Spontaneous sleep interruptions during extended nights. Relationships with NREM and REM sleep phases and effects on REM sleep regulation

Clinical Neurophysiology 113 (2002) 892–900 www.elsevier.com/locate/clinph

Spontaneous sleep interruptions during extended nights. Relationships with NREM and REM sleep phases and effects on REM sleep regulation Giuseppe Barbato*, Charles Barker, Charles Bender, Thomas A. Wehr Section on Biological Rhythm, NIMH, Bethesda, MD, USA Accepted 6 March 2002

Abstract Objectives: There is no agreement in the literature as to whether sleep interruption causes rapid eye movement (REM) pressure to increase, and if so, whether this increase is expressed as shortened REM latency, increased REM density, or increased duration of REM sleep. The purpose of the present study was to examine the effect of different durations of spontaneous sleep interruptions on the regulation of REM sleep that occurs after return to sleep. Methods: The occurrence of spontaneous periods of wakefulness and their effects on subsequent REM sleep periods were analysed in a total sample of 1189 sleep interruptions which occurred across 364 extended nights in 13 normal subjects. Results: Compared with sleep interruptions that last less than 10 min, sleep interruptions that last longer than 10 min occur preferentially out of REM sleep. In both the short and long types of sleep interruptions, the duration of REM periods that ended in wakefulness were shorter than the duration of those that were not interrupted by wakefulness. REM densities of the REM periods that terminated in periods of wakefulness were higher than those of uninterrupted REM periods. The proportion of episodes of wakefulness following REM sleep that were long-lasting progressively increased over the course of the extended night period. The sleep episodes that followed the periods of wakefulness were characterised by a short REM latency. REM duration was increased in episodes that followed long sleep interruptions compared to those that followed short sleep interruptions. REM density did not appear to change significantly in the episodes that followed sleep interruption. Conclusions: REM sleep mechanisms appear to be the main force controlling sleep after a spontaneous sleep interruption, presumably because during the second half of the night, where more sleep interruptions occur, the pressure for non-rapid eye movement sleep is reduced and the circadian rhythm in REM sleep propensity reaches its peak. Processes promoting REM sleep at the end of the night are consistent with the Pittendrigh and Daan dual oscillator model of the circadian pacemaker. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Non-rapid eye movement sleep; Rapid eye movement sleep; Awakening; Circadian rhythm

1. Introduction In humans, natural nocturnal sleep is generally considered to occur in a consolidated ‘monophasic’ period; however, even in young adults, in whom sleep efficiency is high, the sleep period can be interrupted by an estimated 10% of awake time (Dijk and Kronauer, 1999). These sleep interruptions are more frequent in infants (Schulz et al., 1985; Ficca et al., 1999), when sleep is not yet consolidated, and in older subjects (Agnew et al., 1967; Feinberg et al., 1967; Webb and Campbell, 1980; Salzarulo et al.,1999; Dijk et al., 2001). Fragmentation of sleep also occurs in * Corresponding author. Department of Psychology, Second University of Naples, Via Vivaldi 43, 81100 Caserta, Italy. Tel.: 1039-0823-274790, fax: 1039-0823-274792. E-mail address: [email protected] (G. Barbato).

unusual experimental conditions, such as time-free environments (Schulz and Zulley, 1980; Weitzman et al., 1980; Campbell, 1985), forced desynchrony protocols (Dijk and Czeisler, 1995) or long scotoperiods (Wehr, 1992). Spontaneous transitions from sleep to intermittent wake episodes tend mainly to occur out of rapid eye movement (REM) sleep (Langford et al., 1972; Schulz and Zulley, 1980; Weitzman et al., 1980; Barbato et al., 1994; Murphy et al., 2000; Dijk et al., 2001), suggesting that REM sleep may establish physiological conditions that facilitate transitions to wakefulness (Snyder,1966; Broughton, 1968; Lavie et al.,1979; Wehr, 1992). The effect that these sleep interruptions have on REM sleep regulation is a more controversial issue. Brezinova (1974) reported that in middle-aged people, on average, REM cycles interrupted by spontaneous awakening are

1388-2457/02/$ - see front matter. q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S13 88- 2457(02)0008 1-0

CLINPH 2001123

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

20 min longer than intact ones, despite the fact that the periods of intervening wakefulness are only 10 min. Commenting on this result, Brezinova suggested that a new sequence of the non-rapid eye movement (NREM)– REM cycle seems to be initiated from its beginning after such an awakening. Merica and Gaillard (1986) reported a decrease in slow wave sleep in cycles spontaneously interrupted by long periods of awakening compared with what could have been expected in cycles without interruption; however, these long wake interruptions did not appear to reset the cycle clock to zero. Brezinova et al. (1975) awakened subjects for 1 h after the second REM period of the night. They reported that the first NREM period following the awakening was shorter than the corresponding NREM period on uninterrupted nights, and was similar in duration to the first NREM of the night. On average, the first REM period after the awakening was similar in duration to the corresponding period of the uninterrupted night, suggesting that factors controlling its duration were not affected by the awakening. Bunnell et al. (1984) also reported no change in the duration of REM episodes that followed 1 h forced awakenings. On the other hand, they reported a significant increase in eye movement density in these episodes. Campbell (1985) reported a 35% increase in the duration of the first REM episode and no change in REM density, following return to sleep after forced awakenings intervals of 20–120 min. He hypothesised that increased ‘REM pressure’ that results from enforced wakefulness may be expressed either by increased intensity of subsequent REM, as reflected in eye movement measures, or increased duration of the episodes, as reflected by duration of REM, without increased intensity. Miyasita et al. (1989) awakened subjects for 10–90 min. Following a return to sleep, REM latencies exhibited a bimodal distribution with a 25–30 min interval between peaks. About one-third of the sleep episodes were characterised by short onset REM latencies (REM latencies of 25 min or less). Foret et al. (1990) found that 10 min of forced interruption of sleep shortened the inter-REM interval if the awakening occurred during an episode of REM sleep, but that it lengthened the inter-REM interval if it occurred during an episode of NREM sleep. Thus, taken together, there is no agreement among the studies as to whether sleep interruption causes REM pressure to increase, and if so, whether this increase is expressed as shortened REM latency, increased REM density, or increased duration of REM sleep. Differences in duration of sleep interruptions in the different studies is one factor that might account for this lack of agreement. In a previous paper (Barbato et al., 1994) we reported that subjects who rest and sleep in an extended 14 h period of darkness each night show more frequent periods of sustained wakefulness during sleep than occur in conven-

893

tional sleep schedules (Wehr, 1996). The purpose of the present study was to examine the relationship of these sleep interruptions to NREM and REM phases and their effects on the regulation of REM sleep that occurs after return to sleep.

2. Methods The analysis was carried out on 364 extended nights of 13 healthy male volunteers, aged 20–34 years, who participated in an experiment that was designed to investigate the effects of alterations in duration of the photoperiod (the illuminated fraction of the day) on sleep and endocrine systems (Wehr et al., 1993). During one phase of the experiment, the subjects lived for 4 weeks in a winter-type photoperiod in which they went about their usual activities in ambient natural and artificial light for 10 h (08:00– 18:00 hrs) each day and were confined to a dark room for 14 h (18:00–08:00 hrs) each night. The subjects were instructed to remain at bed rest and to sleep whenever possible during the dark period. Polysomnographic sleep was monitored with a Grass 78 D polygraph. The sleep stages were visually scored according to the criteria of Rechtschaffen and Kales (1968). Sleep onset was defined as the first minute of 3 consecutive minutes of stage 2, 3, 4 or REM. A NREM–REM cycle was defined by the successive occurrence of a NREM episode and of a REM episode. The end of the cycle was defined by the occurrence of a new NREM episode or by the occurrence of wakefulness. Since no standard criteria exist for defining a period of wakefulness during the sleep phase, we defined spontaneous awakenings as periods of at least 3.0 consecutive min of wakefulness (stage 0) (Barbato et al., 1994), consistent with criteria previously adopted by Merica and Gaillard (1985). This criterion was adopted to clearly identify periods of sustained spontaneous wakefulness which could have interrupted sleep processes and have an effect on the following period of sleep. Sequences of REM sleep were combined into one REM episode if the interval between them was less than 3 min. This criterion for consolidating sequences of REM sleep is more stringent than previous criteria in literature. The criterion was selected to be consistent with the 3 min criterion for the awake period. Two types of sleep interruption were then defined: short ¼ sleep interruption lasting at least 3 min and not longer than 9.5 min; long ¼ sleep interruption lasting at least 10 min. The 10 min cut-off was selected on the basis of previous studies on the effect of forced awakenings on sleep cycles (Miyasita et al., 1989; Foret et al.,1990). Sleep-onset following a period of sleep interruption was defined as the first minute of 3 consecutive minutes of stage 2, 3, 4 or stage REM. In accordance with previous studies (Miyasita et al., 1989, Bes et al., 1996, Sasaki et al., 2000) a

894

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

Fig. 1. Distribution of short (less than 10 min) and long (equal or longer than 10 min) sleep interruptions across the 28 extended (14 h) nights.

sleep-onset REM period (SOREMP) was defined as one having a REM latency shorter than 25 min. REM sleep variables included: REM latency, the interval between sleep onset and the first epoch of REM sleep; REM duration, the interval from the REM onset to REM offset; and REM density (total REM activity/REM duration). Sleep episodes following the sleep interruptions were categorised as: interrupted short ¼ episodes preceded by periods of short sleep interruption (short interruption: 3– 9.5 min); interrupted long ¼ episodes preceded by a long

sleep interruption (long interruption: at least 10 min); and not interrupted ¼ episodes not preceded by periods of sleep interruptions. The former two groups were also differentiated on the basis of SOREMP’s occurrence. All episodes were classified according to their time of occurrence during the course of the night in 7 consecutive 2 h intervals beginning at lights-off (18:00 hrs). The data corresponding to the periods preceding wakefulness included periods which ended within the 2 h block; the data corresponding to the periods following wakefulness included periods which started within the 2 h block. NREM time (REM latency), REM time and REM density for each 2 h intervals and for the types of episodes (short, long or no interruption) were averaged separately within each subjects over the 28 extended nights. Comparison between the different conditions were made at each time point and for the pooled groups with non-parametric Kruskal–Wallis test. Dunn’s multiple comparisons test was used for post hoc comparisons between the groups. 3. Results Spontaneous sleep interruptions occurred throughout the whole night period. The length (mean ^ SD) of spontaneous short interruptions was 6.2 ^ 1.8 min (range 3– 9.5 min) (n ¼ 382). The length of spontaneous long interruptions was 66.5 ^ 55.6 min (range 10–302 min) (n ¼ 807). Over the 28 days of night sleep extension, the number of short interruptions showed a modest decline, whereas the number of long interruptions was stable following the third day of night sleep extension (Fig. 1). The distribution over the night of awakenings was different for awakenings from NREM sleep and awakenings from REM sleep. Short awakenings from NREM occurred more frequently in the first half of the night (Fig. 2a). Long awakenings from REM showed a progressive increase across the night (Fig. 2b), while long awakenings from NREM showed a major peak at the end of the night (Fig. 2b). Long sleep interruptions occurred more frequently from REM sleep than from NREM sleep (df ¼ 12, t ¼ 3:17, P ¼ 0:008). Short sleep interruptions occurred equally frequently from REM and from NREM sleep. 3.1. Periods of sleep preceding interruptions

Fig. 2. (a) Distribution of short (less than 10 min) sleep interruptions during the 14 h (18:00–08:00 A.M.) long night period. (b) Distribution of long (equal or longer than 10 min) sleep interruptions during the 14 h (18:00– 8:00 A.M.) long night period.

For both short and long sleep interruptions (Table 1), the duration of REM period preceding the interruption was shorter (H ¼ 18:93, P , 0:0001). Post hoc analysis showed significantly shorter REM duration for short interruption (17.7 min)(Dunn’s test: P , 0; 001) and long interruption (18.4 min.)(Dunn’s test: P , 0:01) compared with no interruption (23.3 min.). REM densities were higher in REM periods that ended in wakefulness than in those that were not interrupted by wakefulness (H ¼ 30:17, P , 0:0001). Post hoc analysis

Time

18:00–20:00

20:00–22:00

22:00–24:00

00:00–02:00

02:00–04:00

04:00–06:00

06:00–08:00

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

NREM duration No interruption Short interruption Long interruption

9 2 2

52.1 (4.6) 42.1 (12.9) 45.3 (4.8)

13 8 6

62.7 (3.3) 73.3(15.4) 74.5 (5.2)

13 10 11

73.4 (3.2) 79.2(13.2) 75.3 (4.5)

13 10 13

71.6 (3.8) 67.2 (5.5) 76.6 (4.3)

13 7 12

61.2 (2.4) 65.1 (7.9) 66.3 (3.8)

13 12 13

45.6 (3.6) 47.9 (5.1) 49.3 (4.6)

40.8 (2.3) 43.5 (2.6) 39.2 (1.7)

Kruskal–Wallis

H ¼ 1.3

NS

H ¼ 1.8

NS

H ¼ 0.2

NS

H ¼ 2.6

NS

H ¼ 1.0

NS

H ¼ 0.2

NS

13 10 13 13 H ¼ 1.9

REM duration No interruption Short interruption Long interruption

9 2 2

12.9 (2.6) 16.8 (0.8) 4.5 (0.3)

13 8 6

13.7 (1.0) 12.3(1.9) 18.5 (2.7)

13 9 11

18.6 (1.3) 15.4(1.1) 17.7 (1.8)

13 10 13

24.4 (1.7) 15.3 (1.8) 16.5 (1.3)

13 7 12

26.9 (1.3) 18.1 (2.4) 18.3 (1.3)

13 12 13

35.3 (1.8) 24.0 (3.0) 22.5 (2.5)

28.1 (1.7) 18.8 (2.6) 19.2 (1.2)

Kruskal–Wallis

H ¼ 0.4

NS

H ¼ 4.1

NS

H ¼ 2.3

NS

H ¼ 13.8

P ¼ 0.001

H ¼ 13.9

P ¼ 0.001

H ¼ 12.5

P ¼ 0.002

13 10 13 13 H ¼ 11.6

REM density No interruption Short interruption Long interruption

9 2 2

13 12 13

1.87 (0.15) 2.02 (0.21) 2.35 (0.23)

1.73 (0.17) 1.89 (0.21) 2.26 (0.20)

Kruskal–Wallis

H ¼ 3.2

H ¼ 3.1

NS

13 10 13 13 H ¼ 3.8

0.92 (0.18) 0.65 (0.60) 2.17 (0.83) NS

13 8 6 H ¼ 1.5

1.26 (0.16) 1.85 (0.42) 1.37 (0.25) NS

13 9 11 H ¼ 7.0

1.74 (0.17) 2.25 (0.46) 2.82 (0.36) P ¼ 0.03

13 10 13 H ¼ 10.2

1.96 (0.16) 2.74 (0.25) 2.85 (0.23) P ¼ 0.006

13 7 12 H ¼ 8.6

2.08 (0.16) 2.96 (0.19) 2.54 (0.24) P ¼ 0.01

NS

P ¼ 0.003

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

Table 1 NREM duration (REM latency), REM duration and REM density preceding sleep interruptions

NS

895

896

Time

18:00–20:00

20:00–22:00

22:00–24:00

00:00–02:00

02:00–04:00

04:00–06:00

06:00–08:00

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

s

Mean (es)

NREM duration No interruption Interruption short Interruption long

7 4 3

76.2 (5.5) 56.0 (3.3) 50.5 (19.3)

13 10 8

85.9 (3.8) 49.4 (4.8) 33.7 (7.6)

13 11 10

82.3 (3.8) 40.4 (5.9) 56.0 (4.7)

13 12 13

78.2 (4.0) 51.8 (9.4) 46.2 (2.9)

13 10 13

61.3 (3.2) 29.6 (6.0) 39.7 (1.6)

13 11 13

47.1 (2.5) 20.8 (2.3) 20.0 (1.9)

41.7 (2.7) 22.3 (4.3) 15.3 (3.0)

Kruskal–Wallis

H ¼ 4.5 NS

21.0

P ¼ 0.0001

H ¼ 20.3 P ¼ 0.0001

H ¼ 18.7 P ¼ 0.0001

H ¼ 20.4 P ¼ .0001

H ¼ 24.3 P ¼ 0.0001

13 10 11 13 H ¼ 18.5

REM duration No interruption Interruption short Interruption long

7 4 3

13 10 8

21.9 (2.4) 15.1 (1.9) 14.9 (3.0)

13 11 10

22.3 (1.6) 18.6 (2.1) 25.5 (3.9)

13 13 13

23.3 (1.5) 21.9 (2.6) 24.1 (1.9)

13 10 13

31.0 (2.2) 28.8 (3.7) 33.8 (2.4)

13 11 13

14.4 (1.4) 18.2 (0.0) 21.1 (2.3)

Kruskal–Wallis

H ¼ 1.2 NS

H ¼ 6.3 P ¼ 0.05

H ¼ 2.3

NS

H ¼ 1.5

NS

H ¼ 1.4

NS

H ¼ 10.7 P ¼ 0.005

13 10 12 13 H ¼ 4.9

REM density No interruption Interruption short Interruption long

7 4 3

13 10 8

13 11 10

Kruskal–Wallis

H ¼ 0.6 NS

20.4 (3.5) 14.4 (5.4) 15.0 (5.8)

1.83 (0.42) 1.57 (0.43) 1.23 (0.69)

2.11 (0.16) 2.03 (0.34) 1.70 (0.41)

H ¼ 2.7 NS

H ¼ 0.9

2.23 (0.19)) 13 2.32 (0.28) 13 2.03 (0.22) 13 NS

H ¼ 0.3

2.19 (0.18) 2.26 (0.20) 2.19 (0.19) NS

13 10 13

2.15 (0.17) 2.10 (0.18) 2.16 (0.22)

H ¼ 0.04 NS

13 11 13 H ¼ 0.1

24.4 (1.3) 19.7 (1.9) 28.9 (2.6)

1.94 (0.17) 1.95 (0.25) 2.10 (0.21) NS

13 10 12 13 H ¼ 1.5

P ¼ 0.0001

NS 2.02 (0.26) 2.15 (0.26) 1.82 (0.21) NS

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

Table 2 NREM duration (REM latency), REM duration and REM density following sleep interruptions

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

897

Fig. 4. Distribution of REM episodes (dashed lines) and of the sub-set of SOREMP episodes (solid lines) following short (first graph) and long (second graph) sleep interruptions throughout the extended 14 h night.

Fig. 3. REM latencies of sleep episodes following short sleep interruptions (first graph) and long sleep interruptions (second graph).

revealed significant higher REM densities for both short (2.20) (Dunn’s test P , 0:01) and long interruption (2.44) (Dunn’s test P , 0:001) compared with no interruption (1.69). No significant difference was found between duration of NREM of periods preceding the interruption (H ¼ 1:29, P ¼ 0:52, NS).

(25.0 min.) were longer than those of REM episodes that followed short sleep interruptions (20.1 min.) (Dunn’s test: P , 0:01). The distribution of REM latencies after short sleep interruptions showed a unimodal peak near 5 min (Fig. 3). The distribution of REM latencies after long sleep interruptions showed a bimodal profile with a peak near 5 min (REM latency was shorter than 3 min in 10% of cases) and a second peak near 60 min (Fig. 3). SOREM periods increased in frequency across the night (Fig. 4).

3.2. Periods of sleep following interruptions

4. Discussion

Compared with sleep episodes that were not interrupted by wakefulness (Table 2), sleep episodes that followed either short or long interruptions were characterised by a shorter NREM duration (H ¼ 82:17, P , 0:0001). Post hoc analysis revealed significant shorter NREM duration for both short (37.2 min) (Dunn’s test: P , 0:001) and long (35.6 min.)(Dunn’s test: P , 0:001) interruptions compared to no interruption (66.9 min.) Significant differences between episodes following different lengths of interruption were also found for REM time (H ¼ 11:9, P , 0:003), but not for REM density (H ¼ 1:17, P ¼ 0:56, NS). Post hoc analysis showed that duration of REM episodes that followed long sleep interruptions

Sleep interruptions occurred significantly more frequently out of REM sleep than out of NREM sleep. However, this difference was statistically significant only for interruptions that lasted at least 10 min. Thus, while brief sleep interruptions may occur during any NREM– REM sleep phases, long-lasting ones occur preferentially out of REM sleep episodes. The proportion of episodes of wakefulness following REM sleep that were long-lasting progressively increased over the course of the night, presumably because the homeostatic drive for sleep was decreasing (Feinberg, 1974; Borbe´ ly et al., 1981; Borbe´ ly, 1982). The association of transitions to wakefulness with REM

898

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

sleep is in agreement with hypotheses that REM sleep serves as a ‘gate’ to wakefulness (Snyder, 1966; Broughton, 1972; Lavie et al., 1979). Our finding that REM periods that ended in wakefulness were shorter than those that did not is consistent with observations of Campbell (1985) of sleep in individuals whose sleep–wake cycles were free running in temporal isolation. Murphy et al. (2000) have also recently reported that REM episodes from which spontaneous sleep terminated were truncated relative to those that did not end the sleep period. Our finding that REM density was higher in REM periods that ended in wakefulness than in those that did not is also consistent with previous findings (Barbato et al., 1994). Although higher REM density was associated with a greater likelihood of sleep interruption, the level of REM density did not predict the duration of the sleep interruption that followed. Nevertheless, the association of increased REM density with an increased likelihood of subsequent transitions to wakefulness appears to be consistent with other evidence indicating that heightened REM phasic activity reflects an arousing activation (Aserinsky, 1973; Feinberg et al., 1987). In this regard, Campbell (1985) has suggested that spontaneous transitions from REM sleep to wakefulness tend to occur when the neural activity that induces REM sleep is increasing exponentially in association with phasic activation (Mc Carley and Hobson, 1975; Orem and Keeling, 1980). The lowering of slow wave sleep (SWS) propensity due to chronic sleep satiety in this extended sleep paradigm must have facilitated this activation, increasing the tendency to awaken from sleep and to remain awake for long periods. Consistent with this interpretation is the reduced number of interruptions which occurred in the first 2 days of night extension, when also the total sleep durations were longer compared to the following days of night extension, suggesting that subjects may have been recovering from sleep-deprivation imposed by their habitual sleep schedule (Wehr, 1996). Indices of the REM episodes that occurred after return to sleep were significantly influenced by sleep interruptions. REM latency in sleep cycles that followed a period of sleep interruption was short, regardless of whether the duration of the interruption was short or long, confirming results of previous studies (Brezinova et al., 1975; Campbell, 1987; Miyasita et al., 1989). The frequency of occurrence of SOREM periods was higher in the early morning hours, at the time of the nocturnal rise of the circadian temperature rhythm, consistent with data reported by Bes et al. (1996). Sasaki et al. (2000) also recently presented evidence that the pattern of occurrence of SOREM episodes may be a consequence of the circadian variation in REM sleep propensity, which reaches a peak in the latter part of the night, and of the reduction in the homeostatic drive for slow wave sleep (SWS) that occurs as sleep progresses through the night. Duration of REM episodes that followed short periods of wakefulness were shorter than corresponding episodes that

followed long sleep interruptions. Two previous studies (Brezinova et al., 1975; Bunnell et al., 1984) reported no changes in duration of REM episodes that followed forced awakenings. In contrast, Campbell (1987) reported that of REM episodes that followed forced awakenings, 68% showed an increase in duration, while 16% showed a decrease. REM density showed no significant changes in REM episodes that followed sleep interruptions. Previous studies have shown conflicting results in this regard. Bunnel et al. (1984) reported increased eye movement density during the first REM episode that followed 1 h of forced wakefulness. Campbell (1987) reported no change in eye movement density during the first REM episode that followed forced awakenings, but reported decreased eye movement density during the second REM episode. It is difficult to compare these studies either with each other or with ours, owing to differences in the designs (forced versus spontaneous awakening, presence and quality of task during the waking period). Furthermore, the processes which contributes across the sleep period to sleep and wake regulation can differently influence the present findings. As sleep progresses, homeostatic pressure and NREM sleep pressure decrease, while circadian REM pressure increase. Short-term mechanisms regulating both REM sleep (Vivaldi et al., 1994; OcampoGarces et al., 2000) and process S build up (Achermann et al., 1993) during the wake interruption, can produce shortterm effects which modify the duration and the intensity of the different NREM–REM phases. Confounding variables associated with the special nature of the extended night protocol, should also be considered. Compared to habitual nights, the duration of the nocturnal period of low temperature during extended night increased 2.9 h (Wehr et al., 1993), with the nightly decline in rectal temperature occurring earlier. Also sporadic occurrence of stage 1 sleep during the periods of quiet wakefulness that precede and follow the bouts of sleep cannot be excluded, and this ‘light sleep’ activity might have altered the homeostatic processes previously mentioned. Even when one takes the previous cautious considerations into account, the present results suggest that a sleep interruption does not reset the NREM–REM cycle. Instead, what emerges from the sleep interruption is mainly a tendency to restart the cycle with a rapid progression to REM sleep. This phenomenon might be a consequence of increased REM propensity that results from the interruption of the previous REM episode and from the circadian increase in its level. The tendency for sleep, once it begins, to enter rapidly into REM sleep may also indicate that the drive for sleep at a time of night when the homeostatic drive for SWS is low is primarily an expression of the drive for REM sleep. In a previous paper (Barbato and Wehr, 1998), we proposed that REM sleep propensity is regulated by several processes whose relative importance changes over the course of the night. First, it is regulated by a homeostatic

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

process that compensates for REM sleep interruption or loss and which reaches highest levels at the end of the daytime period of wakefulness. However, this process is mutually antagonistic with a homeostatic process that regulates NREM (SWS) sleep. Because this latter homeostatic process is the principal regulator of NREM sleep, NREM propensity is high at the beginning of the sleep period and declines rapidly thereafter. Consequently, at the beginning of the sleep period, antagonism by NREM sleep is the predominant (negative) force governing REM sleep. As NREM propensity wanes, the homeostatic drive for REM sleep becomes a predominant (positive) force governing REM propensity. Then, as the homeostatic drive for REM sleep wanes, the circadian rhythm in REM sleep propensity, which reaches a peak at the end of the night, becomes the predominant (positive) force regulating REM sleep propensity. This scheme may partially explain the effects of spontaneous waking on subsequent sleep that we observed during extended sleep in the present experiment. The fact that REM sleep periods that terminated in wakefulness were shorter than those that did not may also have contributed to the increased REM propensity that was observed in the sleep cycle that followed the period of wakefulness. According to Benington and Heller (1994), if less REM sleep propensity is discharged in a REM episode, then a shorter interval of NREM sleep will intervene before the subsequent REM sleep episodes is triggered. Circadian variation in REM sleep propensity is probably a much more important factor contributing to the increased REM propensity that was observed in sleep cycles that followed periods of awakening. The circadian rhythm reaches its maximum in the latter part of the night when sleep cycles that followed sleep interruptions typically occurred. Wurts and Edgar (2000) recently reported evidence that the circadian control of REM sleep is much stronger than had previously been expected. In animals, they found that REM tendency after REM sleep deprivation was reduced in suprachiasmatic nuclei (SCN)-lesioned rats. In our study, the circadian influence on REM sleep could have been accentuated by the shift of the temperature curve induced by the dark extension. Wehr et al. (2001) have recently reported evidence that suggests that the human circadian timing system is regulated by two states, one diurnal and one nocturnal, which alternate with one another. The model is similar to the dual-oscillator model of Pittendrigh and Daan (1976), which postulates that the circadian pacemaker consists of two component oscillators, one of which controls physiological and behaviour changes that occur at the beginning of the night and is entrained to dusk, the other of which control changes that occur at the end of the night and is entrained to dawn. According to the dual oscillator model, NREM and REM sleep can be considered as separate processes which belong, respectively, to the dusk and the dawn states of the model. In summary, sustained spontaneous episodes of wakefulness in an extended night protocol occur preferentially out

899

of REM sleep episodes, consistent with the hypothesis that this sleep state is neurophysiologically close to wake. At sleep onset following episodes of spontaneous wakefulness, REM sleep mechanisms appear to be the main forces controlling sleep. Considering that the onset of sleep at the beginning of the night appears to be facilitated by the decline of a circadian rhythm in arousal (Lavie, 1986; Borbe´ ly and Achermann, 1999), sleep seems mainly controlled by two circadian processes, one facilitating sleep onset at the beginning of the night and another promoting REM sleep at the end of the night.

References Achermann P, Dijk DJ, Brunner DP, Borbely AA. A model of human sleep homeostasis based on EEG slow wave activity: quantitative comparison of data and simulations. Brain Res Bull 1993;31:97–113. Agnew HW, Webb WB, Williams RL. Sleep patterns in late middle age males: an EEG study. Electroenceph clin Neurophysiol 1967;23:168– 171. Aserinsky E. Relationship of rapid eye movement density to the prior accumulation of sleep and wakefulness. Psychophysiology 1973;10: 545–558. Barbato G, Barker C, Bender C, Giesen HA, Wehr TA. Extended sleep in humans in 14 hour nights (LD 10:14): relationship between REM density and spontaneous awakening. Electroenceph clin Neurophysiol 1994;90:291–297. Barbato G, Wehr TA. Homeostatic regulation of REM sleep in humans during extended sleep. Sleep 1998;21:267–276. Benington JH, Heller HC. REM sleep timing is controlled homeostatically by accumulation of REM-sleep propensity in non-REM sleep. Am J Physiol 1994;266:R1992–R2000. Bes FW, Jobert M, Muller LC, Schulz H. The diurnal distribution of sleep propensity: experimental data about the interaction of the propensities for slow-wave sleep and REM sleep. J Sleep Res 1996;5:90–98. Borbe´ ly A, Baumann F, Brandeis D, Strauch I, Lehmann D. Sleep deprivation: effect on sleep stages and EEG power density in man. Electroenceph clin Neurophysiol 1981;51:483–493. Borbe´ ly AA. A two process model of sleep regulation. Hum Neurobiol 1982;1:195–204. Borbely AA, Achermann P. Sleep homeostasis and models of sleep regulation. J Biol Rhythms 1999;14(6):557–568. Brezinova V. Sleep cycle content and sleep cycle duration. Electroenceph clin Neurophysiol 1974;36:275–282. Brezinova V, Beck U, Oswald I. Sleep cycle duration and timing of REM periods in interrupted night sleep. Int J Chronobiol 1975;3:81–87. Broughton RJ. Sleep disorders: disorders of arousal. Science 1968;159: 1070–1078. Bunnell DE, Bevier WC, Horvath SM. Sleep interruption and exercise. Sleep 1984;7:261–271. Campbell SS. Spontaneous termination of ad libitum sleep episodes with special reference to REM sleep. Electroenceph clin Neurophysiol 1985;60:237–242. Campbell SC. Evolution of sleep structure following brief intervals of wakefulness. Electroenceph clin Neurophysiol 1987;66:175–184. Dijk DJ, Czeisler CA. Contribution of the circadian pacemaker and the sleep homeostat to sleep propensity, sleep structure, electroencephalographic slow waves, and sleep spindle activity in humans. J Neurosci 1995;15:3526–3538. Dijk DJ, Kronauer RE. Commentary: models of sleep regulation: successes and continuing challenges. J Biol Rhythms 1999;14:569–573. Dijk DJ, Duffy JF, Czeisler CA. Age-related increase in awakenings:

900

G. Barbato et al. / Clinical Neurophysiology 113 (2002) 892–900

impaired consolidation of NonREM sleep at all ciradian phases. Sleep 2001;24:565–577. Feinberg I. Changes in sleep cycle patterns with age. J Psychiatr Res 1974;10:283–306. Feinberg I, Koresko RL, Heller N. EEG sleep patterns as a function of normal and pathological aging in man. J Psychiatr Res 1967;5:107–144. Feinberg I, Foyd TC, March JD. Effects of sleep loss on delta (0.3–3 Hz) EEG and eye movement density: new observations and hypotheses. Electroenceph clin Neurophysiol 1987;67:217–221. Ficca G, Fagioli I, Giganti F, Salzarulo P. Spontaneous awakenings from sleep in the first year of life. Early Hum Dev 1999;55:219–228. Foret J, Touron N, Clodore` M, Benoit O, Bouard G. Modifications of sleep structure by brief forced awakenings at different times of the night. Electroenceph clin Neurophysiol 1990;75:141–147. Langford GW, Meddis R, Pearson AJD. Spontaneous arousals from sleep in human subjects. Psychon Sci 1972;28:228–230. Lavie P. Ultrashort sleep waking schedule. Electronceph clin Neurophysiol 1986;63:414–425. Lavie P, Oksenberg A, Zomer J. It’s time you must wake up now. Percept Motor Skills 1979;49:447–450. Mc Carley RW, Hobson JA. Neuronal excitability modulation over the sleep cycle: a structural and mathematical model. Science 1975;189:58–60. Merica H, Gaillard JM. Statistical description and evaluation of the interrelationship of standard sleep variables for normal subjects. Sleep 1985;8:261–273. Merica H, Gaillard JM. Internal structure of sleep cycles in a healthy population. Sleep 1986;9:502–513. Miyasita A, Fukuda K, Inugami M. Effects of sleep interruption on REM– NREM cycle in nocturnal human sleep. Electroenceph clin Neurophysiol 1989;73:107–116. Murphy PJ, Rogers NL, Campbell SS. Age differences in the spontaneous termination of sleep. J Sleep Res 2000;9:27–34. Ocampo-Garces A, Molina E, Rodriguez A, Vivaldi E. Homeostasis of REM sleep after total and selective sleep deprivation in the rat. J Neurophysiol 2000;84:2699–2702. Orem J, Keeling J. Appendix: a compendium of physiology in sleep. In: Orem J, Barnes CD, editors. Physiology in sleep, New York, NY: Academic Press, 1980. pp. 315–335. Pittendrigh CS, Daan S. A functional analysis of circadian pacemakers in noctural rodents. J Comp Physiol A Sens Neural Behav Physiol 1976;106:333–335.

Rechtschaffen A, Kales A. A manual of standardized terminology techniques and scoring system for sleep stages of human subjects, , National Institute of Health, Publication 204. Washington, DC: US Government Printing Office, 1968. Salzarulo P, Fagioli I, Lombardo P, Gori S, Gneli C, Chiaramonti R, Murri L. Sleep stages preceding spontaneous awakenings in the elderly. Sleep Res Online 1999;2:73–77. Sasaki Y, Fukuda K, Takechi T, Inugami M, Miyasita A. Sleep onset REM period appearance rate is affected by REM propensity in circadian rhythm in normal nocturnal sleep. Clin Neurophysiol 2000;111:428– 433. Schulz H, Zulley J. The position of the final awakening wihin the ultradian REM/NREM sleep cycle. Sleep Res 1980;9:124. Schulz H, Massetani R, Fagioli I, Salzarulo P. Spontaneous awakening from sleep in infants. Electroenceph clin Neurophysiol 1985;61:267– 271. Snyder F. Toward an evolutionary theory of dreaming. Am J Psychiatry 1966;123:121–136. Vivaldi E, Ocampo A, Wyneken U, Roncagliolo M, Zapata A. Short-term homeostasis of active sleep and the architecture of sleep in the rat. J Neurophysiol 1994;72:1745–1755. Webb WB, Campbell SS. Awakenings and the return to sleep in an older population. Sleep 1980;3:41–46. Weitzman ED, Czeisler CA, Zimmermann JC, Ronda J. The timing of REM sleep and its relation to spontaneous awakening during temporal isolation in man. Sleep Res 1980;9:280. Wehr TA. In short photoperiods, human sleep is biphasic. J Sleep Res 1992;1:103–107. Wehr TA, Moul DE, Barbato G, Giesen HA, Seidel JA, Barker C, Bender C. Conservation of photoperiod-responsive mechanisms in humans. Am J Physiol 1993;265:R846–R857. Wehr TA. A ‘clock for all seasons’ in the human brain. In: Buijs RM, Kalsbeek A, Romihn JH, Pennartz CMA, Mirmiran M, editors. Progress in brain research, vol. III. Amsterdam: Elsevier, 1996. pp. 319–340. Wehr TA, Aeschbach D, Duncan WC. Evidence for a biological dawn and dusk in the human circadian time-keeping system. J Physiol (London) 2001;535(3):937–951. Wurts SW, Edgar DM. Circadian and homeostatic control of rapid eye movement (REM) sleep: promotion of REM tendency by the suprachiasmatic nucleus. J Neurosci 2000;20:4300–4310.