Abnormal sleep accompanies many psychiatric conditions, but has long been recognized as a particularly conspicuous feature of affective disorders. More than a mere epiphenomenon, the powerful link between sleep and mood regulation is most dramatically demonstrated by the high efficacy of sleep deprivation in alleviating depression. Indeed, the sleep abnormalities that accompany depression may be due to the same neuropathologies that are responsible for its mood and cognitive symptoms. This powerful link between sleep and mood regulation makes polysomnography (PSG) a useful window into the underlying pathophysiology of depression, yet it is underused, particularly in clinical diagnosis. Recent depression research has emphasized the importance of establishing biologically relevant subtypes of depression with treatment specificity and prognostic value. PSG measures, among other biological markers, may be of importance in establishing these subtypes. Two subtypes of depression that appear to have robust biological differences, the melancholic and atypical subtypes, have recently been shown to have different sleep profiles that can aid in differential diagnosis. Further, routine use of PSG in the workup of a depressed patient would minimize the chances of misdiagnosis in those suffering from primary sleep disorders such as sleep apnea, which can present secondary mood symptoms resembling depression. Increased use of PSG in clinical psychiatric practice would enlarge the body of data available for defining new depressive subtypes in the future. It would also serve an immediate purpose in the separation of atypical, compared with melancholic, depression, and the differential diagnosis of depression from primary sleep disorders.