Although the role of a genetic factor is established in bipolar disorder, causative genes or robust genetic risk factors have not been identified. Increased incidence of subcortical hyperintensity, altered calcium levels in cells derived from patients and neuroprotective effects of mood stabilizers suggest vulnerability or impaired resilience of neurons in bipolar disorder. Mitochondrial dysfunction or impaired endoplasmic reticulum stress response is suggested to play a role in the neurons' vulnerability. Progressive loss or dysfunction of 'mood-stabilizing neurons' might account for the characteristic course of the illness. The important next step in the neurobiological study of bipolar disorder is identification of the neural systems that are responsible for this disorder.