem brains of patients with bipolar disorder may also be an indicator, but such findings are affected by sample pH and thus need to be interpreted with caution. A recently developed animal model carrying mtDNA deletion in neurons suggested that accumulation of mtDNA deletions causes bipolar disorder-like phenotypes. The next step in the study of mitochondrial dysfunction in bipolar disorder should be clarification of how mitochondrial dysfunction, a nonspecific risk factor, can cause specific symptoms of bipolar disorder. Two hypothetical mechanisms are mtDNA neuroplasticity and nonvisual photoreception impairment. Further study of mitochondrial dysfunction in bipolar disorder is expected to be useful for the development of new mood stabilisers.