One night's total sleep deprivation can induce profound antidepressant effects in depressed patients. Following the absence of sleep for one night, 40-60% of depressed patients typically show the transient onset of mood improvement which is then lost following a nap or the next night's recovery sleep. Thus, the sleep deprivation paradigm becomes a non-pharmacological means of acutely inducing altered mood states in a fashion that may yield important information about underlying mechanisms. While the mood changes are usually transient they may, nonetheless, give insights into neural systems involved in mood dysregulation and its acute reversal. Some patients show stable degrees of improvement following sleep deprivations, while others show tolerance and yet others show a relative refractoriness to sleep deprivation-induced mood improvement early in their bipolar depressive episode, but vulnerability to transient or long-lasting switches out of depression later in the course of their episode. These data provide some of the first systematic observations regarding differential neurobiological responsiveness as a function of the course of a bipolar depressive episode. Preliminary data suggest that increases in TSH secretion may be associated with the degree of clinical response to sleep deprivation in our patient population as a whole, as well as individual case studies. Patients with greater degrees of diurnal variation in mood on the baseline day prior to sleep deprivation show greater antidepressant response compared with those without diurnal variation.