For some years now scientists have known that restricting the sleep of very depressed people acted, as Dr. David Sack puts it, "like a sort of window into wonderland," lifting the depression as if by magic.

For a while, the phenomenon was little more than a curiosity -- of genuine, albeit academic, interest, but of little practical use because sleep slammed the window shut. Once the patient went back to sleep, the depression returned.

Now, however, some researchers at the National Institute of Mental Health are finding that by manipulating sleep cycles, they can apparently bring genuine relief to some depressed patients. This relief appears to be as effective and lasting as that produced by antidepressant drugs -- but with no side effects.

Sack, who is a staff psychiatrist at the Clinical Psychobiology Branch of NIMH, describes one 43-year-old teacher who "was so disabled by her depression she couldn't enjoy anything. She couldn't go out. She was hopeless, had suicidal thoughts, no energy. She felt worse, much worse in the morning. She had a long history of depressions and she'd had just about every treatment known to man -- tricyclic antidepressants, MAO inhibitors, lithium. She had an excellent therapist and had been in psychotherapy for many years, but she just didn't respond."

At the NIMH lab, she was treated only by depriving her of a few hours of sleep each night for three nights. "After three days," says Sack, "her mood completely improved and stayed improved. We followed her for the next six months, and although she did relapse after that time, she improved again in reponse to the partial sleep deprivation.

"In her case we felt we had a treatment that we knew was going to be effective in her depression and a treatment we could readminister when she got depressed again.

"We don't know that everyone is going to respond this quickly -- in fact, we don't think they will. But the point is that there was a dramatic improvment over a short period of time and the improvement was sustained.

"It wasn't just a flash in the pan."

Sack and his colleagues -- Drs. Steven James, Norman Rosenthal, Thomas Wehr and Wallace Mendelson -- are starting a new study, hoping to fine-tune the use of partial sleep deprivation to bring this promising non-drug therapy to millions of people for whom life is a perennial downer, a gloom broken in some cases only by episodes of joyless manic highs.

The work by the NIMH team is based on a series of studies involving the relationship between sleep and mood, as well as on the ever-growing knowledge about how the brain works and influences the rest of the body.

Scientists really do not know exactly why humans sleep. "One should say," Sack says, "that it is not entirely understood what the goal or purpose of sleep is, that a major question in the whole sleep research field is what is the function of sleep.

"This intuitive notion that sleep is what makes you feel energetic in the morning is probably not true. I think it is important to know that even if you didn't go to sleep, you would feel tired when you went to bed and you would feel less tired when you got up in the morning. The cycle in fatigue and wakefulness is not dependent on actually attaining sleep.

"Sleepiness and sleep are two different processes."

But whatever goes on during sleep, the sleep disturbances appear to be an integral part of a number of psychiatric disorders, especially depressions, whether they are unipolar (depression alone) or bipolar (depression alternating in cycles with manic highs). Indeed, Sack hypothesizes that it is just as correct to say that "depression is a disorder of sleep and activity as it is to say that depression is a disorder of mood," as it is usually characterized. Depressed patients often sleep intermittently or too much, waken in the night unable to fall back to sleep or are unable to get to sleep.

One study published in the late 1970s hypothesized that it was the period of sleep called REM, or rapid eye movement, that is disordered in depressed persons. Dr. Gerald W. Vogel at Emery University hooked up a group of depressed patients to equipment in a sleep laboratory and woke them every time they went into REM sleep -- up to 30 times a night. The depressed patients gradually improved, but again, this was a therapy that was impractical even in a hospital setting, requiring, for instance, one technician per patient throughout the night.

The scientists are pursuing research leads suggesting that sleep-wake cycles are linked to the regulation of certain hormones and neurotransmitters (brain chemicals) which can affect moods. Also there is evidence that sleep may help synchronize the body's various biological rhythms, those that regulate body temperature and others that govern the production or supression of certain hormones.

"Under normal conditions," says Sack, "these are internally coordinated, kind of hooked up to one another, and the sleep-wake cycle probably gives information back to keep things on schedule."

The sleep disorders prevalent in depressions may interfere with all of this delicate balancing, or, some have suggested, the insomnias that accompany depression may be the body's natural attempt to regulate the out-of-kilter biological clocks.

In any case, says Sack, "we have already been able to confirm that partial sleep deprivation (PSD) does in fact work, that it has an acute antidepressant effect, and by repeating it we believe that it has a cumulative antidepressant effect." Some studies also suggest that a program of medication with partial sleep deprivation may shorten the period it normally takes an antidepressant drug to work, as long as several weeks in some cases.

The NIMH scientists are particularly excited about the prospects for PSD. "We have, I believe, built the groundwork to the point for a treatment which could be clinically applied," says Sack. "It would be the first clinically accessible nonpharmacological treatment for severe depression. Replication would not involve a sophisticated sleep lab or a high tech environment. It is quite likely that in the course of the next couple of years it could be just assimilated into the clinical culture. It will make it much more possible, for instance, for psychologists and non-medically trained clinicians to treat depression, too, because if, after all, this treatment works faster than drugs and works as well, there's not going to be a whole lot of reason to treat people with drugs alone."

Current studies will try to identify those patients most likely to be helped -- probably, the scientists believe, those with histories of cyclic, severe depressions which have responded poorly to current therapies, or those who cannot tolerate or do not wish to take antidepressant medications.

Volunteers for the new PSD studies will spend some time as inpatients at the NIMH unit. Three shifts of nurses will help keep them awake during the prescribed hours -- offering videotapes, support, even walks up and down the corridors, says Sack. (But no coffee, no stimulants.) It is not yet a do-it-yourself project, says Sack, because "it is virtually impossible to stay awake on your own when you want to sleep." More Information

For information call Dr. David A. Sack or Judith Klein, Clinical Psychobiology Branch, National Institute of Mental Health, Bulding 10, Room 4S-239, National Institutes of Health, Rockville, Md. 20205. Phone: 496-2141.