Psychiatric Annals

Sleep Studies in Affective Illness: Diagnostic, Therapeutic, and Pathophysiological Implications

J Christian Gillin, MD

Abstract

The diagnosis of affective illness is rarely made in the absence of altered sleep duration. Since the time of Hippocrates, clinicians have described insomnia as a major and common symptom of depression, although hypersomnia has been observed in a small subgroup of depressed patients.

Inspired by the discovery of Rapid Eye Movement (REM) sleep by Aserinsky and Kleitman1 in 1953, investigators have studied the sleep disturbance of affective illness intensively for the past 20 years orso.2-26 The major sleep abnormalities reported in primary or endogenous depression are summarized in Table 1 and illustrated in Figure 1, which shows the all-night polygraph! cal Iy recorded sleep patterns of a normal subject and an unmedicated patient with severe psychotic depression. As expected from clinical observations, most moderately to severely depressed patients do show reduced total sleep time, increased awake time before, during, and after sleep (ie, increased sleep latency, intermittent wakefulness, and early morning awake time, respectively), and reduced sleep efficiency (the proportion or percent time spent asleep while in bed). Polygraphie recordings, however, make new observations possible; namely, that sleep is shallow (Stages 3 and 4, or Delta Sleep, which correspond roughly to deep sleep, are meager), that the amount of ocular activity during REM sleep (often referred to as REM Density) is increased, and that REM sleep may be redistributed toward the onset of sleep. Kupfer and his associates3'4'16"20 have emphasized that mean REM Latency is short in primary depression, a finding which has been found by several other groups. 2'5-9'15'22"26 When data from individual nights are displayed as a histogram, there are two peaks to REM Latency in depressed patients, with one peak at about 10 minutes and the other at about 50 to 60 minutes.26 Moreover, there are reports that the length of the first REM period is increased and that REM sleep is proportionately more abundant in the first half of the night and less common in the second half of the night in depression compared to what has been reported in normal subjects.18

Vogel has pointed toward a constellation of three REM sleep abnormalities in endogenous depression: a) short REM Latency; b) high REM Density; and c) the abnormal temporal distribution of REM sleep during the night.27 The abnormal distribution of REM sleep toward the beginning of sleep was highlighted by several nights of REM sleep deprivation. Both before, and especially after, REM deprivation, patients with endogenous depression tended to have more REM sleep ("early REM%") early in the night and less REM sleep ("late REM%") later in the night compared with insomniac controls.

The hyperarousal seen in the nocturnal sleep of depressives is also present during the day when patients are requested to nap. Most patients appear to have difficulty sleeping compared with what has previously been reported in normals.28 Thus, even though many patients with depression complain of fatigue and insomnia, they sleep poorly when given the opportunity to nap or sleep at night.

Age-related changes in normal sleep include a gradual reduction in amount of Delta Sleep and REM Latency and increased intermittent wakefulness. The sleep of depressed patients also reflects these age-related changes, especially in measures of sleep continuity.9'29 More than normal subjects, depressed patients show increasing Early Morning Awake time and declining sleep efficiency and total sleep time in each successive age group. Rush et al found empirically in a preliminary study that endogenous depression could be diagnosed with good sensitivity, specificity, and confidence with the following rules: a) if age less than 45 years, diagnosis is depression if REM latency less than 62 minutes; and b) if age greater than…

The diagnosis of affective illness is rarely made in the absence of altered sleep duration. Since the time of Hippocrates, clinicians have described insomnia as a major and common symptom of depression, although hypersomnia has been observed in a small subgroup of depressed patients.

Inspired by the discovery of Rapid Eye Movement (REM) sleep by Aserinsky and Kleitman1 in 1953, investigators have studied the sleep disturbance of affective illness intensively for the past 20 years orso.2-26 The major sleep abnormalities reported in primary or endogenous depression are summarized in Table 1 and illustrated in Figure 1, which shows the all-night polygraph! cal Iy recorded sleep patterns of a normal subject and an unmedicated patient with severe psychotic depression. As expected from clinical observations, most moderately to severely depressed patients do show reduced total sleep time, increased awake time before, during, and after sleep (ie, increased sleep latency, intermittent wakefulness, and early morning awake time, respectively), and reduced sleep efficiency (the proportion or percent time spent asleep while in bed). Polygraphie recordings, however, make new observations possible; namely, that sleep is shallow (Stages 3 and 4, or Delta Sleep, which correspond roughly to deep sleep, are meager), that the amount of ocular activity during REM sleep (often referred to as REM Density) is increased, and that REM sleep may be redistributed toward the onset of sleep. Kupfer and his associates3'4'16"20 have emphasized that mean REM Latency is short in primary depression, a finding which has been found by several other groups. 2'5-9'15'22"26 When data from individual nights are displayed as a histogram, there are two peaks to REM Latency in depressed patients, with one peak at about 10 minutes and the other at about 50 to 60 minutes.26 Moreover, there are reports that the length of the first REM period is increased and that REM sleep is proportionately more abundant in the first half of the night and less common in the second half of the night in depression compared to what has been reported in normal subjects.18

Vogel has pointed toward a constellation of three REM sleep abnormalities in endogenous depression: a) short REM Latency; b) high REM Density; and c) the abnormal temporal distribution of REM sleep during the night.27 The abnormal distribution of REM sleep toward the beginning of sleep was highlighted by several nights of REM sleep deprivation. Both before, and especially after, REM deprivation, patients with endogenous depression tended to have more REM sleep ("early REM%") early in the night and less REM sleep ("late REM%") later in the night compared with insomniac controls.

The hyperarousal seen in the nocturnal sleep of depressives is also present during the day when patients are requested to nap. Most patients appear to have difficulty sleeping compared with what has previously been reported in normals.28 Thus, even though many patients with depression complain of fatigue and insomnia, they sleep poorly when given the opportunity to nap or sleep at night.

Age-related changes in normal sleep include a gradual reduction in amount of Delta Sleep and REM Latency and increased intermittent wakefulness. The sleep of depressed patients also reflects these age-related changes, especially in measures of sleep continuity.9'29 More than normal subjects, depressed patients show increasing Early Morning Awake time and declining sleep efficiency and total sleep time in each successive age group. Rush et al found empirically in a preliminary study that endogenous depression could be diagnosed with good sensitivity, specificity, and confidence with the following rules: a) if age less than 45 years, diagnosis is depression if REM latency less than 62 minutes; and b) if age greater than 45 years, diagnosis is depression if Stage 4 time is less than 20 minutes.

Figure 1. All-night EEG sleep recording demonstrating short, shallow, fragmented sleep as well as short REM Latency and redistribution of REM sleep toward onset of sleep in a severely psychotic unipolar patient and normal control.

Figure 1. All-night EEG sleep recording demonstrating short, shallow, fragmented sleep as well as short REM Latency and redistribution of REM sleep toward onset of sleep in a severely psychotic unipolar patient and normal control.

Table

TABLE 1MAJOR SLEEP DISTURBANCES IN ENDOGENOUS OR PRIMARY DEPRESSION

TABLE 1

MAJOR SLEEP DISTURBANCES IN ENDOGENOUS OR PRIMARY DEPRESSION

Nonsuppression on the dexamethasone suppression test (DST) is a commonly reported biological marker for moderate to severe endogenous depression. Both Feinberg et al6 and Rush et al24 have found short REM Latency in nearly all patients who fail to show suppression on the DST, but many low REM Latency patients showed DST suppression. REM Latency was, therefore, more sensitive than DST "escape" for the diagnosis of endogenous depression. Both groups have also found that the combination of DST and sleep measures separated endogenous from nonendogenous patients with high degrees of sensitivity and specificity.

The extent to which sleep disturbances are statedependent as compared to trait-dependent is unresolved. As Hauri et al1 have shown, unipolar patients in remission had normal REM Latencies, but delayed, shallow, fragmented and more variable sleep from night to night compared with controls. More recently, several investigators have presented preliminary data indicating other persistent abnormalities of sleep in remission, such as occasional nights with short REM Latency (Kupfer, personal communication), increased Stage 1 sleep (Knowles, personal communication), and increased REM Density in bipolar patients.25'30

In general, the magnitude of the sleep disturbances in depressed patients has varied inversely with clinical severity.11' 8,J2 Some of the most dramatic disturbances have been reported in psychotic depressives. Although there is a widespread clinical impression that bipolar depressed patients sleep longer than unipolar depressed patients, EEG sleep studies have not shown the difference in total sleep between these groups of patients to be either robust or even statistically significant. In our own comparison we found only one measure to be statistically different, REM Efficiency (a measure of intrusions of awake and nonREM states into REM sleep).5 It was significantly lower in both unipolar and bipolar depressed patients than normals, but it was still significantly lower in bipolar than unipolar, depressed patients.

MULTIVARIANT DISCRIMINANT ANALYSIS

Since several sleep measurements appear to be abnormal in primary depression, attempts have been made to separate different clinical groups (ie, primary vs secondary depressives or depressives vs normals) by statistical techniques which simultaneously use several measures rather than one (multivariant as compared with univariant techniques). Coble et al,3 for example, found that the combination of two variables, REM Latency and REM Density, discriminated patients with primary depression compared to those with secondary depression. These findings were replicated by Feinberg et al.6 In addition, Kupfer's group found that discriminant analysis, based upon the duration of the first REM period and sleep efficiency, separated patients with and without delusions of guilt and sin.18

In an effort to obtain some data on the specificity of the sleep abnormalities in primary depression, we compared unmedicated depressed patients, normal controls, and patients with primary insomnia (persistent complaints of insomnia five nights a week for two years or more in the absence of obvious psychiatric or medical illness or substance abuse).8 Compared with normal controls, depressed patients, as expected from previous studies, showed significantly less total sleep, longer sleep efficiency, more early morning awake time, more intermittent awake time, less delta (Stage 3 and 4) sleep, less sleep efficiency, and shorter REM Latency. Compared with insomniac patients, the depressed patients showed significantly greater early morning awake time, shorter REM Latency, and greater REM Density. When the sleep data from these three groups were subjected to multivariate discriminant analysis, 83 of the 1 15 subjects were correctly "diagnosed," 100% of the 41 normals, 69.9% of the 56 depressed, and 88.9% of the 18 insomniacs. In order to validate this analysis, we studied a second group of depressed patients with the discriminant functions generated by the first phase of the study and successfully classified 15 of 18 patients (83.3%). (The values for successful classification reported here are slightly different from those in the original report, since we subsequently learned that the original program inadvertently eliminated all decimal places from the entered data.*)

Taking our data from the NIMH, Feinberg et al at the University of Michigan have used it to create discriminant functions which successfully identified patients with primary or endogenous depression.6 They found a discriminant function using REM Latency and four measures of sleep continuity (awake, total recording period, early morning awake time, and sleep efficiency) from the NIMH patients which separated unipolar endogenous and nonendogenous patients in Ann Arbor. Bipolar endogenous patients were less well discriminated from nonendogenous patients. These findings, using these techniques, support the hypothesis that patients with moderate to severe depression in the different parts of the world share a common set of sleep disturbances. These observations suggest, therefore, that the sleep disturbances of depression could reflect some common underlying pathophysiological process which is directly or indirectly associated with etiological factors of depression. Further evidence supporting such intimate relationships between sleep and affective illness come from studies suggesting clinical benefits which result from deliberate manipulations of sleep.

IS SLEEP "BAD" FOR DEPRESSION?

Five major types of sleep alteration have been reported to improve patients with primary or endogenous depression: total sleep deprivation, partial sleep deprivation (either first or second half of night), selective REM sleep deprivation, and phase advance of the sleep period (Table 2).

Since the initial reports of Schulte33 and, later, Pflug and Tolle,34 over 25 studies have been published suggesting that about one-third to two-thirds of endogenously depressed patients benefit from a night of total sleep deprivation,3 The benefits are occasionally dramatic and enduring but, more frequently, are transient and mild. Endogenous patients with diurnal mood swings appear to respond better that those with neurotic depression. When clinical improvement does occur, it is often at 4 to 7 AM (ie, after only about four to eight hours of sleep deprivation). In a single case study, Knowles et al reported that a 15-minute nap reinduced depression following clinical improvement induced by sleep deprivation.36 Moreover, we found that responders to sleep deprivation showed improvement in their sleep (ie, increased total sleep time, reduced sleep latency) whereas nonresponders did not show any significant sleep change during recovery as compared with baseline before sleep deprivation.37'38 Although responders and nonresponders did not differ from each other in any univariate measures prior to sleep deprivation, responders were more likely to be classified as "depressed" by our previously described multiple discriminant analysis.8 Following sleep deprivation, half the responders who had been classified as depressed prior to sleep deprivation were classified as normal. Thus, the clinical changes appeared to be associated with sleep changes in the appropriate direction, and clinical benefits were most likely in patients with more severe depressive-like sleep patterns and more severe clinical ratings. In this regard, it is interesting that Nasrallah et al have reported that the antidepressant effects of sleep deprivation are more common in patients with dexamethasone nonsuppression39 and that sleepdeprivation-induced clinical improvement reverses dexamethasone non-suppression.40

In addition to total sleep deprivation, partial sleep deprivation of either the first or second half of the night has been reported to have antidepressant effects. 2 While the two forms of partial sleep deprivation have not formally been compared, sleep deprivation in the second half is said to be more efficacious, better tolerated and equally as good as total sleep deprivation.42

The clinical usefulness of sleep deprivation therapy has been limited by the fact that many patients relapse following a night of recovery sleep. Recently, however, van Bemmel and van den Hoofdakker found that partial sleep deprivation on the recovery nights prevented relapse and augmented improvements induced by a night of total sleep deprivation.43

In another design, Wehr et al have reported that a persistent six-hour phase advance of the sleep period (ie, moving sleep onset from 1 1 PM to 5 PM and wake up from 7 AM to 1 AM) had antidepressant effects. Although the number of patients studied with this technique is still small, it appears that the clinical benefits appear less rapidly but last longer than with total sleep deprivation, provided, of course, that the patients remain on the new sleep-wake schedule.

Finally, it appears as if selective REM sleep deprivation has antidepressant effects in endogenous depression. Vogel, the primary architect of this hypothesis, has suggested two types of evidence: I) antidepressant drugs work by producing sustained, significant reduction of REM sleep; and 2) deprivation of REM sleep by the awakening technique has antidepressant effects, whereas similar awakenings from nonREM sleep had no antidepressant effects?7'45 In support of the first argument, a review of the literature shows that nearly all of the recognized classical antidepressant medications (tricyclics, monoamine oxidase inhibition) reduce REM sleep time with chronic administration and, conversely, non-antidepressant drugs (alcohol, barbiturates, etc.) do not.35 This is not to say that all antidepressant therapies suppress REM sleep. Electroshock therapy, for example, has not been shown to do so consistently in depressed patients.45'46 One possible pharmacological exception to the above generalization could be para-chlorophenylalanine (PCPA), an inhibitor of serotonin synthesis, which Wyatt et al showed to reduce REM sleep by about 50% over several weeks of treatment.47 Not only is PCPA not antidepressant, but Shopsin et al 48 have shown that it reverses the antidepressant effects of tranylcypromine and Carpenter49 suggested that it might induce depression in patients with carcinoid syndrome.

In support of the second argument, Vogel et al conducted a heroic experiment in which 34 endogenously depressed patients were manually REM deprived nearly every night.43 Approximately half the patients were discharged from the hospital with REM deprivation for 6 to 1 2 weeks as the only form of therapy. Patients who failed to respond to REM deprivation did not respond to Imipramine (250 mg/ day for four weeks). Thus, the results suggest that REM deprivation, whether by awakenings or by pharmacological means, has antidepressant effects in about half the patients.

Table

TABLE 2EXPERIMENTAL MANIPULATIONS OF SLEEP-WAKE CYCLE REPORTED TO HAVE ANTIDEPRESSANT EFFECTS

TABLE 2

EXPERIMENTAL MANIPULATIONS OF SLEEP-WAKE CYCLE REPORTED TO HAVE ANTIDEPRESSANT EFFECTS

It may seem paradoxical that various deliberate disturbances of sleep improve rather than exacerbate the clinical state of patients with primary or endogenous depression. After all, conventional wisdom has held that troubled sleep causes troubled minds, and, therefore, tranquil sleep should help troubled minds. These conventional notions should not be easily dismissed, but they may be wrong in emphasizing total sleep time as the key variable. A rationale for the phase advance therapy is that it brings the sleep period into proper alignment ("phase") with an underlying circadian process which has been phase advanced or moved forward relative to clock time, perhaps reflected by the circadian temperature or REM system. Vogel et al have also suggested this as a possible mechanism by which REM deprivation has antidepressant effects.2 They showed that patients who improved with REM deprivation increased "late REM%" more rapidly during the early weeks of treatment than patients who did not respond. Likewise, it has been hypothesized that there is a critical period during the 24-hour day (perhaps the hours at the end of the normal sleep period) when sleep may have a depressogenic effect.60 AU of these hypotheses await critical evaluation.

HOW SPECIFIC ARE THE SLEEP DISTURBANCES OF DEPRESSION?

From the preceding discussion it is evident that the sleep disturbances are relatively sensitive markers of clinical depression; that is, most patients with moderate or severe endogenous or primary depression will have short REM Latency, reduced total sleep time and delta sleep, etc. The next question is, how specific are these changes; that is, do patients with other diagnoses have these same sleep changes? In answering this question, it is evident that there are other psychiatric conditions in which the entire constellation of these sleep abnormalities are not found. For example, most psychotic schizophrenic patients and insomniacs do not show short REM Latency,22 although both groups may show low amounts of Delta sleep. 8'22,3S Over the years, however, there has been increasing evidence that short REM Latency may be present in conditions other than primary and endogenous depression, some normal and some pathological. For example, short REM Latency was reported in normals who nap during the morning, who are in time-free environments ("free-running"), who have previously been deprived of REM sleep, and who are elderly.35 Likewise short REM Latency occurs in narcoleptics, and, more recently, has been described in some patients with anorexia nervosa,51 acute schizophrenia,52 chronic schizophrenia in the elderly,53 psychotic schizoaffective schizophrenia,54 and atypical or secondary types of depression.2 In some of these instances, however, issues of diagnosis or concomitant major affective illness may have obscured the interpretation of the results. More recently, however, we have found that both adolescents55 and adults56 with severe obsessive-compulsive disorder have sleep disturbances which are virtually identical to those seen in depressed patients. Not only is REM Latency short and, in the adults, Delta Sleep reduced, but total sleep time and sleep efficiency is low and sleep latency is increased, In the case of the adults with obsessivecompulsive disorder, no major differences in sleep existed between those with and without a history of or concurrent major depressive symptoms. Moreover, many of the adult patients with obsessive-compulsive disorder "escaped" on the DST, j ust as do patients with endogenous depression.57

While these results with obsessive-compulsive patients are preliminary and require replication, they do suggest that these sleep disturbances may be sensitive but not necessarily specific markers of endogenous and primary depression as currently defined. In neither of these studies then did first REM periods or REM Density increase in the patients compared with normals.

METHODOLOGICAL ISSUES

These studies should serve to remind us that more data are still needed comparing unmedicated well-diagnosed depressed patients with age-matched controls of various sorts, including normals, differing types of depression, other psychiatric or medical illnesses. There are surprisingly few studies comparing depressed patients and normals, and very few in which more detailed measures have been made, for example, of REM period duration or REM density. Moreover, the definitions of commonly used sleep measures are not in agreement. In a survey of the definition of REM Latency in different laboratories, Knowles et al found different answers in each laboratory.58 Different investigators defined sleep onset variously, and either included or excluded intermittent wakefulness between the onset of sleep and REM. If the most well known sleep measure is not defined, what can we say about other measures?

It is not surprising that duration of the first REM period in both normals and patients should vary from study to study. For example, in normals it was reported to be 15 minutes by Hartmann12 and 21 minutes by Duncan et al.15 In primary and endogenous depressives, it varied from 16 minutes to 27 minutes in a series of studies.5'12'18'27'54'59 Measurement of the duration and REM density of any given REM period is complicated by the increased intrusions of wakefulness and nonREM sleep into REM periods;5 investigators do not always report whether or not they exclude these intrusions when calculating REM duration.

Studies of REM Density have been hampered by the absence of an agreed-upon method of scoring. Kupfer and Gillin, who were trained by Frederick Snyder, have used the same scale based upon a score of 0 to 8 of ocular activity per minute of REM sleep. Other methods a re also in use. Some progress with these problems may be expected with new technological methods, such as computer-assisted measurements of eye movements during REM sleep.20

HOW TO INTERPRET THE EEG SLEEP CHANGES IN DEPRESSION?

What clues to underlying pathophysiology do these objective EEG sleep changes provide? Given the methodological problems reviewed and gaps in needed information, it would be premature to reach any firm conclusions, but it is not too early to think theoretically, if only to provide direction in unexplored areas (Table 3).

Table

TABLE 3SOME INTERPRETATIONS OF SLEEP DISTURBANCE IN DEPRESSION

TABLE 3

SOME INTERPRETATIONS OF SLEEP DISTURBANCE IN DEPRESSION

There have been at least five major theoretical models suggested. First, Frederick Snyder32 proposed the REM Deprivation Hypothesis; that is, that many of the sleep phenomena observed during the deepest or waning phases of depression (ie, short REM Latency, enhanced REM Density, etc.) could have resulted from prior deprivation of REM sleep in the early (waxing) phases of depression when insomnia developed for nonspecific reasons. No longitudinal data exist during the early waxing phases of depression to evaluate this hypothesis. Longitudinal studies later on in the course of the illness are not necessarily consistent with it.4'12'13'21'" Hartmann considered the hypothesis only to reject it, and to suggest instead that depressed patients actually exhibit "increased REM pressure."12

Secondly, there have been a variety of chronobiological hypotheses, such as the phase-advance hypothesis (alluded to earlier) which postulates that the shift of REM sleep to the beginning of the night in depression reflects a phase advance of an oscillator driving the propensity for REM sleep with respect to the sleep-wake (day-night, light-dark) cycle.44'50'61'62 In addition, there are other hypotheses suggesting that patients with affective illness are "free running" or suffer from internal desynchronization. " Preliminary studies, which have examined the circadian organization of REM sleep and temperature in depressed patients have not found convincing evidence consistent with these hypotheses.28'67

Thirdly, and in some ways similar to the chronobiological models, is Vogel's hypothesis that the sleep of the depressed patients resembles that of normal subjects who are sleeping on an extended sleep regime. In essence, the sleep of depressed patients resembles that of normals who are nearing sleep satiety. This hypothesis could be consistent with suggestions that depressed patients are overaroused.35 Vogel also proposed a theoretical model suggesting that neurophysiological mechanisms inhibiting REM sleep could be defective in endogenously depressed patients.

More recently, Borbely and Wirz-Justice68 have interpreted the sleep of depressed patients in light of Borbely's69 two-process model for normal sleep regulation. This model proposes that sleep is regulated by two processes, a circadian factor ("C") which reflects sleep propensity around the clock, and a homeostatic process ("S") which reflects the increasing propensity to sleep with increasing wakefulness and which may be mediated by a postulated sleep factor. Borbely and Wirz-Justice suggest that Process S accumulates slowly or is deficient in some manner in depression.

Finally, there have been suggestions, based on neurophysiological70'71 and neurochemical evidence,72'73 that the short REM Latency, enhanced REM Density, reduced total sleep time, and other features may result from enhanced cholinergic or decreased aminergic functional activity, or a combination of both. Studies in animals suggest that REM sleep may be induced by pharmacological agents which mimic acetylcholine or block noradrenergic neurons;70-73 furthermore, some studies of single cell activity suggest that neurons which are presumptively cholinergic increase their firing rate in association with REM sleep, especially rapid ocular activity, and conversely, some noradrenergic and serotonergic neurons decrease their firing during REM sleep.70'71 In man, REM sleep can be induced by intravenous infusions during non REM sleep of physostigmine, an anticholinesterase, or arecoline, a direct muscarinic agonist. Scopolamine, a muscarinic antagonist, blocks the effects of arecoline and delays REM sleep.74

In an effort to model the sleep disturbances of depression, we tried to create a state of cholinergic supersensitivity in normal volunteers by administration of scopolamine for three consecutive mornings. The subjects showed, by the third day, short REM Latency, enhanced REM Density, decreased total sleep time and sleep efficiency, and overall sleep patterns identified by our multivariant discriminant analysis as predominantly "depressed."75

THE CHOLINERGIC REM INDUCTION TEST

In order to test for cholinergic supersensitivity in patients with affective disorder, we developed the Cholinergic REM Induction Test (CRIT). An intravenous dose of placebo or arecoline is administered during nonREM sleep 25 minutes following the end of the first REM period to a subject pretreated with a peripherally active anticholinergic agent. The time from the infusion until the second REM period is measured. The results indicate that patients with major affective disorder, both in remission and when ill, enter REM sleep considerably faster following arecoline than do agematched normal controls; patients and controls did not differ following saline.30'74' 6 The results were the same in patients in remission as those who were ill, indicating that it was independent of clinical state. Furthermore, they were the same in patients who had been taken off pharmacological treatment two weeks previously as those who had never received somatic therapy or who had discontinued treatment four months or more previously. Preliminary analysis of the data from a twin study indicates that the response to the cholinergic REM Induction Test is strongly influenced by genetic factors.77

To our knowledge, no attempts have been made in other laboratories to replicate these results from the cholinergic REM Induction Test. Recently, however, Berger et al78 reported that depressed patients were more likely to awaken to an intravenous infusion of physostigmine (0.5 mg five minutes after sleep onset) than normal controls. Since we had previously found that physostigmine could induce either wakefulness or REM sleep, depending upon the dose and the timing during nonREM sleep, 4 these results are further evidence of enhanced responsivity to cholinergic agonists in depressed patients.

The mechanisms by which arecoline induces REM sleep more rapidly in depressed patients than controls is unknown. It could reflect muscarinic receptor supersensitivity in some crucial area thought to be involved in sleep regulation (ie, locus coeruleus, raphe nuclei, or fastigial tegmental gigantocellular neurons). It could also, for example, reflect a functional deficiency in an aminergic system. In support of this hypothesis, we found a significant negative relationship between response on the Cholinergic REM Induction Test and behavioral activation produced by amphetamine.30

The response on the Cholinergic REM Induction Test correlated not with REM Latency but with REM Density of the first REM period in both normals and remitted bipolar patients.30 Animal studies tentatively suggest that REM Density could be directly correlated with cholinergic mechanisms70 and inversely with serotonergic mechanisms.79

The rapid Cholinergic REM Induction seen in depressed patients appears to be consistent with recent neuroendocrine findings of Risch et al.80 They found that patients with affective illness showed more pronounced elevation of serum prolactin, ACTH, and beta-endorphin than psychiatric patients without affective illness in response to physostigmine.

In summary, these results suggest that patients with major affective disorders may have an inherited supersensitivity to cholinergic, muscarinic stimulation, which is independent of clinical state. This hypothesis is consistent with the earlier hypothesis of Janowsky et al82 who proposed that depression results from an increased ratio of cholinergic to noradrenergic functional activity, but extends it by suggesting a state independent, perhaps predisposing, role for cholinergic influences, and a statedependent role for aminergic influences. Further work is needed to test these hypotheses.

SUMMARY

Modern EEG sleep studies show both transient and abiding abnormalities of sleep in patients with major affective disorders. They provide hope for future progress not only in diagnosis but prognosis, not only pathophysiology but therapy.

ACKNOWLEDGMENT

The author expresses his deep appreciation to William E. Bunney, Jr., M. D., Richard J. Wyatt, M. D. and Robert A. Cohen, M. D., Ph.D., who supported the Unit on Sleep Studies, NIMH; Natraj Sitaram, M. D., who was involved in all aspects of the cholinergic studies; John Nürnberger, M. D. and Elliott Gershon, M. D., who made major contributions to the Cholinergic REM Induction Studies; to his colleagues Robert Post, M. D., Dennis Murphy, M. D., Robert M. Cohen, M. D., Wallace Mendelson, M. D., and Thomas lnsel, M. D.; and to Wallace Duncan, Angela Moore, Carol Vanskiver, Debra Garnett, Julie Blendy and Fredric Storch for devoted technical assistance which made these studies possible. Joan Harris expertly edited and prepared the manuscript.

REFERENCES

1. Aserinsky E, Kleitman N: Regularly occurring periods of ocular motility and concomitant phenomena during sleep. Science 1953; 1 18:273-274.

2. Akiskal AH. Rosenthal TL. Haykal RF, et al: Characterological depressions. Arch Gen Psychiatry 1980; 37:777-783.

3. Coble P, Foster FG, Kupfer DJ: Electroencephalograph ic sleep diagnosis of primary depression. Arch Gen Psychiatry I976; 33:1124-1127.

4. Coble PA. Kupfer DJ. Spiker DG. et al: EEG sleep in primary depression: A longitudinal placebo study. J Affective Disord 1979; 1131-138.

5. Duncan WB. Pettigrew KA. GiIUnJC: REM architecture changes in bipolar and unipolar patients. Am J Psychiatry I979; 136:1424-1427.

6. Feinberg M. Gillin JC. Carroll BJ. et al: EEG studies of sleep in the diagnosis of depression. Biol Psychiatry 1982;17:305-316.

7. Gillin JC. Mazure C Post RM. et al: An EEG sleep study of a bipolar (manic-depressive) patient with a nocturnal switch process. Biol Psychiatry 1977;12:711-718.

8. Gillin JC. Duncan W. Pettigrew KD, et al: Successful separation of depressed, normal and insomniac patients by EEG sleep data. Arch Gen Psychiatry 1979; 36:85-90.

9. Gillin JC Duncan WC, Murphy DL. cl al: Age related changes in sleep in depressed and normal subjects. Psychiatry Res 1981; 4:73-78.

10. Gresham SC. Agnew WF. Williams RL: The sleep of depressed patients. Arch Gen Psychiatry 1965; 13:503-507.

11. Hartmann E. Verdone P, Snyder F: Longitudinal studies of sleep and dreaming patterns in psychiatric patients. J Nerv Ment Dis 1966; 142:177-126.

12. Hartmann E: Longitudinal studies of sleep and dream patterns in depressive patients. Arch Gen Psychiatry 1968; 19:312-329.

13. Hauri P, Chernik D. Hawkins D, et al: Sleep of depressed patients remission. Arch Gen Psychiatry 1974; 31:386-391.

14. Hauri P. Hawkins D: Phasic REM. depression and the relationship sleeping and waking. Arch Gen Psychiatry 1971; 25:56-63.

15. Hawkins DR, Mendels J: Sleep disturbance in depressive syndromes. Am J Psychiatry 1966; 123:682-290.

16. Kupfer DJ: REM Latency: A psychobiological marker for primary depressive disease. Biol Psychiatry 1972; 11, 159-174.

17. Kupfer DJ , Foster FG. Coble P, et al: The application of EEG sleep in the differential diagnosis of affective disorders. Am J Psychiatry 1978; 135:169-74.

18. Kupfer DJ. Broudy D, Coble PA. et al: EEG sleep and affective psychosis. J Affective Disord 1980; 2:17-25.

19. Kupfer DJ. Spiker DG. Coble PA. et al: Sleep and treatment prediction endogenous depression. Am J Psychiatry 1981; 138:429-434.

20. Kupfer DJ. Shaw DH. Ulrich R. et al: Application of automated analysis in depression. Arch Gen Psychiatry 1982; 39:569-573.

21. Mendels J, Hawkins DR: Sleep and depression: IV Longitudinal studies. J Nerv Ment Dis 1971; 153:251-272.

22. Mendelson WB. Gillin JC. Wyatt RJ: Human Sleepandits Disorders. New York, Plenum Publishing Corp 1977.

23. Post RM, Stoddard FJ. Gillin JC, et al: Slow and rapid alterations in motor activity, sleep, and biochemistry in a cycling manic-depressive patient. Arch Gen Psychiatry 1970; 34:470477.

24. Rush AJ. Giles GE, Roffwarg HP. et al: Sleep EEG and dexamethasone suppression test findings in unipolar major depressive disorders. Biol Psychiatry 1982: 17:327-341.

25. Schulz H. Trojan B: A comparison of eye movement density in subjects and in depressed patients before and after remission. Sleep Research 1979; 8:49.

26. Schulz H. Lund R, Cording C. et al: Bimodal distribution of REM sleep latencies in depression. Biol Psychiatry 1979; 14:595-600.

27. Vogel GW, Vogel F. McAbee RS, et al: Improvement of depression by REM sleep deprivation. Arch Gen Psychiatry 1980; 37:247-253.

28. Kupfer DJ. Gillin JC. Coble PA. et al: REM sleep, naps, and depression. Psychiatry Res 1981; 5:195-203.

29. Ulrich RF, Shaw DH, Kupfer DJ: Effects of aging on EEG sleep in depression. Sleep 1980; 3:31-40.

30. Sitaram N. Nürnberger JL Gcrshon ES. Gillin JC: Cholinergic regulation of mood and REM sleep: Potential model and marker of vulnerability to affective disorder. AmJ Psychiatry 1982; 139:571-576.

31. Spiker DG, Coble P. Cofsky J. et al: EEG sleep and severity of depression. Biol Psychiatry 1978; 13:485-488.

32. Snyder F: Dynamic aspects of sleep disturbance in relation to mental illness. Biol Psychiatry 1969; 1:119-130.

33. Schulte W: Zum Problem der Provokation und Kupierung von Melancholischen Phasen. Schweizer Arch Neurol Neurochir Psychiatr 1971; 109:427-435.

34. Pflug B, Tolle R: Disturbance of 24 hour rhythm, endogenous depression, and the treatment of endogenous depression by sleep deprivation. In Pharmacopsychiatry 1971; 6:187-196.

35. Gillin JC. Sitaram N, Wehr TA. et al: Slcepand affective illness, in Ballinger J, Post RM (eds): Biology of Major Affective Disorders. Baltimore. Williams and Wilkins Co, to be published.

36. Knowles JB, Southmayd SE. Delva N, et al: Five variations of sleep deprivation in a depressed woman. Br J Psychiatry 1979; 135:403-410.

37. Duncan WC, Gillin JC. Post RM, et al: Relationship between EEG sleep patterns and clinical improvement in depressed patients treated with sleep deprivation. Biol Psychiatry 1980: 15:879-889.

38. Gemer RH. Post RM, Gillin JC. et al: Biologiealand behavioral effects of one night's sleep deprivation in depressed patients and normals. J Psychiatr Res 1979; 15:21-40.

39. Nasrallah H. Coryell WH: Dexamethasone nonsuppression predicts the antidepressant effects of sleep deprivation. Psychiatry Res 1982; 6:61-64.

40. Nasrallah HA. Kuperman S. Coryell W: Reversal of dexamethasone nonsuppression with sleep deprivation in primary depression. Am J Psychiatry 1980; 137:1413-1414.

41. Schilgen B, Tolle R: Partial sleep deprivation as therapy for depression. Arch Gen Psychiatry 1980: 37:267-277.

42. Goetze V. foelle: Antidepressant effects of partial sleep deprivation during first half of the night. Psychiatr Clin 1981: 14:129-149.

43. Van Bcmmel AL. van den Hoofdakker RH: Maintenance of therapeutic effects of total sleep deprivation by limitation of subsequent sleep. A pilot study. Acta Psychiatr Scand 1981; 63:453-462.

44. Wehr TA. Wirz-Justice A, Goodwin FK, et al: Phase advance of the circadian sleep-wake cycle as an antidepressant. Science 1979; 206:210-213.

45. Vogel GW. Thurmond A. Gibbons P, et al: REM sleep reduction effects on depression syndromes. Arch Gen Psychiatry 1975; 32:765-777.

46. Green WJ, Stajduhar PP: The effect of ECT on the sleep dream cycle in psychotic depression. J Nerv Ment Disord 1966; 143:123-134.

47. Wyatt RJ. Sjoerdsma A, Snyder F: Effects of parachlorphenylatanine on sleep in man. Elect roencephabgr Clin Neurophysiol 1969; 27:529-532.

48. Shopsin B. Friedman E. Gershon S: Parachlorophenylalanine reversed tranylcypromine effects in depressed patients. Arch Gen Psychiatry 1976; 33:811-819.

49. Carpenter WT: Serotonin in affective disorders. Ann Intern Med 1970; 73:613-619.

50. Papousek M: Chronobiologische Aspekeic der Zyclothymie. Fortschr Neurol Psychiatr 1975; 43:381-440.

51. NeilJF. MerikangesJR, Foster FG. et al: Waking and all-night sleep EEG's in anorexia nervosa. Clin Electroencephalogr 1980; 11:9-15.

52. Stern M, Fram D. Wyatt RJ, et al: All night sleep studies of acute schizophrenics. Arch Gen Psychiatry 1969; 20:470-477.

53. Jus K, Bouchard M, Jus AK. et al: Sleep EEG studies in untreated, longterm schizophrenic patients. Arch Gen Psychiatry 1973: 29:386-390.

54. Kupfer DJ, Broudy D. Spiker DG. et al: EEG slecpand affective psychoses: I. Schizoaffective disorders. Psychiatry Res 1979; 1:173-178.

55. Rapoport J. Elkins R, Langer DH. et al: Childhood obsessive compulsive disorders. Am J Psychiatry 1981; 138:1545-1555.

56. Insel TR. Gillin JC. Moore AM. et al: The sleep of obsessive-compulsive disorder patients. Arch Gen Psychiatry. 1982; 39:1371-1377.

57. Insel TR. Kalin NH, Guttmacher L B. et al: The dexamethasone suppression test in patients with primary obsessive-compulsive disorder. Psychiatry Res 1982: 6:153-160.

58. Knowles .IB. McI can AW. Cains J: Del'initmns of REM latency. Some comparisons with particular reference to depression. Biological Psichiatri' 1982: 17:993-1002.

59. Foster FG. Kupfer DJ, Coble PA. et al: Rapid eye movement sleep density: An objective indicator in severe medical-depressive syndromes. Arch Gen Psychiatry 1976; 33:1119-1123.

60. Wehr TA, Wirz-Justice A: Internal coincidence model of sleep deprivation and depression. Presented at the European Sleep Meeting, Amsterdam, September 1980.

61. Weitzman ED. Kripke D, Goldmacher D, et al: Acute reversal of the sleep waking cycle in man: Effects on sleep stage patterns. Arch Neurol 1970; 22:483-489.

62. Kripke DF. Atkinson M: Circadian phase disorders in manic depressives. New Research Abstracts, Presented at the Annual Meeting of the American Psychiatric Association Meeting. Detroit. Michigan. May 1974.

63. Halberg F: Physiological considerations underlying rhythmometry, with special reference to emotional illness, in Symposium Bel-Air III. Geneva, Massan et Cie, 1968, pp 73-1236.

64. Rockwell DA, Winget CM. Rosenblatt LS. et al: Biological aspects of suicide: Circadian disorganization. J Ner\' Ment Dis 1978; 166:851-858.

65. Wehr T. Goodwin FK: Desynchronization of circadian rhythms as a possible source of manic-depression cycles. Psychopharmacol Bull 1980; 16:19-20.

66. Kripke DF, Mullaney DJ, Atkinson M, et al: Circadian rhythm disorders in manic-depressives. Biol Psychiatry 1978; 13:335-344.

67. Avery DH. Wildschiodtz G. Rafaelsen OJ: Nocturnal temperature in affective disorders. J Affect Dis 1982: 4:61-71.

68. Borbdy AA, Wirz-Justice A: A two-process model of sleep regulation: II. Implications for depression. Human Neurobiology, 1982; 1:205-210.

69. Borbely AA: A two-process model of sleep regulation: 1. Physiological Basis and Outline. Human Neurobiology. 1982; 1:195-204.

70. Hobson JA. McCarley RW, McKénna TM: Cellular evidence bearing on the pontine brain-stem hypothesis of desynchronized sleep control. Prog Neurabiol 1976; 6:280-376.

71. McCarley RW: REM sleep and depression: Common neurobioiogical control mechanisms. Am J Psychiatry 1982; 139:565-570.

72. Karczmar AC, Longo VG, Scott de Carolis: A pharmacological model of paradoxical sleep: The role of cholinergic and monoamine systems. Physiol Behav 1970;5:175-182.

73. Gillin JC. Mendelson WB, Sitaram N, et al: The neuropharmacology of sleep and wakefulness. Ann Rev Pharmacol Toxicol 1978; 18:563-569.

74. Sitaram N, Gillin JC: Development and use of pharmacological probes of the CNS in man: Evidence of cholinergic abnormality in primary affective illness. Biol Psychiatry 1980; 15:925-955.

75. Gillin JC. Sitaram N, Duncan WC: Muscarinic supersensitivity: A possible model for the sleep disturbances of primary depression? Psychiatric Res 1979; 1:17-22.

76. Siiaram N. Nürnberger JI. Gershon ES, et a): Fastercholinergic REM sleep induction in euthymic patients with primary affective illness. Science 1980; 208:200-202.

77. Nurberger Jl, Siiaram N, Gershon ES. et al: A twin study of cholinergic REM induction. Biological Psychiatry, to be published.

78. Berger M, Land R, vonZerssen D: Der Einfluss eines chloninergikumus auf den REM Schlaf gesunder Probanden und depressiver Patienten, in Beckman H (ed): Fortschritte psychiatrische Forschung. Bern. Huber. 1982, pp 323-327.

79. Siegel JM, Wheeler RL. McGinty DJ: Activity of medullary reticular formation neurons in the unrestrained cat during waking and steep. Brain Res 1979; 179:49-60.

80. Risen SC. Kalin NH, Janowsky DJ: Cholinergic challenges in affective illness: Behavioral and neuroendocrine correlates. J Clin Psvchopharmacol 1981; 1:186-192.

81. Janowsky DL, El-Yousef MK, Davis JM: A cholinergic adrenergic hypothesis of mania and depression, lancet 1972; 2:632-635.

TABLE 1

MAJOR SLEEP DISTURBANCES IN ENDOGENOUS OR PRIMARY DEPRESSION

TABLE 2

EXPERIMENTAL MANIPULATIONS OF SLEEP-WAKE CYCLE REPORTED TO HAVE ANTIDEPRESSANT EFFECTS

TABLE 3

SOME INTERPRETATIONS OF SLEEP DISTURBANCE IN DEPRESSION

10.3928/0048-5713-19830501-02

Sign up to receive

Journal E-contents