Total sleep deprivation for one night induces temporary remissions in sixty percent of patients with major depression; it can also induce mania in bipolar patients. Furthermore, as little as two hours of sleep can trigger depression in patients who have improved after sleep deprivation. Thus sleep appears to have a depressant effect and wakefulness a mood-elevating effect. These observations have practical implications for the management of affective illness. For example, some patients' depressions can be treated with sleep deprivation, and sleep disruption is sometimes an identifiable and preventable cause of mania. The purpose of this project is to attempt to identify biological mechanisms of the antidepressant effects of sleep deprivation. Identification of biological mechanisms would increase understanding of the pathogenesis of depression and mania and could be expected to lead to new, rapidly acting drug treatments for depression and mania. The current project is designed to test a hypothesis that thermoregulatory mechanisms underlie the mood-altering effects of sleep and sleep deprivation. The hypothesis is based on our observation that many of the physiological responses to sleep resemble responses to heat exposure. Like heat exposure, sleep onset stimulates secretion of sweat, prolactin (PRL) and growth hormone (GH), and it inhibits metabolic heat production and secretion of TSH and triiodothyronine (T3). Like cold exposure, sleep deprivation has opposite effects. We are testing the hypothesis that the heat-like property of sleep and the cold-like property of sleep deprivation are responsible for their clinical and neuroendocrine effects in depressed patients. According to this hypothesis a warm environment, compared with a cool environment, should blunt the antidepressant and neuroendocrine effects of sleep deprivation. Patents and normal individuals are sleep-deprived on one occasion in an ambient temperature of 90 degrees F, and on another occasion in an ambient temperature of 62 degrees F. In twelve patients studied so far, the results are consistent with the predictions based on our hypothesis, but some additional subjects are needed to permit definitive statistical evaluation of the insults.
