go to content


Vadim S. Rotenberg, E-mail: vadir@post.tau.ac.il, L. KAYUMOV., P. INDURSKY, J. HADJEZ, R. KIMHI, P. SIROTA, A BICHUCHER

Journal of Psychosomatic Research, 1997, 42:565-576.


Twenty-seven depressed patients and 10 healthy subjects were investigated in the sleep laboratory during two to three consecutive nights. Eleven of the 27 patients demonstrated the first night effect" (group I) and 11 other patients demonstrated a clear absence of the 'first night effect' (group II). Five of the 27 depressed patients were omitted from the study because they did not fit criteria for first night effect. The 10 healthy controls demonstrated a first night effect. In group I the duration of the first rapid eye movement (REM) sleep episode was increased on the first night and on the second night the REM sleep latency was decreased, whereas REM sleep duration and eye movement (EM) density was increased. The number of the short sleep cycles (less than 40 minutes) was greater in group I versus group II and the percentage of slow-wave sleep (SWS) was also higher in group I. In depressed patients with the first night effect" the enhanced REM sleep requirement is satisfied not only by an increased REM sleep duration but also by the improved REM sleep quality that is crucial for adaptation. The adaptive role of the increased first REM period and the increased EM density in this period is very limited.

Keywords:"First night effect", EM density, REM sleep, Slow-wave sleep.


Sleep structure in depression, in comparison to healthy individuals, is characterized by reduced slow-wave sleep (SWS), shortened REM sleep latency, redistribution of REM sleep throughout the night (REM sleep in the first cycle is relatively increased), and increased eye movement density that is also predominantly more evident in the first cycle [1-3].

Recently, a meta-analysis has shown that the polysomnographic differences between affective disorders and other diagnostic categories (such as schizophrenia, insomnia, and narcolepsy) are less obvious [4] and sleep variables do not reliably distinguish affective disorders from other illnesses. Nevertheless, there is no doubt that the combination of the aforementioned altered sleep variables, although not specific for depression, discriminate mental pathology from mental health. At the same time, the functional significance of the redistribution of REM sleep and of the increased REM density is not clear. There is contradiction in the literature as to the

relationship between REM sleep variables and clinical state in depressed patients. REM sleep variables have been shown to display a normalization during remission [5,6], but this finding has not been confirmed in other investigations[7-9]. Thus, the alteration in REM sleep variables may be a trait or state marker of psychopathology [2], or may play a definite role in its pathogenesis. The latter is supported by the use of REM deprivation as an effective treatment for depression [10-11].

Alternatively, REM sleep redistribution may also reflect an attempt to compensate for depression [12]. This proposition ties in with theories concerning the adaptive function of REM sleep in healthy subjects. One of these theories [13] states that REM sleep is responsible for adaptation of recent information, especially that information which is not relevant to the previous experience and, in the process of such adaptation, the familiar defense mechanisms are used to adapt to new, stressful events. Another theory [14-16] suggests that REM sleep is necessary for the restoration of search activity and for the compensation of the renunciation of search in the previous wakefulness, because renunciation of search is extremely deteriorative for mental and somatic health. Evidence of the deteriorative effect of REM sleep deprivation [17] is in agreement with this concept.

Increased REM sleep pressure could either reflect the brain's attempt to compensate for the depressed state (a clear example of renunciation of search), or could simply be a marker of this state. An investigation of patients under natural stressful conditions may help shed further light on this issue. The first night effect is a good example of such natural stress. The first night effect is an alteration of sleep structure on the first night in a sleep laboratory, in comparison to the second night. In healthy subjects, REM sleep latency is increased on the first night of investigation and REM sleep percentage is slightly decreased [18,19]. Sleep latency and the number of sleep stage shifts toward superficial sleep or wakefulness are also increased, whereas sleep efficiency index is decreased [18,19]. All the aforementioned changes probably reflect the increased state of vigilance caused by the unfamiliar sleeping environment. Although most healthy subjects demonstrate first night effect, it was found only in 50% of in-patients suffering from major depression [20, 21]. Depressed patients who do not show first night effect often demonstrate a paradoxical evolution of REM sleep with an increase in REM sleep latency from the first to the second night. It has also been shown that depressed patients who display first night effect have a better clinical outcome [21]. The purpose of this study was to compare sleep structure and the interrelationship between objective and subjective sleep variables on two consecutive nights in depressed patients with and without first night effect.

Our assumption was that the peculiarity of REM sleep alteration in depressed patients with and without first night effect may elucidate the functional meaning of REM sleep in depression. We have formulated following hypotheses: (1) sleep structure of patients with the first night effect is more similar to sleep structure of healthy subjects than to sleep structure of patients without first night effect; and (2) REM sleep distribution and its correlations with other sleep variables are different in both groups of patients.


Twenty-seven patients diagnosed with depression (25 with major depression and 2 in the depressive phase of lifetime bipolar mood disorder) were investigated clinically and in the sleep laboratory. The group consisted of 11 men and 16 women with a mean age of 49.87,3 years. The clinical evaluation of depression was performed by psychiatrists using DSM-III-R. The 21-item Hamilton Rating Scale for depression [22] was applied a day before the first polysomnography to score the severity of the present depressive episode. All patients were physically healthy and there was no nonpsychiatric medical condition responsible for their depression. A clinical interview was used to exclude other diagnoses associated with altered sleep (current alcoholism, apnea. myoclonus). Patients were free from psychotropic medication for at least 10 days before the investigation, and they had previously not been treated with neuroleptic agents or other medications with long half-lives. A control group was comprised of six men and four women (mean age 40.5 10 years) without any sleep complaints or any medical problems. On clinical interview, they denied having any relatives with mood disturbances. Sleep data were collected from all subjects on two consecutive nights. During the sleep study, the lights were turned off between 10:00 and 10:30 PM and turned on between 6:00 and 6:30 AM. For the control group, this sleep/wake routine was almost normal and corresponded to their bedtime and awakening schedule during the work week. Even when patients and healthy subjects awakened spontaneously they were not allowed to get out of bed before 6:00 AM to rule out a possible additional sleep period in early morning hours. During the day, patients were supervised by the ward staff to prevent napping.

Frontal electroencephalogram (EEG), submental electromyography (EMG), and electro-oculogram (EOG) for EM detection were monitored using a Neurofax EEG-4400 electroencephalograph. EEG and EOG were recorded at a low filter setting of 0.3 Hz.

Polysomnograms were analyzed according to the international criteria [23]. The investigator who analyzed polysomnograms was blinded with respect to subjects. Sleep onset was defined as the first 30-second epoch of stage I, stage II, SWS, or REM sleep that appeared after wakefulness, and was followed by no more than 1 minute of wakefulness in the first 10 minutes of sleep. The definition of sleep onset at the appearance of stage I is relatively rare, however, this criterion has been used previously [4]. This criterion was used because the first night effect can influence the duration of stage I in the first cycle and, as a result, REM sleep latency can be changed dramatically. REM latency was defined as the time from the onset of sleep until the appearance of the first REM episode. REM percentage was defined as the percentage of total sleep time spent in REM sleep. Eye movements (EM) in REM sleep were counted visually. The criterion for EM detection was a minimum 25-mV deflection of the tracing above baseline [24]. EM density was calculated as EM frequency for every 1 minute of REM stage. Fragments of wake-fulness or non-REM sleep occurring within a single REM episode were not analyzed for EM activity and were not included in the REM episode length when a net EM/REM episode variability was computed. To be considered as a REM episode, the REM sleep duration had to exceed 1 minute. The REM episode was defined as singular if there were no more than 10 minutes of intervening wakefulness or non-REM sleep incorporated into the REM episode. These criteria were applied in patients and controls to estimate the fragmentation of REM episodes and to compare such fragmentation in both groups.

All patients were divided into two groups based on the presence or absence of the first night effect. According to our criteria, first night effect was present if REM sleep latency on the first night was at least 30 minutes longer than REM sleep latency on the second night. REM sleep latency was arbitrarily used as a main criterion of first night effect, as supported by the literature [20,21,25], whereas the decrease of sleep efficiency and SWS reduction may or may not be present [25]. The first night effect was defined as absent if REM sleep latency was of similar duration or was less than 10 minutes longer on the second night.

On the morning after the final awakening in the laboratory, each subject had to complete a 13-item questionnaire that assessed their subjective feelings about sleep duration, sleep latency, number of awakenings, sleep depth, and the state after sleep (mood, refreshment, etc.; see Appendix).

Results in the text are expressed as meanSD. Statistical significance of the results was determined using the paired Student's f-test. Intercorrelations between objective sleep variables and correlations between subjective reports and sleep variables were also determined.


Five of the depressed patients who had REM sleep latency of more than 10 minutes on the first night, but less than 30 minutes longer on the second night, were excluded from the analysis because they did not fit our criteria for the first night effect.

Based on the results of their polysomnogram 11 of the remaining patients were assigned to group I (first night effect). The other 11 patients were assigned to group II (no first night effect). Each group contained six women and five men. The mean age of group I was 49.47.3 years, and of group II was 51.56.8 years. The difference in age between the control group (40.5 10 years) and both experimental groups was not significant. The mean score on the Hamilton Rating Scale for Depression was 41.26.7 (group I) and 38.35.9 (group II). This difference was not statistically significant.

Sleep variables of groups I and II and the control group are presented in Table I. On the first night, the percentage of slow-wave sleep (SWS%) in group I was not significantly different from SWS% in the control group. However, the distribution of SWS in these two groups was different. In the first two cycles, SWS% was higher with the control group, whereas in the last two cycles it was significantly higher in group I. On the second night, SWS% was also not significantly lower in group I, and the distribution of SWS was the same: in the first two cycles it was higher in the control group, whereas in the last two cycles it was higher in group I. Thus, SWS in group I was only slightly reduced when compared with the control group, a finding atypical for depression, but it was shifted to the second part of the night. In contrast, in group II, SWS% was reduced for both nights in comparison to the control group. In group II, SWS was significantly reduced (p<0.01) in the first two cycles, whereas there was no significant difference in SWS between both groups for the subsequent two cycles. Thus, there was a normal distribution of SWS in group II, and the decreased SWS% was a result of the reduction of SWS in the first two cycles. We should emphasize that seven patients in group I demonstrated a prominent increase in SWS in the last two cycles, attributable to the average SWS% of group I being significantly higher than SWS% of group II (p<0.05).

On the first night, REM sleep latency in group I was similar to that of the control group. On the second night, REM sleep latency in both groups was decreased when compared with the first night, an expected finding because group I and the control group both demonstrated the first night effect. However, the decrease in REM sleep latency on the second night was more prominent in group I, resulting in REM sleep latency on the second night being significantly lower in group I than in the control group. In group II, the first night effect was absent and there was a tendency toward paradoxical REM sleep evolution; that is, on the first night, REM sleep latency in group II was reduced in comparison to group I and the control group, whereas on the second night REM sleep latency in group II was not significantly greater than that of the control group, but was significantly greater than that of group I (p<0.01).

REM% on the first night was not significantly increased in group II when compared with the other groups, but on the second night was significantly higher in group II than in the control group (p<O.Q5). In group I, REM% tended to be reduced on the first night in comparison to the second, but this difference was not significant.

In the two groups of patients, REM sleep was shifted to the first cycle in both nights when compared with the control group, a finding typical for depression, but with this shift being especially prominent on the first night for group I. The proportion of the first REM sleep episode to total REM sleep duration was the highest on the first night of group I, although, as previously mentioned, REM sleep on this night was slightly reduced in comparison to the second night.

Table I. Sleep variables in the control group and in depressed patients with and without first night effect

Sleep variables Group I (First night effect present) Level of significance Control group Level of significance Group II (First night effect absent)
First night

LPREM (min)


NS 113.234.4



10.0± 9.7


12.4 6.3

P <0.01 4.7 67


17.8± 16.8 P <0.01 46.1 23.7 P <0.01 11.9 13.1


17.1± 27.9 P <0.05 6.8 8.0 NS 8.2 24.5
REM% 18.2± 10.9 NS 19.4 4.0


23.7 8.8



P <0.01 10.0 P <0.01 33.5
Second night
LPREM (min) 35.8 41.8 P <0.01 69.3 16.9 NS 86.5 ±62.9
SWS% 11.3 13.6 NS 15.9 6.1 P <005 5.5 ±5.8
SWS1,2min 17.6 24.0 P < 0.01 51.6 13.8 P <0.01 18.2± 17.4


20.5 27.9 P <0.05 9.0 12.5 NS 4.6± 8.7
REM% 24.7 10.2 NS 19.0 3.8 P <0.05

25.4± 8.4



P < 0.02 9.5 P <0.01 32.1

LPREM-REM sleep latency; SWSslow-wave sleep; 1, 2, 3, 4sleep cycles; REMttotal REM sleep in minutes.

Table II. Eye movement density in the control group and in depressed patients with and without first night effect

Sleep cycles Group I Level of significance Control Level of significance Group II
First night
1 6.9 5.6 p < 0.02

2.7 2.5

p < 0.05

5.0 3.7


5.7 3.5


3.7 2.3


4.4 3.8


5.7 6.4


5.0 4.4


4.0 4.4


4.6 7.6


5.8 3.7


6.1 5.2

Second night


5.9 4.5

p < 0.02

1.5 1.7

p < 0.02

4.6 2.5


7.3 3.9

p < 0.01

3.7 2.1

p < 0.02

5.0 2.4


4.9 4.1


5.9 4.1

p < 0.05

3.0 2.3


6.9 4.8


4.9 3.3


6.8 4.3

The distribution of EM density for the different cycles is presented in Table II. On the first night, the first cycle EM density was significantly higher in both groups of depressed patients than in the control group (p<0.02 and p<0.05, respectively). On the second night, EM density in the depressed patients was increased not only in the first, but also in the second cycle, and that increase was especially prominent in group I. Groups were also different according to the number of short sleep cycles (less than 40 minutes, Table III). Patients in group I, in comparison to the control group, showed a significant increase in the number of short cycles, whereas the difference between the number of short cycles in group II and the control group was less prominent.

Correlations between sleep variables were also different in groups of patients. In group I, SWS in the third cycle correlated positively with EM density in the second REM period (r=0.50, p<0.01), whereas in group II it correlated positively with EM density in the same cycle (r=0.57,p<0.01). It is necessary to take into consideration that, in group I, SWS was relatively increased in the third cycle, whereas in the third cycle group II SWS was low and EM density was the lowest. In group I, the duration of the first REM period correlated negatively with EM density in the third and fourth cycles (r=-0.59, p<0.01), whereas in group II REM sleep duration of the first cycle correlated positively with EM density in the same cycle (r=0.55,p<0.01). In group I, the subjective estimation of sleep duration and feeling of being refreshed after sleep correlated positively with SWS duration (r=0.48; r=0.45). In the control group, the subjective estimation of sleep duration also correlated positively with SWS, whereas in group II the subjective estimation of sleep duration correlated with REM duration in the first cycle (r=0.40).

In addition to all aforementioned differences in sleep variables, we have confirmed the findings of Ansseau et al. [21], who showed that groups of patients are different in their resistance to treatment. All patients in group I went into remission without electroconvulsive therapy (), however, this was true for only one patient in group II (p<0.01).

Table III. Percentage of sleep cycles < 40 minutes in the total number of cycles in the control group and in depressed patients with and without first night effect

Variables Group I Level of significance Control Level of significance Group II
Number of cycles 96 0 91   111
Percent of cycles<40min 18% p<0.001 6% p<0.05 14%


Our initial hypothesis has been partly confirmed by the results of the present investigation. The SWS% and REM% in group I (first night effect), but not group II, was more similar to that of the control group. SWS was directly related to the estimation of sleep quality in group I only and in healthy subjects SWS correlated positively with the estimation of sleep duration. At the same time, the groups of patients were different in their distribution of REM sleep and SWS. In group I, REM sleep was shifted to the first cycle on the first night, and SWS on both nights was more often shifted to the last cycles. On the second night, EM density in group I achieved its maximum not in the first cycle (as is customary in depression) but in the second cycle. Moreover, this increase in EM density in the second cycle was related to the redistribution of SWS.

How are all these differences related to the first night effect? What is the meaning of the first night effect in depression? The literature and the results of the present investigation suggest that: (1) First night effect is a normal reaction of healthy subjects to a new unusual sleeping condition. In the present investigation, eight healthy subjects demonstrated a prominent first night effect. Thus, depressed patients able to present such a response may be more similar to healthy subjects than those patients who do not display a normal reaction to the new environment. (2) Depressed patients with first night effect have been shown to have a better clinical response to treatment (they do not need for the positive outcome) and have a shorter duration of the disease and of the current depressive episodes [21]. (3) In our investigation, SWS was higher in patients who demonstrated first night effect. SWS is a marker for a successful outcome of depression [26], as well as a successful adaptation to shift work [27]. According to the present investigation, the first night effect in depression is characterized by some specific changes in the REM sleep variables and sleep structure. REM sleep on the second night in group I was increased in comparison to the first night; REM sleep latency was significantly decreased (in comparison not only to the first night of the same group, but also to the second night of group II); and EM density was increased. Thus, it appears that there was an increased REM sleep pressure and REM rebound effect on the second night of group I after mild REM sleep suppression on the first night. A typical first night effect in healthy subjects is not accompanied by the rebound effect of REM sleep. Thus, our data confirm the increased REM sleep requirement in depression. Even a small (not significant) REM sleep reduction on the first night was shown to be compensated by REM rebound, and the first REM sleep episode on the first night of group I was also significantly increased. The increased REM sleep pressure in group I was confirmed by the increased number of short cycles in this groupeven in the middle of the night REM sleep in this group appears earlier than in the group without first night effect and earlier than in healthy subjects.

The increased REM sleep pressure in depression was shown previously [6]; however, our findings emphasize that this pressure is especially high after the first night effect, and is accompanied by REM rebound. If the first night effect is a marker of the neurological protective reserves in depressed patients, it is reasonable to consider REM rebound effect as an attempt at psychophysiological adaptation. Such an attempt is evident either in the redistribution of REM or in the alteration of REM sleep quality. The increased duration of the first REM sleep episode may reflect an effort to satisfy the increased REM sleep pressure. This tendency is especially strong when REM sleep duration is partly reduced and delayed due to the first night effect. However, lengthening of the first REM episode was typically not exclusive to group I. In our investigation the first REM episode was more prolonged compared with the second REM episode in 43% of all cases in group 1, and in 40% in group II. Therefore, increased duration of the first REM episode was more prominent in the relatively more adapted group of patients; such an increase, by itself, is not a marker of adaptability, and is not sufficient to satisfy the increased REM sleep requirement. The functional meaning of the first REM sleep episode is presumably less important than the functional meaning of subsequent REM episodes. For instance, the second REM sleep episode seems to have some advantages to psychophysiological adaptation, at least in comparison to the first REM sleep episode, because the second REM sleep episode was increased in healthy students in the process of adaptation to stress [28]. However, REM sleep quality, and especially quality of the second and third REM sleep episodes, may be even more important than the duration of REM sleep.

It is possible to assume that EM density is related to the adaptive function of REM sleep. EM density correlates with the subject's active participation in his dream events [29], and such participation is very important for the restoration of search activity. After REM sleep with an increased EM density, mood improves in a positive direction [30]. The adaptation to stress of bereavement correlates with an increase in EM density [31]. According to our results, EM density in almost all cycles, except the first, correlates with SWS in the third cycle in both groups. In depression, the episodic increase of SWS in the second part of the night is predicted by the increased EM density exactly before "this explosion" of SWS [12], whereas REM sleep duration does not predict such increase of SWS. Thus, REM sleep quality (reflected in an increased EM density) seems to be more important than REM sleep duration. However, the increase of EM density in the first cycle is typical for all depressed patients, as opposed to healthy subjects, and seems to be less important than the increase of EM density in subsequent cycles. In group I, on the second night, EM density is increased predominantly in the second REM sleep episode, and exactly in this group, EM density in the second cycle correlates positively with SWS in the subsequent cycle. This correlation confirms the adaptive nature of the increased EM density. Our findings are in agreement with those of Buysse et al. [32] who showed that depressed patients without recurrence of depression after psycho-therapeutic treatment showed, at baseline, an increased EM density from night 1 to night 2, whereas those with recurrence showed no change, in a manner similar to patients of our group II.

We do not wish to imply that the increased duration of the first REM sleep period and the increase of EM in this period are absolutely irrelevant for the process of adaptation. According to Cartwright and Lloyd [33] depressed patients who had an enhanced dreamlike quality of mentation during the first REM sleep period showed decreased Beck Depression Inventory scores at follow-up assessment. Thus, the first REM sleep episode can also contribute to adaptation if the psychic activity in this period is intense. However, it is presumably more difficult to achieve such a level of mental activity in the first REM sleep episode than in the subsequent episodes. In normal sleep, dreamlike mental activity is almost absent in the first REM sleep episode. In the most depressed patients, the relative increase of EM density in the first REM sleep period is not sufficient enough for the restoration of mental health. Moreover, in both groups, EM density in the first REM sleep episode correlates with the subjective estimation of sleep latency resulting in an underestimation of sleep in the first cycle. The positive role of REM sleep in sleep estimation in healthy subjects [34], supports our proposal that there is a functional deficiency of the first REM sleep episode in depression.


Sleep Questionnaire (for presentation in the morning after polysomnography)

1. How long have you slept?

More than 9 hours; 8.5 hours; 8 hours; 7.5 hours; 6.5 hours; 6 hours; 5.5 hours; 5 hours; less than 5 hours; no sleep at all

2. How long did it take your to fall asleep?

More than 1 hour; 1 hour; 50 min; 40 min; 30 min; 20 mm; 10 min; 5 min; less than 5 min

3. How many awakenings have you had during sleep? 0; 1; 2; 3; more than 3

4. How long did it take you to go back to sleep after each awakening?

Less than 1 min; 1-5 min; 6-10 min; 11-20 min; more than 20 min (please specify to what awakening (first, second, etc.) belongs the duration of wakefulness)

5. Are you refreshed after sleep? Totally; partly; not refreshed

6. Was your sleep in the first part of the night: Deep; moderate; superficial

7. Was your sleep in the second part of the night: Deep; moderate; superficial

8. Have you experienced dreams during your sleep? Yes; no.

9. Do you remember some of your dreams? Yes; no

10. Have you been:

(A) An observer; (B) An active participant in your dream content

11. Your dreams were predominantly:

Pleasant; neutral; unpleasant; frightening

12. Was it interesting for you to watch dreams: (A) Yes; () n; () no answer

13. Does your sleep mood after sleep become:

Better than in the evening; the same; worse than in the evening


1. Coble PA, Kupfer DJ, Shaw DH Distribution of REM latency in depression. Biol. Psychiatry 1981,16: 453-466.

2. Kupfer DJ, Ulrich RF, Coble PA, Jarrett DB, Grochocinski VJ, Doman J, Matthews G, Borbely AA. The application of automated REM and slow wave sleep analysis (normal and depressives). Psychiatric Res. 1984;13:325-334.

3. Reynolds CF III, Kupfer D. Sleep in depression. In: Williams RZ, Karacan I, Moore CA, eds. Sleep disorders, diagnosis and treatment. New York: John Wiley 1988:147-164.

4. Benca RM, Obermeyer WH, Thisted RA. Gillin JC. Sleep and psychiatric disorders: a meta-analysis. Arch Gen Psychiatry 1992:49:651-670.

5. Riemann D, Berger M. EEG sleep in depression and in remission and the REM sleep response to the cholinergic agonist RS 86. Neuropsychopharmacology 1989:2:145-152.

6. Thase ME, Reynolds CF III, Frank E, Jennings JR, Nofzinger E, Fasiczka BA, Garamoni G, Kupfer D. Polysomnographic studies of unmedicated depressed men before and after cognitive behavioral therapy. Am. J. Psychiatry 1994:151:1615-1622.

7. Rush AJ, Erman MK. Giles DE. Schlesser MA. Carpenter G, Vasavada N, Roffwarg HP. Polysomnographic findings in recently drug-free and clinically remitted depressed patients. Arch. Gen. Psychiatry 1986:43:878-884.

8. Thase ME, Simons AD. The applied use of psychotherapy in the study of the psychobiology of depression. J. Psychother. Pract. Res. 1992:1:72-80.

9. Buysse DJ, Kupfer DJ, Frank E, Monk TH. Ritenour A. Electroencephalographic sleep studies in depressed outpatients treated with interpersonal psychotherapy: II. Longitudinal studies at baseline and recovery. Psychiatry Res. 1992;40:27

10. Vogel GW, Thurmond A. Gibbons P, Sloan K. Boyd M, Walker M. REM sleep reduction effects on depression syndromes. Arch Gen Psychiatry 1975;32:765-77.

11. Vogel GW, Vogel F, McAbeer RS, Thurmond AJ. Improvement of depression by REM sleep deprivation. Psychiatry 1980:37:247-253.

12. Rotenberg VS. Sirota P. Abramov S, Steinbock M, Shamir D, Melamed Y, Elizur A. REM density in REM sleep and the distribution of delta sleep. J Sleep Res 1994;3(suppl l):222.

13. Greenberg R, Pearlman C. Cutting the REM sleep nerve: an approach to the adaptive role of REM sleep. Persp Biol. Med. 1974;17:513-521.

14. Rotenberg VS. Search activity in the context of psychosomatic disturbances, of brain monoamines and REM sleep function. Pavlov. J. Biol. Sci. 1984;10:1-15.

15. Rotenberg VS. REM sleep and dreams as mechanisms of the recovery of search activity. In: Moffitt A, Kramer M, Hoffmann R, eds. The function of dreaming. New York: University of New York Press 1993.

16. Rotenberg VS. The revised monoamine hypothesis: Mechanism of antidepressant treatment in the context of behavior. Integr. Psychol. Behav. Sci. 1994;29:182-188.

17. Rechtschaffen A, Gililand M, Bergmann B. Winter J. Physiological correlates of prolonged sleep deprivation in rats. Science 1983;221:182-184.

18. Agnew H, Webb W, Williams R. The first night effect: an EEG study of sleep. Psychophysiology 1966;2:263-266.

19. Browman C, Cartwright R. The first night effect on sleep and dreams. Biol Psychiatry 1980; 15:809-812.

20. Reynolds CF III, Newton TF, Shaw DM, Coble PA, Kupfer DJ. Electroencephalographic sleep findings in depressed outpatients. Psychiatry Res. 1982:6:65-72.

21. Ansseau M, Kupfer DJ, Reynolds CF III. Internight variability of REM latency in major depression: implications for the use of REM latency as a biological correlate. Biol Psychiatry 1985;20:489-505.

22. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960:23:56-62.

23. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects: Department of Health, Education and Welfare. National Institutes of Health Publication 204. Washington, DC: U.S. Government Printing Office 1968.

24. Benson KL, Zarcone VP. Rapid eye movement sleep and eye movements in schizophrenia and depression. Arch Gen Psychiatry 1993;50(6):474-82.

25. Schmidt HS, Kaelbling R. The differential laboratory adaptation of sleep parameters. Biol Psychiatry 1971:3:33^5.

26. Reynolds CF III, Hoch CC, Buysse DJ, George CJ, Houck PR, Mazumdar S, Miller M, Pollock BG, Rifai H, Frank E. Sleep in late-life recurrent depression: changes during early continuation therapy with nortriptyline. Neuropsychopharmacology 1991;5:85-96.

27. Rotenberg VS. The competition between SWS and REM sleep as index of maladaptation to shift work. Homeostasis 1992:33:235-238.

28. Rotenberg VS, Arshavsky VV. REM sleep, stress and search activity. Waking Sleeping 1979; 3:235-244.

29. Rotenberg VS. Functional deficiency of REM sleep and its role in pathogenesis of neurotic and psychosomatic disturbances. Pavlov J Biol Sci 1988;23:l-3.

REM sleep in depressed patients

30. Kupfer DJ, Hemnger GR REM activity as a correlate of mood changes throughout the night Arch Gen Psychiatry 1972 27 368-373

31. Reynolds CF III, Hoch CC, Buysse DJ, Houck PR, Schlernltzauer M, Pasternak RF, Frank E, Mazumdar S, Kupfer DJ Sleep after spousal bereavement a study of recovery from stress Biol Psychiatry 1993,34791-797

32. Buysse DJ, Kupfer DJ Frank E, Monk T, Rifenour A Do electroencephalographic studies predict recurrence in depressed patients successfully treated with psychotherapy9 Depression 1994,2 10V108

33. Cartwnght RD, Lloid SR Early REM sleep a compensatory change in depression7 Psychiatry Res 199451245-252

34. Rotenberg VS The estimation of sleep quality in different stages and sleep cycles J Sleep Res 1993,2 17-20