• Improvement on the Hamilton Depression Scale (6-Item Version) from Baseline to Follow-up [ Time Frame: Baseline, 24 h (after sleep deprivation night), 48 h (after the first recovery night) ] [ Designated as safety issue: No ]
  

• Number of responders (50% reduction on the HAMD-6) in the modafinil group versus the placebo group [ Time Frame: Baseline, 24 h (after sleep deprivation night), 48 h (after the first recovery night) ] [ Designated as safety issue: No ]
  
• self-rating scale of global mental state (Befindlichkeitsskala (Bf-s)) and the Stanford Sleepiness Scale. [ Time Frame: Baseline 24 h (after sleep deprivation night), 48 h (after the first recovery night) ] [ Designated as safety issue: No ]
  
• Polysomnography (among others assessing sleep onset latency, sleep efficiency, sleep states, wakefulness after sleep onset) [ Time Frame: Baseline 24 h (after sleep deprivation night), 48 h (after the first recovery night) ] [ Designated as safety issue: No ]
  
• Neuropsychological battery (including Zahlenverbindungstest, California Verbal Learning Test, d2 Attention Stress Test, Farb-Wort-Interferenztest, CS fine motor task [ Time Frame: Baseline 24 h (after sleep deprivation night), 48 h (after the first recovery night) ] [ Designated as safety issue: No ]
  

In about 60 % of depressed patients, sleep deprivation has an acute therapeutic effect. The effect appears within very few hours; this is different from all other antidepressant treatments. However, in most cases patients relapse after the subsequent night's sleep (Wu & Bunney 1990).

The antidepressant mechanism of action of sleep deprivation is not known as yet; several hypotheses are presented and discussed in various review articles (Wiegand 1995; Kasper & Möller 1996; Wirz-Justice & van den Hoofdakker 1999; Gillin et al. 2001; Ringel & Szuba 2001; Giedke & Schwärzler 2002). At present, the following general hypotheses are most discussed:

1. Sleep deprivation exerts its effect by activating or intensifying an antidepressant "pro-cess". This "process" may be dopaminergic or serotonergic transmission, thyroid function etc.
2. Sleep deprivation inactivates a hypothetic "depressiogenic" substance which is produced during sleep.
3. Sleep deprivation reduces the central nervous cholinergic transmitter activity and restores the balance between cholinergic and aminergic transmitter systems.
4. Sleep deprivation acts by preventing sleep during a "critical" or "vulnerable" phase of circadian rhythms; there are various chronobiological assumptions which can specify such a "critical phase".

Several studies tried to indentify predictors of response to sleep deprivation in order to clarify the mechanism of action. Among clinical predictors is a symptom pattern with "endogenous" or "melancholic" traits and the presence of pronounced diurnal variations of mood, and a behaviour pattern pointing to an elevated level of arousal or activity. Another predictor is a pronounced sleep disturbance during the baseline night. Among the many neuroendocrine and neurohumoral factors which have been studied, only elevated thyroid hormones turned out to be a predictor for response to sleep deprivation. PET and SPECT studies have convergently demonstrated an elevated metabolism in parts of the limbic system (e.g., the anterior cingulum) at baseline in responders. These findings do not yet allow conclusions with respect to the neurotransmitter systems involved.

Wiegand et al. (1993) investigated whether scheduled daytime naps can induce relapses after successful sleep deprivation therapy. The timing of the nap turned out to be a crucial factor; nap sleep duration and sleep structure during naps were less important.

The majority of studies in this field suffers from a methodological problem: there is no objective continuous polysomnographic measurement of sleep. The continuous absence of sleep during the sleep deprivation period is thus not documented. It is known from sleep deprivation studies in healthy probands that during prolonged sleep deprivation, short sleep episodes ("microsleep") occur frequently. Hemmeter et al. (1998) were the first to demonstrate that also in depressed patients undergoing sleep deprivation, microsleep occurs and tends to prevent the antidepressant effect. Data from a recently finished study of our group point into the same direction (partly published in Wiegand et al. 2002).

To further elucidate this question, an experimental procedure appears useful where the occurrence of sleep episodes during the sleep deprivation period is suppressed as far as possible by the vigilance enhancing drug modafinil.

The study aims to investigate whether the administration of modafinil or placebo during a 40 hour sleep deprivation period in depressed patients can intensify the antidepressant effect of the sleep deprivation.

This study is a basic science study that aims to provide information on the therapeutic mechanism of sleep deprivation in depression and on the reoccurence of depressive symptoms in case of intermittent short sleep episodes.

Primary Hypothesis:

There is a significant reduction on the HAMD-6 scale between baseline and 24 h later (at the mornings before and after one night of sleep deprivation)

Secondary Hypotheses:

1. The amount of "responders" (50% of reduction on the HAMD-6 scale) is significantly greater in the modafinil than in the placebo group.
2. The primary hypothesis and the first secondary hypothesis are also assessed by a self-rating scale of global mental state (Befindlichkeitsskala (Bf-s)) and by the Stanford Sleepiness Scale.
3. The overall amount of sleep assessed by polysomnography is smaller in the modafinil group as compared to the placebo group during the 40 h sleep deprivation period.
4. The group differences in HAMD-6 ratings are parralleled by differences in the overall amount of sleep during the 40 h sleep deprivation period.
5. There are group differences (Modafinil versus Placebo) on a comprehensive neuropsychological battery taken at baseline and 24 h later (at the mornings before and after one night of sleep deprivation)