In the last 30 years, it has been convincingly demonstrated that sleep in major depression is characterized by disturbances of sleep continuity, a reduction of slow wave sleep, a disinhibition of REM sleep including a shortening of REM latency (ie the time between sleep onset and the occurrence of the first REM period) and an increase in REM density. Furthermore, manipulations of the sleep-wake cycle like total or partial sleep deprivation or phase advance of the sleep period have been proven to be effective therapeutic strategies for patients with unipolar depression. The database concerning sleep and sleep-wake manipulations in bipolar disorder in comparison is not yet as extensive. Studies investigating sleep in bipolar depression suggest that during the depressed phase sleep shows the same stigmata as in unipolar depression. During the hypomanic or manic phase, sleep is even more curtailed, though subjectively not experienced as disturbing by the patients. REM sleep disinhibition is present as well. An important issue is the question, whether sleep-wake manipulations can also be applied in patients with bipolar depression. Work by others and our own studies indicate that sleep deprivation and a phase advance of the sleep period can be used to treat bipolar patients during the depressed phase. The risk of a switch into hypomania or mania does not seem to be more pronounced than the risk with typical pharmacological antidepressant treatment. For patients with mania, sleep deprivation is not an adequate treatment–in contrast, treatment strategies aiming at stabilizing a regular sleep-wake schedule are indicated.