ReviewSleep deprivation in depression: what do we know, where do we go?
Introduction
Thirty years ago, the first experimental total sleep deprivation (SD) in a depressed patient with severe insomnia revealed an unexpected and paradoxical improvement the following day (Pflug and Tölle 1971). The remarkable transformation of often deeply depressed, psychotic, suicidal patients in the course of a few hours into their normal premorbid “selves” convinced many psychiatrists at that time of the extraordinary importance of this phenomenon. Many studies followed, resulting in widespread consensus that SD can have antidepressant effects. It also turned out that subsequent sleep tends to reverse this improvement. The original interest and investment in clinical research and application faded away.
In our opinion, two factors may be responsible for the current lack of interest. First, considerable relapses were frequently observed after recovery sleep. Second, the dominance of pharmacology and neurochemistry in research on pathogenesis and therapy of psychiatric disorders may be responsible. It is difficult to obtain funding for non-pharmacological and non-neurochemical clinical research (the same is true for another efficacious antidepressant modality, light treatment). Nevertheless, the rapidity and the magnitude of the clinical changes brought about by SD and sleep still remain highly intriguing and may provide clues for understanding the pathophysiology of depression. In fact, it is surprising that no pharmaceutical company has focused on this model in the search for that much-needed rapid-acting antidepressant; don’t clinicians want a drug that works within a day?
This review will not be an update of previous summaries Wu and Bunney 1990, Kuhs and Tolle 1991, Elsenga 1992, Leibenluft and Wehr 1992. It reiterates and extends the issues brought up in recent detailed reviews Wirz-Justice 1995, Van den Hoofdakker 1997. We wish to reintroduce SD on the clinical and scientific agenda. A renewed attention is justified, not only because of its intrinsic importance, but also because of new developments in the clinical application of SD and in theoretical concepts of the mechanisms underlying the regulation of mood.
Section snippets
For whom does sleep deprivation work?
There is an extraordinarily broad response to SD in depression, irrespective of the syndromal classification. Patients with “endogenous” characteristics (even those with severe psychotic features) respond more often than those showing “neurotic” features (75% vs. 48%, Wu and Bunney 1990; 68% vs. 33%, Elsenga 1992). There is some empirical evidence that bipolar depressed patients respond more often than recurrent unipolars Szuba et al 1991, Barbini et al 1998. Although approximately 25% of
Mechanisms underlying the effects of sleep deprivation
The central characteristic of the mood disorders is by definition the disturbance of one of the most complex and fundamental processes regulating our interaction with the environment: the regulation of mood. SD is a complex intervention with impact on a host of levels of functioning, covering the entire area from social interactions to molecular processes. Thus, the mechanisms explaining SD response can be looked for on many levels.
Let us begin by narrowing these down. Psychological mechanisms
Models of sleep regulation
The two process model of sleep regulation postulates interaction of a homeostatic process S and a circadian process C (Daan et al 1984). Sleep need is represented by process S and is reflected in EEG slow wave activity (SWA), i.e., power density in the frequency range below 4Hz. The level of sleep need depends on the duration of prior wakefulness and sleep: it increases exponentially with increasing duration of wakefulness and decreases during non-rapid-eye-movement (NREM) sleep. Sleep onset
Internal coincidence
The “internal coincidence hypothesis” (Wehr and Wirz-Justice 1981), an extension of the phase advance hypothesis for depression (Wehr et al 1979), assumes that the phase-angle between an advanced circadian pacemaker and the sleep-wake cycle is depressogenic. Depressive patients sleep at the wrong biological clock-time, like shift workers or transmeridian travellers. The second assumption is that SD is effective because the coincidence of sleep with the critical phase is avoided. Recovery sleep
Manipulations of the sleep-wake cycle other than total sleep deprivation: duration or timing?
The models presented essentially postulate two factors that may be responsible for changes in depressive state after SD and recovery sleep: manipulations of sleep duration (and therefore of the level of process S) or manipulation of sleep timing (and therefore the interaction between processes S and C). Which is crucial?
For this question, partial SD has served as a theoretical tool. The antidepressant effects appear equivalent to those of total SD (reviewed in Van den Hoofdakker 1997). Although
Mood regulation
Since mood is such a complex phenomenon, modulated moment to moment by myriad factors within and without (from hormones to social situations), insight into its regulation is difficult. Yet surprisingly, mood, as sleep, follows similar laws of nature. Self-rated mood undergoes a clear circadian rhythm, that can be revealed in a constant routine protocol (e.g., Wirz-Justice 1995). Mood is also dependent on a homeostatic component. The exponential increase of sleep need during wakefulness has a
Dysregulation of mood in depression
As shown in the former paragraph, the instability of mood can be explained by an instability of phase relationships between the need for sleep (process S) and processes controlled by the circadian system. In healthy subjects, some phase relationships are favorable, others unfavorable, and this may also be the case in depressed patients. Sleep deprivation may produce its antidepressant effects by increasing sleep need and thus changing the phase relationship between sleep need and processes
Back to the clinic
Given that in healthy subjects each moment’s mood is physiologically based on the interaction of time of day and elapsed time awake, it is evident that mood can be influenced in many ways. It can vary from day to day because of a good night’s sleep or a bad one, because of a nap, different times of outdoor light exposure, the use of drugs, etc., not to mention psychological events. This complexity is daunting, but the very fact that something so subjective as mood can be reliably measured and
Acknowledgements
We are extremely grateful to Marijke Gordijn for her important and creative contributions, and Domien Beersma and Els Hiddinga for the data in Figure 1.
References (69)
- et al.
The unipolar-bipolar dichotomy and the response to sleep deprivation
Psychiatry Res
(1998) - et al.
Neuroendocrinological investigations during sleep deprivation. I. Concentrations of thyrotropin, thyroid hormones, cortisol, prolactin, luteinizing hormone, follicle-stimulating hormone, estradiol, and testosterone in patients with major depressive disorder at 8am before and after total sleep deprivation
Biol Psychiatry
(1990) - et al.
Prolongation of the antidepressant response to partial sleep deprivation by lithium
Psychiatry Res
(1986) - et al.
Can non-REM sleep be depressogenic?
J Aff Disord
(1992) - et al.
REM sleep deprivation during 5 hours leads to an immediate REM sleep rebound and to suppression of NREM sleep intensity
Electroenceph Clin Neurophysiol
(1990) - et al.
Dopamine agonist amineptine prevents the antidepressant effect of sleep deprivation
Psychiatry Res
(1996) - et al.
Observed behavior as a predictor of the response to sleep deprivation in depressed patients
Psychiatry Res
(1989) - et al.
The relationship between tiredness prior to sleep deprivation and the antidepressant response to sleep deprivation in depression
Biol Psychiatry
(1995) - et al.
Effect of partial sleep deprivation on sleep stages and EEG power spectraevidence for nonREM and REM sleep homeostasis
Electroenceph clin Neurophysiol
(1990) - et al.
Rate of switch from depression into mania after therapeutic sleep deprivation in bipolar depression
Psychiatry Res
(1999)