Citation
The effects of extended bed rest on human sleep patterns

Material Information

Title:
The effects of extended bed rest on human sleep patterns
Creator:
Campbell, Scott S ( Scott Searcy ), 1952-
Publication Date:
Language:
English
Physical Description:
viii, 60 leaves : 28 cm.

Subjects

Subjects / Keywords:
Biological rhythms ( jstor )
In rem ( jstor )
Periodicity ( jstor )
Philosophical psychology ( jstor )
Rapid eye movement sleep ( jstor )
Sleep ( jstor )
Sleep patterns ( jstor )
Sleep stages ( jstor )
Wakefulness ( jstor )
Young adults ( jstor )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Psychology thesis Ph. D ( lcsh )
Sleep ( lcsh )
Sleeping customs ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1981.
Bibliography:
Includes bibliographical references (leaves 55-59).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Scott S. Campbell.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
0028132575 ( ALEPH )
07904894 ( OCLC )

Downloads

This item has the following downloads:


Full Text











THE EFFECTS OF EXTENDED BED REST ON HUMAN SLEEP PATTERNS











BY

SCOTT S. CAMPBELL











A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY
OF FLORIDA IN PARTIAL FULFILLMENT OF' THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY











UNIVERSITY OF FLORIDA 1981











































Copyright 1981

by

Scott S. Campbell














ACKNOWLEDGEMENTS



The author would like to thank the members of his supervisory committee for their assistance and criticism during the preparation of this dissertation. The author would especially like to express his deep appreciation to Dr. Wilse B. Webb whose encouragement, guidance, time, and patience provided a constant source of motivation and inspiratiOnr Thanks likewise go to Briggs, Lewis, Alton, and Gregg whose technical assistance and friendship were invaluable. And final thanks to my parents for their support, both moral and financial, without which the completion of this dissertation would have been impossible.















TABLE OF CONTENTS


CHAPTER PAGE

ACKNOWLEDGEMENTS ................ ............ iii

ABSTRACT ........ ....................... ..vii

ONE INTRODUCTION ..... ................. l...

Sleep as a Biological Rhythm . ...... I
Uniphasic or Ultradian ? ..... .......... 4

TWO METHODS ........ ................. ..12

THREE RESULTS ......... .................... 15

Initial Sleep Period ............ .... 15
Bedrest Period ...................... ..17
Circadian Considerations ............ ...24

FOUR DISCUSSION ........ .................. 33

Total Sleep Time .................. 35
Uniphasic or Ultradian? ... .......... ..49

APPENDIX

SLEEP QUESTIONNAIRE .... .............. ...53

REFERENCES ........ ....................... ...55

BIOGRAPHICAL SKETCH ..... .................. ..60












iv















LIST OF TABLES


TABLE PAGE

I Comparisons of sleep stage characteristics for 1
baseline (Night 2), bedrest period, and normative group.

II Individual and group statistics for the bedrest 20
period (BP)

III Distribution of bedrest sleep periods by the 21
presence or absence of REM and Stage 4 sleep.

IV Comparison of sleep variables as a function of 25
night (11 PM 7 AM) and day (7 AM 11 11M).

V Sleep onset times of longest sleep episodes for 28
Night 2 (baseline), 3, and 4.

VI Comparison of sleep variables between first and 31
second 24 hours of bedrest period.























V













LIST OF FIGURES


FIGURE PAGE


1 Placement of sleep/waking episode across 60 18
hours for individuals and group.

2 Example of sleep across BP (#932). 19

3 Total number Of subjects asleep, for each hour 26
across the bedrest period.

4. Proportions of TST spent in REM sleep, as a 29
function of time of day.





























vi















Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


THE EFFECTS OF EXTENDED BED REST ON HUMAN SLEEP PATTERNS

By

Scott S. Campbell

August 1981

Chairman: Wilse B. Webb
Major Department: Department of Psychology

That the sleep response is one of numerous physiological functions which may be classified as a biological rhythm is well-established. Less well-founded, despite its popularity, is the notion that the

sleep of adult humans maintains an inherent tendency to occur monophasically within each 24 hours, typically during the nocturnal portion of the 24-hour day.

This dissertation addresses the hypothesis that self-imposed and/ or sociophysically mediated behavioral controls have a significant

influence on the manifestation of such a uniphasic sleep pattern,, and that attenuation of behavioral controls may result in a "loosening"s of the circadian framework of typical adult sleep.

Each of 9 young adults was electrographically recorded over a

period of 60 hours, during which behavior was restricted primarily -to lying quitely in bed; i.e.,' subjects were not permitted to read, write, vii








listen to music, watch TV, exercise, etc. Meals were served at random intervals, and subjects had no knowledge of time of day throughout the bedrest period (BP).

The data indicate that a reducton in behavioral controls, which typically influence the sleep process, results in substantial alterations in the placement and periodicity of sleep. Sleep episodes became

evenly dispersed across the 24-hour day, resulting in an approximate 6 hour sleep/waking cycle.

While the average sleep period duration decreased, mean total

sleep time for the bedrest period (28 h) exceeded, by more than 40%, that which would be expected under the assumption of a uniphasic sleep! waking rhythm. The continued tendency toward a nocturnal weighting of total sleep was clear, however, since 55% of TST for BP occurred during the night (11PM to 7AM).

Within sleep structure (e.g., sleep stage percentages, mode of

appearance) was affected only slightly by the looseningn" of the sleep response. REM sleep retained its tendency to occur differentially across the 24-hour day; stage 4 sleep appeared -to maintain its relationship to prior wakefulness.

It is suggested that the sleep/waking patterns observed in this

study may be further manifestations of an inborn (albeit, highly. dampened) ultradian sleep/waking rhythm.










viii














CHAPTER ONE
INTRODUCTION



Slee AsaBiological Rhythm. That the sleep response is one of a vast array of physiological functions which maintain a reliable tendency to repeat at regular intervals has been demonstrated on numerous occasions over the past two decades (see for example., Aschoff, 1965; Richter, 1965; Symposia on Quantitative Biology, 1960; Webb, 1971; Webb and Agnew, 1974a). That the modulation of the sleep/waking cycle is controlled by innate and autonomous mechanisms seems, also, to have been well established.

In a free-running (time free) environment, subjects display the continuation of a sleep/waking rhythm which deviates slightly from the standard 24 hours (usually longer). The only plausible explanation for such a finding, according to Aschoff (1965, p. 1427) is

"that this rhythm is not imposed on the organism by the environment, but is truly endogenous. We are dealing with a self-sustaining oscillation which is free-running under constant conditions and has

its own inherent frequency."

The autonomous nature of the sleep/waking rhythm may also be demonstrated in the free-running environment, relative to other

rhythms. The vast majority of rhythms which are in phase with the sleep/waking cycle in the 24-hour environment become desynchronized upon introduction into the time free environment (Aschoff, 1978; Aschoff and Wever, 1976; Wever, 1973). An example of such
1









desynchronization is that which occurs between the temperature and sleep/waking rhythms. When both rhythms are entrained to 24 hours,

the maximum temperature typically occurs just prior to retiring, and the minimum temperature occurs toward the end of the sleep period. However, under free-running conditions, maxima and minima of rectal temperatures become independent of sleep onset and termination.

In addition, the sleep/waking rhythm may be entrained to artificial light/dark regimens while typically in-phase rhythms maintain their free-running periodicities. For example, throughout 10 days of entrainment to a 3 hour sleep/waking cycle, subjects' temperatures continued to exhibit characteristic circadian periodicities (Weitzman, Nogeire, Perlow, Fukushima, Sassin, McGregor, Gallagher, and

Hellman, 1974). Such data "contradict any hypothesis of a simple 'cause--and-effect' relationship between the activity-rest cycle and the rhythm of body temperature, and suggest a system of two coupled oscillations . with a mutual phase relationship which depends on the conditions" (Aschoff, 1978; p. 740).

Further evidence of the sleep/waking cycle's autonomy relative to other biological rhythms may be seen in the results of sleep deprivation studies. In the absence of sleep (for more than 70 hours) measures of performance, temperature and endocrine function all showed continuation of endogenous ciradian rhythms (Murry, Williams, and Lubin, 1958; Froberg, Karlsson, Levi, and Lidberg, 1975, as cited by Aschoff, 1978).

Examinations of the maturation of the human sleep/waking rhythm have also provided evidence for its innate and autonomous nature.





3



After observing the development of an initial 25-hour periodicity in an infant maintained on a self-demand sleep-wakefulness and feeding schedule, Kleitman and Engelmann (1953) concluded that "there may be a 'natural' rhythm which only has to be adjusted to the 24 hour alternation of night and day" (p. 280). Other researchers have also observed the development of adult-like, primarily monophasic, nocturnally placed sleep under similar conditions of self-demand feeding (Parmelee, Wenner, and Schulz, 1964; Salzarulo, Fagioli,

Salomon, Ricour, Raimbault, Ambrosi, Cicchi, Duhamel, and Rigoard, 1900).

While the sleep/waking rhythm clearly appears to be inborn and independently regulated, it may, at the same time be entrained by exogenous Zeitgebers, to frequencies other than those characteristic of the free-running environment. It has been demonstrated that

human subjects are capable of maintaining sleep/waking cycles of various lengths, quite efficiently within a range of about 20 to 27 hours, as the result of entrainment to appropriate light/darkregimens. It has been further shown that the sleep/waking rhythm characteristic of a free-running environment may be regulated by "rigid control of

the sleep portion of the cycle particularly of the wakeup time" (Webb and Agnew, 1974b; p. 701).

Several investigators also point to the developmental literature

as a source of evidence for the entrainability of the sleep/waking rhythm (Ellingson, 1975; Gesell and Armatruda, 1945; Kleitman, 1963; Sander, Stechler, Burns, and Julia, 1970; Sterman, 1979). As expressed by Sterman (p. 213), "the development of sleep is a postnatal





4



phenomenon, subject both to the dictates of physiological maturation and the influences of environment. Such factors as geography and social custom entrain the physiological substrates of the sleep -waking cycle to determine the behavioral patterns which will come to characterize these states in a given culture."

Indeed, the very fact that adult humans maintain a 24-hour

sleep/waking periodicity, in spite of an inherent rhythm with a frequency longe' than 24 hours, is clear evidence of the entrainability of the rhythm.

It appears then, that the oscillation between sleep and wakefulness, characteristic of adult human subjects, meets the criteria necessary for its designation as a biological rhythm. Sleep is an inborn, independently regulated, exogenously entrainable, regularly occurring response.

Up phasic or Ultradian? While sleep and wakefulness clearly

comprise a physiological system which may be classified as a biological rhythm, determination of the rhythm's inherent periodicity remains less clear, since few studies have been conducted which adequately address this point. It is generally assumed that the sleep/waking cycle is anuniphasic one. In addition to the obvious support for such a notion, i.e., that adult humans typically obtain sleep in a single, nocturnal block, results from studies in which the normal sleep period has been dispersed or eliminated have also been cited as evidence for this view. In such experiments, performance measures are usually lowest and subjective reports of sleepiness are





5


usually greatest during the interval in which the subject would normal ly be sl eepi ng (Coiquhoun, 1971; Wei tzman et al. 1974).

It has further been reported (Weitziman et al., 1974) that "a

striking persistence of a circadian pattern of total sleep time" (p. 1020) was exhibited throughout 10 days of entrainment to a 3-hour day. In this study, subjects were encouraged to sleep 1 hour out of each 3 hours. The authors reported that the 4 sleep periods beginning at 3 AM, 6 AM,. 9 AM, and noonaccounted for 70% of the total sleep time.

The decreased sleep efficiency associated with dispersed and

displaced sleep schedules has also been seen as lending support to

the notion of an inborn, nocturnally placed, uniphasic system of sleep and wakefulness. Studies in which the normal, nocturnal sleep period has been shifted to a time typically employed for work or other waking activities, indicate that such sleep episodes are characterized by significantly greater proportions of wakefulness (stage 0) after initial sleep onset, as compared to baseline (normal) sleep (Webb, Agnew, and Williams, 1971). Such findings suggest that the maintenance of sleep is more difficult when the sleep period is shifted from its typical, nocturnal placement. Further, as experimentally circumscribed 'days' deviate from the normal 24 hours, sleep efficiency (amount of time in bed spent asleep), becomes less. Webb and Agnew (1975) found that Subjects maintaining 'a sleep/waking schedule of 3 hours sleep/6 hours wakefulness only achieved 80% sleep efficiency when compared to baseline conditions (8 hours sleep! 16 hours wakefulness). Weitzman et al. (1974) reported,56% sleep





6




efficiency for subjects attempting a 3-hour cycle compared with 90% efficiency for the baseline condition.

It has also been noted in dispersed, as well as displaced sleep, that the intrasleep structure of the "new" episodes differs in several details from that seen in normally occurring nocturnal sleep. For

example, Weitzman et al. (1974) observed "early" REM periods (preceded by less than 10 minutes stage 2 at sleep onset) and "sleep onset" REM periods (preceded only by wakefulness or stage 1) in over half of all REM episodes recorded over the 10 days of 3-hour sleep/wake scheduling. Results of the displacement of subjects' typically nocturnal sleep into day sleep include significant increases in the amount of stage I and reductions in the amounts of stages 3 and 4 (Webb and Agnew, 1978). Such findings lend further support to the view that these sleep episodes differ substantially from endogenously modulated sleep periods.

It may be argued, however, that studies of sleep dispersion and and displacement do not adequately address the question of the inherent frequency of the sleep/waking cycle. Rather, such studies more specifically test the entrainability of the presumed uniphasic rhythm. Failure to effectively impose a shortened regimen upon a hypothesized uniphasic rhythm does not necessarily rule out the existence of an ultradian sleep/waking rhythm. Yet, data supporting such a possibility are not extensive.

One type of evidence used to support the suggestion that the

human sleep/waking rhythm is inherently ultradian is provided by the analysis of sleep patterns of individuals participating in Polar expeditions. Lewis (1961; Lewis and Masterton, 1957) has suggested





7



that given a situation in which adherence to customary circadian scheduling is less stringently controlled, a more accurate reflection of the inborn sleep/waking rhythm may emerge.

In the Polar regions the photoperiod is not diurnal, but seasonal. Further, expeditions into these regions typically maintain little resemblance to an organized community, in that there are few circumscribed daily routines, no specific feeding schedules, few timetables (Lewis, 1961). Sleep records from 5 Polar campaigns suggest that the organization of sleep/waking patterns was a direct function of the stringency with which the "community" was structured. For example, members of the British North Greenland Expedition (1952-54) were allowed to sleep as long as they liked, at virtually any time they had the inclination to do so. Under such conditions, sleep durations per 24 hours did not increase substantially over baseline amounts. Yet, the length and placement of individual sleep episodes showed considerable deviations from members' normal sleep habits recorded in England. The tendency toward polyphasic sleep placement was most extreme during the three months of continuous light or darkness. "During these periods .. men used any of the 24 clock-hours for sleep, The winter months were characterized by many interruptions to sleep and by the taking of naps" (Lewis, 1961, p. 325). Although no data were presented relative to the rhythmicity of these multiple sleep episodes, it appears that the uniphasic "nature" of the sleep

process loosened somewhat under conditions of limited environmental cues and negligible behavioral and cultural restraints.





8



Within the more structured hours of daily life, it has been

suggested that the ultradian nature of sleep and wakefulness may be manifest in the form of naps (Webb, 1978). Webb's proposal that naps are part of an inborn sleep/waking rhythm is based on the contention that certain characteristics of these intermittant sleep episodes

match features which are common to biological rhythms in general. Specifically, it is maintained that naps are: 1) temporally, repetitive, 2) species specific, 3) often developmental, 4) innate and unlearned, 5) endogenous, and 6) adaptive. While evidence seems to support the first four characteristics, and only post hoc speculation can address the sixth, Webb points out that experimental support for the claim that naps are endogenously mediated is extremely limited, and that further data are "likely to be slow forthcoming." This is due primarily to the fact that virtually all procedures utilized for the examination of presumed biological rhythms have prohibited other-than-circumscribed sleep episodes as a consequence of experimental design considerations. Such experiments typically focus on the interactions of major components of sleep, waking and performance. As such, "the very nature of naps in these designs are uncontrolled and interfering sources of experimental variance" (Webb, 1978 p. 316). Two primary designs of this type are the free-running and bedrest procedures.

It has been noted that the sleep/waking rhythms maintained by subjects in free-running environments provide excellent support for the notion that such rhythms are innately and independently controlled. Under such conditions, time cues are removed from the




9



environment, but other behavioral controls are typically left Unaltered. That is, subjects are askIed to lead a "normal" life, including eating three meals a day in normal sequence, having the opportunity for exercise and other activities such as reading, writing, watching TV, etc., arid maintaining circumscribed sleep schedules (no naps). As a result, these subjects have a tendency to carry on day-to-day life in a manner to which they have grown accustomed over the years. Clearly, such circumstances may strongly bias behavior toward a circadian schedule, perhaps entraining an inherently ultradian rhythm -to an environmentally and culturally modul ated ci rcadilain frainewoirk. These factors notwi thstandi ng, the sleep patterns displayed by subjects in time -Free environments have been seen as providing evidence not only for the biological basis of the sleep/waking rhythm, but for -its circadian (uniphasic) nature, as well.

The latter inference seems tenuous, at best. For', when controls on -the number and placement of sleep episodes are reduced in freerunning environments, evidence seems to indicate that the "normal sleep pattern" (single sleep period of from 7 to 10 hours with occasional napping) is modified. Webb and Agnew ('1974a) reported that under such circumstances "a significant increase-in both long and short periods of sleep, compared to -the 24-hour environment" (p. 620) resulted. While neither the placement nor the periodicity of the increased episodes was analyzed, the authors concluded that such changes in the sleep response showed the extent to which the uniphasic sleep pattern is'disrupted by reduced controls on sleep p1 acement.





10


While studies employing a bedrest design may clearly eliminate some behavioral controls left in tact by the free-running design (most notably physical exercise), this procedure is also typically

characterized by experimentally circumscribed sleep periods as a consequence of full schedules of psychological, psychophysiological and performance testing (see for example, Ryback & Lewis, 1971; Ryback, Lewis & Lessard, 1971).

One recently published study (Nakagawa, 1980) however, does

provide some insight into sleep/waking cycles in the absence of cultural mediation and behavioral controls. In this study, subjects were required to remain lying quietly in bed for 10 to 12 hours immediately following a full night's sleep. The original aim of the study was to "elucidate changes in daytime states of consciousness" by observing "the EEG patterns of (awake) subjects in restrained postures under minimally changing experimental conditions" (p. 524).

Despite explicit instructions to remain awake during the 10 to 12 hour period, howeVer, Subjects experienced "an uncontrollable

desire to fall asleep," resulting in sleep/waking cycles of approximately 4 hours across the experimental period. Again, it appears that the uniphasic placement of the sleep process loosened in response to a reduction in behavioral controls.

In summary, most designs employed in the investigations of

the human sleep/waking rhythm are inadequate for that purpose since they include strict controls over waking and sleeping as a necessary part of the procedure. However, results from studies in which such controls are relaxed indicate that subjects have a tendency to








relinquish their typical uniphasic sleep placement in favor of a more dispersed, polyphasic sleep system.

The present study attempted to address the hypothesis that

behavioral, cultural and environmental factors serve to entrain an inherently ultradian sleep/waking rhythm to a uniphasic, 24-hour framework by imposing upon the organism a series of behavioral controls. In the experimental environment, such controls are frequently mimicked by circumscribed sleep/waking schedules, feeding regimens and testing procedures. These external controls "may create incompatible responses which suppress the expression of (the) rhythmic response or create such a 'noise' background as to not permit measurement" of the hypothesized ultradian rhythm (Webb, 1978 p. 315). Therefore, byminimizing "incompatible responses" and "noise background" the present study sought to "emancipate" the hypothesized ultradian rhythm from control by a self-imposed and/or sociophysically mediated circadian framework. By so doing, it was proposed that the presence of such a rhythm may be more readily observed and measured. Specifically, subjects were required to maintain relatively static, basal levels of behavior over a period of almost three days, during which time their sleep/waking behavior was the primary

variable of interest.














CHAPTER TWO
ME TH0ODS



*rhe subjects for this study were 10 young adults (5 males and 5 females) ranging in age from 18 to 25 years (mean = 20.4). Subjects were selected after questionnaire (Appendix 1) responses revealed normal sleep patterns and the absence of chronic health problems and/or acute illness.

Each subject slept for 2 nights in the laboratory immediately prior to the beginning of 60 consecutive hours of bedrest. Between

the 2 nights in the laboratory subjects were permitted to maintain their normal daily schedules with the stipulation that no naps be taken and no drugs (including alcohol and caffeine) be ingested during that period.

For each session, 2 subjects reported to the laboratory approximately 1 hours prior to their normal bedtimes (about 11 PM) for electrode placement. Five millimeter gold disc electordes were applied over Cambridge electrode jelly and held in place by coilodian saturated gauze strips. EEG and electrooculogram (EOG) were continuously monitored for each subject using a Grass Model VI electrophysiograph. The EEG of each subject was recorded on three channels between electrode sites Fpl F7, P1 T5, and C3 A2. The EG was monitored on a separate channel by electrodes placed on the external canthus of each eye.

12









At the completion of electrode placement each subject retired to a private, electrically shielded, sound attenuated, temperature controlled room. Each subject's record was scored in one-minute epochs, following the methods described by Agnew and Webb (1972a). Cross-scoring of random 2-3 hour segments of the sleep portions of each record revealed a between-scorer reliability of greater than 90%.

Sleep onset was determined by the appearance of the first spindle or K-complex (Agnew & Webb, 1972b). A sleep period was considered as Such when sleep continued, uninterrupted by more than

20 minutes of wakefulness, ,or at least 30 minutes. A waking period was defined as Stage 0 EEG for at least 20 minutes.

Timing of the 60 hours of bedrest began with the spontaneous awakening of each subject from his/her second night of laboratory sleep and the subject's verbal report that he or she felt rested and had had enough sleep. At this time subjects were given an opportunity to eat breakfast and use the bathroom. No intentional time cues were given the subjects upon awakening nor at any other time 'throughout the bedrest period, with the exception that each subject was told when hie or she was "past the halfway point" at a random time after 30 hours had past. However, uncontrollable, general time cues from exogenous sources (eg. nearby building construction, general building noise during working hours, homecoming festivities during one session subjectss 945 & 946)) were present in varying degrees from session to session.

During the 60 hours of bedrest, subjects were allowed minimal

exogenous stimulation. Activities such as reading, writing, listening






14


to music or watching television were prohibited. Conversationsduring the period were limited to brief dialogues with the experirienter at meal -times. No instructions were given relative to when or when not to sleep. Subjects were asked to lie as quietly as possible, but were permitted to change positions in bed when inclined to do so. Closed circuit TV monitors were employed to insure adherence to the experimental instructions. Subjects were permitted to go to the bathroom (in room) at their convenience and were allowed to sit up iii bed during meals.

Meals were served unsystematically (always during an ongoing waking period, as determined by the EEG) and consisted oi a choice of several types of sandwiches, vegetable dishes, desserts and noncaffeinated beverages. A maximum of I- hour was permitted for the completion of each meal. Several types of fruit and a pitcher of ice water were also available to each subject ad libitum.

Illumination in each room was provided by a 60 watt lamp placed on a table by the subject's bed, and control of light and control of light and darkness was at the discretion of each subject.













CHAPTER THREE
RESULTS

One of tChe 10 subjects terminated participation prior to the end of the bedrest period. Since only 12 hours of bedrest were re.corded for this subject, the results presented here are those of the

9 subjects who completed the session.

The criterion used here to define a "sleep period" (.i.e., 30 minutes of sleep, uninterrupted, by more than 20 minutes of wakefulness) accounted for 91.3% of all sleep recorded. Eight sleep episodes had durations of less than 30 minutes (mean = 7.25 min., range =2 to 18 mn.), and therefore were not considered "sleep periods" in the following analyses.

Initial SieeD Period. The mean total sleep time (TST) for the sleep period immediately preceding the onset of the bedrest period (Night 2), was 8.72 hours, with a range of from 5.82 hours to 10.43 hours. Sleep stage percentages for the period were generally within normal limits. However, as shown in Table 1, there was a significantly greater proportion of Stage 0, t (df =39) =2,32, p.<05 (all

levels of significance are based on a two-tailed test), and a smaller percentage of stage 1, t(39) = 5.04, p.,001, and stage 3, t(39) =2.20, p.<05, in this sample as compar-ed with a normative

group of 16 males and 16 females of approximately the same age (mean= 24 years) previously recorded in the same laboratory (Williams, Agnew

and Webb, 1964; 1966). "Chopping" each record to an artifical maximum 15








16




w C'Q C\J LO LO I-- Izjm Ln t m
m C"i m cl cl; C C
C\J C\J C\J m C)
V) C) C) C


CL CL 0CD C) cli co C\i co ko + + +
1 1 r- 7 9 C L + +
F- +
0 V) tzd- LO -j- LO M LO m t I

cd Gi
>
+ 4-5
+ ro
+ +
(\I I'D rl_ Ln t-- m m ::I,
-0 CD m C"i m co Irr co m 0
(D I- 2n
-0 V) r 1 If; L
ZA E
0


4-)
+ a)
+ CC"i + co
I-- Oll vo LO (D C\i r-- 1-d- 4- LO
C-r CY) C) LO CY) CD LO I:zr m cri
tu
C/) co r_ C Lf; C ki
W Lr) Lf) Lr)
LO
V) CF)
ai

S- + +
+ + co
+ + + 4--) 10
C\j I:zj- Cl) C\d co Zt 00 r0 4-)
C r- C C9
V) LO c". C'O
4-3
FE:
4S ro

4-)
u
ro
ro + >)
CD + +
L) M Ln -d- 00 (7) lz:t CY) 00
m CV) LO Q0 0) LC) Ln C\j a)
>
IM V) C\; C 0 C C cl (1)

4--)
L) V)
ro
L)

C\j C:) 00 m S4- :3 Lr) Ln kJo 00 t-- LO m cy") 40 0 F- (zs M cn
r C6 r- L a) V) -Y
C\J

0 (U
>



E
0 0
u 9-CD
>
C\j 0- Cld 4--)
4--) CY) CL Ln -a
LLJ ro 11 4.) 0 +-) a) 0
__j F. = -a -C _g_ Sco t- o) u U) -0
ct o (3)
F-





17



of 8 hours for Night 2 (Jte. scoring no more than the first, 460 minutes of each record) did not significantly change the mean proportions of sleep stages as compared to 'intact' flight 2, but such a procedure did reduce the proportions of stages 0 and 3 so that they were not significantly different relative to the normative group.

.Bedrest Period. Although timing of the bedrest period (BP) began with the spontaneous awakening -from flight 2 and the subject's acknowledgement that (s)he felt awake and rested, later analysis of the records of three subjects indicated -that, the 'spontaneous awakenings' were less than 20 minutes in length (7, 9, and 11 minutes) and were followed by substantial sleep episodes meann = 2.11 hours). These episodes, therefore, were included in -the initial sleep period (Night 2), and as a result, the bedrest periods of the 3 subjects were somewhat under 60 hours (mean = 55.32 hours). The mean duration of BP for the entire group was 58.8 hours (range =53.1 to 62.2 hours).

Referring to Table I it can be seen that the sleep stage percentages averaged from sleep periods occurring during BP differed only slightly from those recorded during Night 2. There were nonsignificant increases in the proportions of stages 0, 1, and 2, while nonsignificant decreases in SWS and REM sleep were observed as compared to Night 2. Similar trends (all significant, however) were

observed relative to the normative group, with the exception of stage 1, which comprised a smaller percentage of TST across BP than in -the normative sample.

The distribution of sleep and wakefulness across BP, for all nine subjects, is presented in Figure 1, Figure 2 presents,, i~n










18





Le)





ra


























O)




LO
LLJ


LLI cc
LL
0 (Tj
(A

0










0 cu (Do
(3)
CL ci
(3) 4-
0 V) C\j
"P -P
9= a) C)

Qj








Cli




co C W n 0 E LLrn z

S103rens
O 638AnN 13eNnN 103rens









19







4)




CD







(A
r

4
0 (A



(U







C\i








Ln co
< Ul)
















(A 3: 0 Ln

0 QJ X SLLJ ro


z C"i C)



LL.









LO C"i (Y) C\j co C-)
C L 0) M M (.0 --1- 0')
LU
fa: .;I. Lo C k C C C
C\j C\j C\j cli C\j C\j



00 (Y) C) Q-) m QD t
UO CY) CO tD ( o t.0 co 00 00 Lr)
V) cl; C C C C C C C 0
C"i



cy') co Oo (.0 ( o C ') r Oo co zt
00 m C) LO m rl- C\i m CY) C\i
cl; k o;



00
C\j C\j LO 00 Ol 00 m CY)
m r^l OD LO C\j C\i C) C\i I:zj- m
V) C cl; cl C C C; 0 C6 4
E.0 k.0 kr) LO LO LO t-0 LO LO




tD) o co C:) m Lr) C) co r
Q) 4--) Cl) <- C) r- (.0 r- CD Cllj co
lx cl; C L 1; cy; C



C)
-P m CY) I-- C\i cv) r- ( o I-- -zi- C) m 00
4--) CD P-- Ln (.0 C\j Lf) CQ
V)
cl cl; C\




4--5 CU 4-3
Ln h: m
4-) (11 C:) d- m qd- 00 t m C) co
m t-- co 00 C) co C) CY) LO ko
-P CL
V) (1) C) L C




0) m C\j C o co co co C) LO 0 C



0

73 CD 4J -C
_0 1,- 1-1 kD (Y) CY) LO t o co 00 m LO
co -j- Ln 0) 0) m CY)
> 0- (3)
CL) -i C C- C C\ C (Y; C cl cl




S:
CD CD co C) m
V) 1 C C : C C C
F- r m CY) C\j m Oo LO t.0 qzi- r-- Ls')
LU CY) C\j m m C\i C\j (\j
cn

4--)
0 C\j co -,I- r-_ I-- r-I 00 C) LO C)
C\i m LO CD 0') 00
(1) (1) C C C C C cy; r 06 C
Do D- kD t D ko k-D ko t.0 LO LO LO LO





4-3
0 co m C:) W) .a ro
(L) C\j C\j co m m Izi- -zj- 42) V)
0') CY) o) 0") m 0') 0*) CF)







21







E EE
C) C) cx CD clM F- C)

G) C) C'
E
V) CD
m CD 0- U-) CL m
C) C\j CD CD Lr)
u CD C; :: C ;
(; IR rLU
F 'o F E
cu (D CL CD n- CD
C) (Y) C) C\] (Z)
C)
r
L)
Q) E co F 0- F F F
co C:) Sn L.0 ro CL r-L Ln
C) C) C) tc:j- C) C-) LO C)
ro CY) C) LO
C\j
Iz C) C ; rc-D C r-o






CL
4-)- 0 1.0 V) S- Ln C
ro C) S- LO S- I-- S- E
4- oi
-0 ro ITO co (.0 C\j Lil)
S- t-- CD CD co
C -P r- (D --o 0
W 4--) 4-) -P
C) r-I
00 Ln lzt Co
co CY)






4->
V) CY) r- r- CD
C\j
Z,




4
0 >1 00
C)

4-1


S
4- m 0)
4-) 41- LO
0
fn Fuj
LLJ
-j
S
F



1 <





22



detail, the sequencing of sleep and wakefulness of one subject (932) across the bedrest period. Individual and group statistics for the entire bedrest period are presented in Table II. BP was characterized by an alternation of waking periods with an average length of 2.7 hours (range =.4 to 10.2 hours) and 'Sleep periods with a mean duration of 2.99 hours (range =.5 to 11.6 hours). Subjects spent almost half of BP asleep (47.6%) with a mean total sleep time of 28 hours (range =20 to 37.3 hours), Sleep was dispersed across an average of 9.3 episodes during the bedrest period (.range = S to 13).

These results are in contrast to what would be expected under the assumption of a uniphasic model of sleep. Under such an aSSuffip tiori, one would predict 2 discrete sleep episodes across the 60 hours, placed within the normal nocturnal periods (approximately 11 PM to

7 AM), with a mean TST for the bedrest period of about 16 hours (about 27% of BP).

Sleep periods during BP could be divided into four general

categories based on the presence or absence of stage 4 and REM sleep, as shown in Table TIT. Seventy percent (59) of all sleep periods recorded over the bedrest period were 'full blown' sleep episodes, in that all stages of sleep were observed. Of the remaining 25 sleep periods, 3 (all day episodes) showed neither REM nor stage 4 sleep. Nine sleep periods (8 days, 1 night) exhibited an absence of REM sleep and 13.(12 days, 1 night) showed no stage 4 sleep.

The mean period length of those episodes with neither REM nor stage 4 was 61 minutes. The mean duration of 'noREM' sleep periods was 65 minutes, and the 'noStg4' sleep episodes averaged 84 minutes





23



duration. This is compared to a mean duration of 225 minutes (.3.75 h) 'for those sleep episodes which contained both stage 4 and REM.

The occurrence of these events in terms of sidereal time was also examined, The three periods containing no REM and no stage 4 occurred after noon and before the onset of the night phase (11 PM/). All but 2 of the 'noREM' periods occurred between 2 PM and 10 PM. The 'noStg4' episodes occurred across the entire day phase (7 AM to 11 PM) with only one such episode occurring at 'night' (6:40 AM).

Regarding the mode of appearance of REM sleep episodes, of 153 REM periods recorded over BP, only 3 were characterized by onsets at: the beginning of a sleep period. Sleep onset REM periods (SOREMPS) were defined as those REM episodes occurring at the onset of sleep which were not preceded by more than 1 minute of stage 2. Two sleep episodes containing SOREMPS also were characterized by the absence of stage 4 sleep. The SOREMPS were contributed by 3 subjects, had a mean duration of 15.3 minutes (9, 10, and 27 minutes), and occurred at approximately 6:30 AM, 6:30 PM, and 1:30 PM, respectively.

In an effort to further elucidate the factors involved in the cycling of REM sleep, two examinations were made. In the first procedure, the REM sleep cycle (defined as the interval between the onset of one REM period and the onset of the next) was calculated for each subject's baseline night (Night 2). The onset points of the cycle were then superimposed across each subject's entire bedrest period (including waking episodes) -to test the hypothesis that the REM cycle was independent of pri.or sleep time. Also, the cycle was





24



superimposed across each subject's compressed bedrest period, in which waking intervals were removed to obtain a continuous sleep period, to test the possible sleep dependency of the REM sleep cycle. The initial onset of the superimposed cycle corresponded to the onset of the last REM sleep period of Night 2.

The relationship between the superimposed REM cycle and the occurrence of actual REM episodes was examined by calculating the mean difference between each onset point of the hypothesized cycle and W, closest REM epoch within that sleep period. The average difference between each onset of the superimposed cycle and the appearance of an actual REM epoch, across the intact bedrest period, was 15.8 minutes CSD = 3.7 min). The mean difference between the values for the 'compressed' BP was 14.5 minutes (SD 4.7 min). The difference between the two conditions was not significant.

To further test the sleep dependency of the cycle, the mean REM cycle length of the 'compressed" BP, for the group, was determined. Again, beginning the cycle with the onset of the last REM period of Night 2, mean cycle length for the 'compr essed' BP was 96.84 minutes WD = 10.42 min). This is compared with a-mean REM cycle length of 96,56 minutes (SD = 12.76 min) for Night 2.

Circadian Considerations. There were considerable differences in several measures as a function of time of day. For comparative purposes, waking and sleep periods were considered daytime events if they were initiated between 7 AM and 11 PM. "Nighttime events were defined as those episodes with onsets between 11 PM and 7 AM.





25


TARLE IV, Comparison of sleep variables as a function of night
(11 PM 7 AM) and day (7 AM 11 PM).



Overall Night Day

Mean TST 27,99 h 15,34 h 12,65 h
(5.39) (3.30) (4.91)

Mean Sleep 2.99 h 5.52 h 1.93 h
Period Length (2.35) (2.65) (.98)

Total # Sleep 84 25 59
Episodes

Sleep/Wake 6.10 h 7.81 h 5.25 h
Cycle Length (2.68) (3.26) (1.74)

% Stage 0 3.59 3.50 3.64
(2.28) (2.26) (3.58)

% Stage 3.28 3.88 2.81
(1.87) (2-79) (2.06)

% Stage 2 58,41 56.37 60.80
(4.94) (5.76) (6.22)

% Stage 3 4.33 4.13 4.94
(1.24) (1.58) (1.7S)

% Stage 4 9.86 8.89 11.01
(4.54) (5.37) (4.28)

% Stage REM 20.47 23.13 17.54
(2.94) (4.01) (3.87)


Mean p <.001 t test, two-tailed
(SD) p < .01







26



co












-po



09
QI)




C)
iS
o



C
CC)



(0



V) (41)
CL (4- )
Co )









00















L 0(0
LO4 CIO





27



Of 84. total sleep episodes recorded for the group across the bedrest period, 59 (70.2%) occurred during the day and 25 (29.8%) occurred at night, However, over half of the total sleep time (54.7%) during BP was obtained during episodes initiated between 11 PM and 7 AM.

Comparisons between day sleep periods (DS) and night sleep periods (NS) of' temporal measures and sleep stage percentages are shown in Table IV, The most notable difference in measures of DS and NS was with regard to mean sleep period duration. NS episodes continued for a mean duration 5,52 hours (SD = 2.65 h) compared to an average length of 1.93 hours (SD = .98 h) for sleep episodes initiated during the day (t (82) = 6.58, p<.001). Conversely, the mean duration of waking episodes was 1.47 hours (SD = 1.04 h) for nighttime events and 3.12 hours (SD = 2.4 h) for those initiated during the day (t (82) = 4.29, p4.001). Consequently, the sleep/ waking cycle length (defined as the interval between the onset of one sleep episode and the onset of the next) increased significantly from day (5.25 h, SO = 1.74) to night (7.81 h, SD = 3,26) t (73)

2.89, P<.01.

Figure 3 presents, by hours, the number of subjects sleeping. across the bedrest period. The group tendency toward nocturnal sleep placement is evident. Of 144 total subject-hours available for sleep during the 2 night periods (11 PM to 7'AM), 104 (72.31/0) were used for that purpose. In contrast, of the 279"subject hours comprising the first and second day periods (.8 AM to 11 PM.and 7 AM to 11 PM) less than half 05.2%). were spent sleeping.







28




U) r- co







co 0-) :::k

m 0) C\J




CI) Ln U')

C\j
V



V)
LO co C)
CD L.0 (11)
o') C\j Ilj




0) rl_ C\i 00
CD LO -f C\i
C\i
0
(4-
(A
Gi C C"i m
..0 tl- LO tzl- CY)
0
C\J

Gj

ai CD LO
G) to -at LO
m






0 C) -:ir CQ Lo cy') C)
(4-- C;





4-J CY) C\j CY) Lr)

CY) C\j




Cl- -:2c
co CD r"
C) co : I- m CD
L/) m





Lli 4--)
.co








29




















C\i
a) u
< 4--) CO
zi- co
E
.- S
E 0
Ll
4
t-0 co
C 0 (D
0
C\l 4- QJ U)
C\j < u
4
0


-Cl 4-)
>
00


4
Cl- C\j a
co -1 :E IV
uj a) 4--)


CL

4-)

G-1

Z: 2E 0- m ,-,I- co


4- ro 40 4--) L) c C- ro (n (Ii Q)
(-) 0) _z S- U
(1) V) 0
4-3 0S
o LLJ

>) E s- ro 14- .CL. -0 0 4-3


C)

C C\j
cr)
<
co



LO C:) C:) m
C\j


Wl INIMId





30



The possible presence of a free-running sleep/wake rhythm was,

examined by comparing the mean sleep onset time of Night 2 (baseline) and the mean onset times of the longest sleep periods of Nights 3 and 4. Thle mean sleep onset times for the group exhibited a tendency to, occur later on each succeeding night--slightly under 1 hour later, on Night 3 (as compared to baseline) and 42 mliffutes later onl Night

4 (as compared to Night 3). However, as shown in Table V only 2 subjects (929 and 937) demonstrated consistent free-running patterns across the three nights. Four additional subjects delayed the onset of' their major nocturnal sleep onset on Night 4 (relative to Night 3).

Sleep stage percentages were not affected substantially as a function of time of day, While there was slightly less stage 1 (21.8 vs 3.9%) and slightly more stage 2 (60.8 vs 56..4%) and stage 4 (11 vs 8.9%) during the daytime sleep periods, only the difference in REM sleep percent was significant, with REM sleep comprising just over 23% of NS episodes and only 17.5% of DS periods (t (82) =5.91, p<.001). The difference in REM percent appeared to be a consequence not only of the difference in sleep period length but also of the maintenance of a circadian pattern of REVS across the the bedrest period, as shown in Figure 4. The largest proportion of 1ST spent in REM sleep occurred during those sleep episodes initiated between 4 AM and 8 AM. The proportion of REM sleep decreased across thle-next three 4-hour time blocks until reaching a low of 11.57% of TST for sleep episodes initiated between 4 PM and 8 PM. The percentage of REM sleep then exhibited a steady increase across the subsequent 4-hour time blocks (8 PM to 12 MID and 12 M to 4 AM).





31.


TABLE V!, -Comparison of sleep variables, between first andI second 24
hours of badrent period.,




Overall 1st 24 [IRS 2nd 24 [IRS


Mean TST 27,99 h 14.46 h ** 10.97 h
(5.39) (2.28) (2,36)

Mean Sleep 2.99 h 3.54 2.82 h
Period Length (2.35) (1.92) (1.44)

Total # Sleep 84 39 33
Episodes

Sleep/Wake 6.10 h 5.19 h 5,64 h
Cycle Length (2.68) (220) (2.04)

% Stage 0 3.59 2.94 4.48
(2.28) (2.23) (4.25)

% Stage 1 3.28 3.03 3.74
(1,87) (1,93) (2.66)

% Stage 2 58.41 .59.69 56.68
(4.94) (4.03) (8.50)

% Stage 3 4.33 4.15 4.66
(1.24) (1.40) (1.58)

% Stage 4 9.86 8.39 11.39
(4.54) (3.60) (6.88)

% Stage REM 20.47 21.77 19.02
(2.94) (2.56) (4.82)


Mean **p< .01
(SD) t test, two-tailed





32


In addition to day/night differences, there was alsho disparity in temporal measures as a function of elapsed time within the bedrest period. Table VI compares rsT, sleep period duration, sleep/waking cycle length, and sleep stage percentages recorded during the first 24 hours of bedrest with those recorded during the second 24 hours. Total sleep time 'for the first 24, hours (14.5 h) was significantly longer than TST for the second 24 hours (10.9 h) (t (16) =3.19, p< .01). While the decrease in TST was,, in part attributabl e to a reduction in the total number of sleep episodes during the second 24 hours (39 vs 33), a decline in the mean sleep period length across BP (3.54 h vs 2.82 h) also contributed to the difference.

Despite the decrease in overall sleep period duration, the

sleep/waking cycle length remained essentially unchanged throughout the bedrest period. For purposes of comparison, only those sleep/ waking cycles initiated and completed within one or the other 24 hour periods were considered. The mean duration of sleep/waking cycles occurring during the first 24 hours was 5.2 hours (SD = 2.1 h), while the mean cycle length of those completed during the second 24 hours was 5.6 hours ('SD =2.4 h).

There were no significant differences in sleep stage proportions between the 24-hour intervals. However, trends were observed relative to stages 4 and REM. Stage 4 increased (from 8.4. to 11.4%) from the first 24 hours to the next; this despite the fact that there were more sleep episodes containing stage 4 in the first 24 hours (32 vs 27). Conversely, REM sleep decreased by over 10% from the first to the second 24 hours..















CHAPTER FOUR
DISCUSSION~

The results of this study substantiate and further detail previous reports that a reduction in the control of the sleep response results in a 'loosening' of the uniphasic, nocturnally placed sleep pattern typically observed in adult humans. At t"he same time, however, the results underscore the tendency of sleep to be differentially weighted toward a definite and persistent nocturnal placem,.ent.

On the one hand, the data suggest a quadriphasic sleep pattern, with the onset of sleep episodes occurring at approximately 6 hour intervals throughout the bedrest period. While such a rhythmicity proved to be highly variable within subjects, from -one sleep period to the next, the relatively small between-subjects variability emphasizes the regularity of such a rhythm within the group. That the number of sleep periods was proportionately placed, relative to day and night, further accentuates the uniform distribution of sleep episodes across the 24-hour day. The 'day' phase (7 AM to 11 PM) comprised approximately 73% of the total bedrest period. Correspondingly, 70% (59) of the total number of sleep episodes recorded- across BP were initiated during the day.

On the other hand, comparison of the mean duration of those day sleep periods with the average length of sleep periods initiated at night clearly illustrates the substantial orientation toward the 33





34


nocturnal placement of total sleep time. Over half (55%) of the total sleep time recorded during BP was obtained during sleep periods initiated between 11IPM and lAM, despite the fact that 'night' comprised only 27% of the total bedrest period.

These findings are in accord with those reported by Weitzman. et al. (1974) relative to the placement of the sleep of subjects maintained on a 3-hour 'day' In that study, the sleep of 7 subjects exhibited a strong tendency to occur, during the scheduled sleep periods beginning at SAM, 6AM, 9AM, and noon. While the authors note the shift away from typical sleep placement (i.e., 11PM 7AM) during the 10 days Of u]itradian Cycling, they emphasize th2 "highly resistant nature" of the circadian loading of the sleep response.

Webb and Agnew (1974a) also noted a tendency for subjects to

maintain a circadian, primarily nocturnal, placement of total sleep time in a free-running environment. The authors cite three aspects of their findings which suggest a continuation of "the entrained influence

of the original 24-hoiur cycle." First, several subjects showed sleep onset times during the first 3 or 4 days which were significantly earlier than would be predicted by a pure free-running hypothesis. Secondly, throughout the 14 days of the study, there were tendencies on the parts of 5 of the 7 subjects for sleep onset times to 'regress' periodically, thereby -increasing the probability that 'sleep onset would occur during the nocturnal phase (between midnight and 7AM). Finally, there were two occasions in which subjects skipped sleep periods, the predicted onsets of which would have occurred during the day phase (specifically, noon anid 4PM), and instead delayed initiation of the next sleep periods to nocturnal placements (11PM and 3AM, respectively).





35


The tendency toward a free-running sleep/waking cycle was not apparent in the present study. While the mean sleep onset times of

the major nocturnal sleep episodes became successively later across the three nights, only 2 individuals exhibited such a tendency. Other subjects delayed sleep onset on Night 3, only to advance onset time on Night 4. Still others exhibited a reverse trend. Such 'scalloping' of the expected free-running rhythm in the present study may reflect -the continued inclination of some subjects to maintain their sleep in its typical nocturnal placement.

Total Sleep ime. The results of the present study are also

consistent with previous findings relative to the amount of total sleep obtained. In summing up the findings on studies of unrestricted sleep, Webb and Agnew (1975a) noted that "it may now be stated with ccnf'idence that permitting subjects to sleep ad lib under conditions which have not increased the sleep need by prior deprivation or increased energy

expenditure will consistently result in increased sleep length," (p, 369). In addition, sleep logs maintained across several weeks reveal that average weekday ('restricted') sleep and weekend ('unrestricted') sleep differs by an hour or more (Johns, Gay, Goodyear, and Masterton, 1971; Webb, 1981; White, 1975).

The subjects in the present study were instructed,' on Night 2, to sleep until -they felt well-rested and no longer sleepy. The mean sleep

duration for Nigrit 2 (8.7 hrs) was within minutes of the mean total sleep time (8.8 hrs) reported for 4 subjects (10 nights each) permitted to sleep as long as they wished, following 10 nights of 'restricted' sleep which averaged 7.4 hrs per night (W4ebb and Agnew,






36


1974a) The mean sleep length for Night 2 is also in agreement with self-reported extensions in weekend sleep lengths (about 8.5 hrs).

The mean total sleep time reported here is, however, almost an

hour less than that of young adults (9.6 hirs) studied by Verdone (1968) and Webb and Agnew (1975a). It seems likely that the mean total sleep time for Night 2, like typically reported weekend sleep, was a consequence of the relaxation of self-imposed controls on sleep length (e.g., alarm clocks), rather than an expression of the complete satisfaction of sleep 'need'. Several subjects in the present study were rather impatient to terminate Night 2 and to initiate the bedrest period. This is illustrated by the case of one subject who reported,

during a brief (3 mim) awakening, that she felt rested and had "'had enough sleep." The report came after approximately 2 hours had elapsed in Night 2. Clearly, such eagerness to 'get on with it' may have resulted in the termination of sleep prior to the spontaneous ending of

sleep due to satiety.

A more accurate measure of maximum sleep length may be derived

from the examination, across the bedrest period, of sleep durations per 24 hours. Beginning with the termination of Night 2 (typically around 8AM) the first.24-hour period was characterized by an average of 14.5 hours of sleep, the second 24 hours contained an average of almost 11 hours of sleep. Such figures are consistent with previously reported 24-hour sleep amounts. For example, Aserinsky (1969) instructed subjects to "make every attempt to sleep" while confined to bed for 30 hours (interrupted only twice for meals). During the first 24 hours (midnight to midnight) subjects obtained an average of*14.4 hours of sleep. During the 24-hour period beginning at 7:30AM, after a full





37


night's sleep (mean =7.04 hrs.), the mean sleep duration was 12.93 hrs. This figure is in remarkable agreement with the average TST for each of the first 2 24-hour periods, subsequent to Night 2, in the present study (12.71 hrs.). In light of such findings, it appears that 12 to 13 hours of sleep per 24 hours may be considered a close approximation of the maximum capacity for sleep (Aserinsky, 1969). However, at 'least three lines of evidence suggest -that substantially shorter sleep times may represent the maximum point of sleep 'need'. First, it has been proposed that REM density may be a reflection of "satisfaction of a sleep need, or . the buildup of a pressure to awaken" (Aserinsky, 1969, p. 155). It was found in that study that REM density reached maximum values after 7.5 to 10 hours of sleep. Secondly, under coniditions of perceptual deprivation, subjects averaged about 12 hours of sleep during the first 24 hours, but subsequently reduced sleep lengths until they approached baseline levels (about 7.5 hirs.) by the third or fourth day of deprivation (Potter and Heron, 1972)-~ Also in the present study, total sleep time during the second 24 hours of bedrest decreased. almost 25% when compared with that of the first 24 hours. It might be speculated that the decline would have continued, if bedrest had continued.

Nevertheless, with the exception of the perceptual deprivation study cited above, all values of sleep satiation or fulfillment of sleep "need" substantialy exceed the 7- to 3-hour sleep durations typically reported for normal young adults. The interpretation of such findings may be generally separated into two opposing views. It has been suggested, on the one hand, that the extended sleep observed in unrestricted sleep regimens is a consequence of decreased sensory input (see for example, Nakagawa, 1980; Heron, 1957). Proponents of




38


this view have speculated that an important -function of the sensory and perceptual systems is to maintain a waking level of arousal in, the reticular activating system. In studies comparable to the present one, where conditions are maintained at a relatively constant level, and

sensory stimulation is not only static but also depressed relative to normal levels, input from peripheral afferent pathways may be insufficient -to furnish the appropriate level of RAS activation. "In general, appropriate sensory stimulation is needed to maintain an arousal state in animals, but repetitive monotonous stimulation habituates the arousal reaction "(Nakaguawa, 1980, p. 532). In short the, according to

this view, extended sleep is the result of neurological "boredom" and/or habituation. Implied in such a view is the notion that the extended sleep time observed under unrestricted or bedrest conditions is not

representative of natural sleep tendency (or requii~ement), but rather, is simply one consequence of the organism's decreased requirement to process environmental stimuli and respond accordingly (Heron, 1957).

This is analagous to Flanigan's (1971) suggestion that the extended sleep observed in nonhuman subjects maintained under con-trolled and highly -favorable conditions (i.e., limited activity, plenty of food, no

predators) may reflect the upper limits of sleep that are organismically possible, rather than that which is typical in the animal 's natural habitats.

In contrast, it has been Suggested (see for example Webb and Agnew, 1975a) that the extended sleep lengths characteristic of unrestricted sleep conditions are, in fact, "normal" and that typically reported sleep durations of 7 to 8 hours may actually be the product of chronic sleep deprivation. In support of such a notion is the finding that





39.

only one third of a group of individuals who completed 6-month sleep diaries reported that they usually awakened spontaneously in the morning. Further, half of the respondents in the study reported that they

were typically not well-rested upon awakening (Kleitman, Mullin, Cooperman, and Titelbaum, 1937). The extension of sleep on weekends reported

by most individuals also suggests a 'need' for more than the typically obtained weekday mean, Further, it has been shown that after 3 consecutive nights of sleep, restricted to a period between 11PM and 7AM (mean = 7.5 hrs), subjects slept an average of 2 hours longer on the 4th night when allowed to awaken spontaneously (Webb and Agnew, 1975a). The subjects in that study were not encouraged to extend sleep; indeed, they were "unaware that they would be allowed to sleep longer."

Similarly, the subjects in the present study were given no in-structions relative to the maintenance of sleep and wakefulness during the bedrest period, yet they also obtained substantially more sleep than self-reported normal sleep durations. This seems to be suggestive of an inclination toward, if not an actual 'need' for, more sleep than is typically obtained. Of interest is the observation that the extended

sleep durations resulted not from a lengthening of sleep episodes initiated at night, but rather from the addition of day sleep periods.

Nocturnal sleep (11PM to 7AM) per 24 hours did not exceed normal night sleep (7.8 h). However, day sleep episodes contributed an average of approximately 5 hours to each 24-hour total. It might be speculated

that the usual sacrifice of such day sleep periods is the source of our chronic sleep deprivation. On the other hand, the bedrest environment

in the present study was highly monotonous, with novel sensory input severely curtailed. It is, therefore, likely that such conditions






40


contributed somewhat to the observed tendency for extended sleep times, as-well.

Intrasleep Structure. The mean sleep stage proportions of the

combined sleep periods during BP did not differ significantly from those of the baseline night (Night 2), and were within normal limits reported for young adult populations, although REM and slow wave sleep (SWS) were significantly reduced when compared to the normative sample used here.

Such findings are in agreement with those reported by Webb and Agnew (1977) relative to the distribution of sleep stages in experimentally controlled, polyphasic sleep regimens. That is, despite influences -from changes in sleep length, p-1ior wakefulness and sleep onset times associated with various schedules of dispersed sleep, "when the total sleep obtained during these altered regimens is compared with baseline sleep on a 24-hour regimen, the basic structure of sleep persists" (p. 448).

Also, in accord with previous reports of dispersed and nap sleep (Maron, Rechtschaffen and Wolpert, 1954; Moses, Hord, Lubin, Johnson, and Naitoh, 1975; Nakagawa, 1980; Webb and Agnew, 1977; Weitzman et al. 1974) the occurrence of REM sleep maintained a circadian pattern across the bedrest period. Sixty percent of total REM time occurred during sleep episodes initiated between 11PM and 7AM. Further, while only 9 of the 84 sleep periods recorded during BP contained no REM sleep, 8 of those sleep episodes occurred during the day, and 7 occurred between 2PM and 10PM. Finally, Figure 4 graphically points out the tendency of REM sleep to occur differentially at intervals across the 24-hour day, with maxima and minima of the circadian oscillation occurring between

4 and 8AM and 4 and 8PM, respectively. Such data seem to underscore the tendency for REM sleep to occur differentially across the 24 hours.





41


The decrease in REM percent during the bedrest period, compared to baseline, also appears to have been the consequence of the circadian placement of REM sleep. While total REM time was differentially obtained within a relatively circumscribed time block, total sleep time was influenced somewhat less by such contingencies. In addition, opportunities for day sleep, when REM was less likely to occur, occupied a larger percentage of the bedrest period than did night sleep time. As a result, over the bedrest period, absolute sleep time was augmented to

a greater extent than was REM time, thereby lessening the proportion of TST spent in REM sleep.

The decreased proportion of REM sleep recorded over the bedrest

period may also be attributable, in part, to the decrease in mean sleep period length, relative to baseline, associated with BP. It has been shown that the amount of TST spent in REM sleep becomes proportionately less with the reduction of sleep length (Webb and Agnew, 1977). Since mean sleep period length during the bedrest period was almost two..thirds less than that for the baseline night (2.99 hrs. vs 8.75 hrs.), a corresponding decline in the proportion of REM sleep for the majority of sleep periods during BP would also be expected.

The observed decrease in REM percent, however, may also be a

reflection of the decreased mobility and/or lack of novel perceptual stimulation characteristic of the bedrest situation. Several investigations of sleep patterns during prolonged bedrest have reported similar declines in REM sleep (Aserinsky, 1969; Nakagawa, 1980; Ryback and Lewis, 1971). In addition, Adey, Bors and Porter (1968) reported decreased

porportions of REM sleep -in quadraplegic patients, relative to normal amounts. The authors proposed that the diminished REM sleep was related





42


to "loss of most motor performance in the waking state," as well as the "absence of highly focused attention in complex motor activity." The finding by Ryback and Lewis (1971), that subjects who were permitted no exercise while confined to bed exhibited decreased REM sleep, while those who were allowed to exercise during bedrest did not exhibit such a decline, also provides some support for this notion. In the present study, the finding that REMS percent decreased (albeit slightly) from the first to th second 24 hours, may likewise be interpreted as supporting an immobilization/sensory habituation effect on the amount of REM sleep obtained.

A rather unexpected fli ding in the present study was that of an increase (though not significant) in the proportion of SWS, as the bedrest period progressed. In their examination of patients with high crevical lesions, Adey et al. (1968), reported the diminution of SWS as well as decreased REM percentage. As with REM sleep, the authors concluded that such a decline was the consequence of immobility. Thus, the findings of the present study appear contrary to what would be expected by an immobilization hypothesis.

On the other hand, the results reported here are in accord with the findings relative to SWS reported by Ryback and Lewis (1971). The bedrest subjects in that study exhibited increases fn SWS with the nonexercise group showing the larger rise. These results were interpreted by the authors as indicating that SWS may serve to "repair or maintain the muscular system" in response to the effects of atrophy, as well as hypertrophy. Thus, it would appear that a change in the proportion of SWS, in either direction, may be explained by one or the other version of an immobilization hypothesis.




43


However, it seems somewhat unparsimonious to have to employ both versions to satisfactorily explain the data in terms, of immobilization. In the present study, SWS comprised a smaller proportion of TST during BP than during the baseline night. The hypothesis proposed by Adey et ai. is useful in explaining such results. Yet, within the bedrest period, there was an increase in the percentage of SWS from the first to the second 24 hours. These findings clearly require the Ryback and Lewis version.

Several observations seem to support an alternative explanation for the increase in SWS across the bedrest period, however. It has been shown (webb and Agnew. 1971, p. 1354) that "longer periods of

wakefulness before sleep result in greater amounts of stage 4 sleep in the first 3 hours of sleep." In the present study, the mean inteval between sleep episodes increased by some 70% (from 1.65 hrs. to

2.82 hrs.) from the first to the second 24-hour period. At the same time, the proportion of SWS increased almost 25% (from 12.9% to 16%). By the same token, periods of wakefulness between sleep episodes averaged about 1 hour longer during the day than at night (3.32 hrs. vs

2.29 hrs.). Correspondingly, SWS comprised a slightly greater proportion of day sleep than of night sleep episodes (16.2% vs 13.2%). The relationship between prior wakefulness and subsequent amounts of SWS seems apparent.

Such an association would also account for the decline of SWS

from Night 2 to the bedrest period. The amount of wakefulness preceeding the baseline night (14 to 17 hrs.) far exceeded the duration of waking intervals characteristic of BP. Consequently, the subsequent sleep periods reflected this difference by exhibiting differential





44


proportions of SWS sleep. Inconsistent with a prior wakefulness hypothesis, however, is the observation by several investigators that there is a SWS 'kick' in the records of extended sleep. That is, extended sleep is oftLen characterized by the reappearance of substantial SWS episodes in the late morning. The possible significance of such a finding is considered later in the discussion.

Results of the examination of the influence of prior sleep on the REM sleep cycle are best described as equivocal. Two examinations were made, one to test the notion that the REM cycle is sleep-dependentIC, the other to examine the possibility that the REM cycle is, instead, a sleep-independent rhythm based on elapsed cluck tLime, regardless of stat e. Support -for both views may be found in the literature. For example, Globus (1966) found that the cycling of REM sleep demonstrated "at least in part" a tendency -to be a time--locked, rather than a sleep-dependent, rhythm in the nap sleep of 2 subjects recorded for a total of 107 REM periods. This conclusion was based primarily on thle observation that correlations between sleep onset and REM onset decreased at certain times -in the afternoon (between noon and 3 PM). That is, during this interval sleep onset got progressively later while REM onset remained constant relative to real time. Yet, prior sleep was not considered and, therefore, the interpretation of such findings is problematic.

In contrast, Moses et al. (1975) and Moses, Lubin, Johnson, and Naitoh (1977) reported findings which suggest that the REM sleep cycle is a sleep-dependent phenomenon. In the 1977 study, the nap sleep of

25 subjects was 'compressed,' by subtracting all intervening waking periods, and then examined as a single sleep episode. The resulting

mean REM cycle lengths did not differ significantly from those recorded





45

during baseline sleep. The same procedure was carried out in the present study, and similar results were obtained. Using the-onset of the last REM period of Night 2 as a starting point, the mean REM cycle length was determined for subjects' 'compressed' bedrest periods, and compared to the mean cycle length for Night 2, The mean cycle length for the compressed period differed from the cycle length for Night 2 by less than 1 minute (96.84 min vs 96.56 min). Such evidence is strongly suggestive of a-sleep-dependent-cycle.

In a second test, each subject's mean REM cycle length for Night 2 was superimposed on both his/her 'intact' bedrest period (i.e., including waking periods) and compressede' bedrest period. The mean deviation of an actual REM occurrence, from the onset of the hypothesized cycle, was used as a measure of 'goodness of fit.' The mean deviation across the 'compressed' bedrest period was 14.5 minutes. In other words, the actual occurrence of REM periods during BP differed by an average of about 15 minutes, as compared to expected occurrences, when intervening waking episodes were disregarded. Again, such a finding is supportive of a sleep-dependent REM sleep cycle. However, a similar finding that

the mean deviation of actual REM sleep from hypothesized REM periods across the 'intact' bedrest period (15.8 min), precludes the acceptance of the sleep-dependence of REM sleep at the absolute exclusion of the 'real time' model.

Numerous studies of napping and dispersed sleep (Carskadon and Dement, 19.75: Globus, 1966; Moses et al., 1915; Nakagawa, 1980; Webb et al., 1966) have reported perturbations in the mode of appearance of REM sleep episodes, characterized by the onset of REM sleep within minutes of the initiation of a sleep period. The reported frequencies





46


of occurrence Of Suich sleep onset REM periods (SOREMPS) have ranged from as few as 311 out of '107 REM episodes (Globus,. 1965) to as many as 21 of 34 REM episodes (Moses et al,, 1975). Maron et al. (1964) reported tLhe Occurrence of no REM onset periods in the examination of 18 subjects who napped in either the afternoon (about 1 :30PM) or evening (about 7:30PM). Yet, the overall rate of occurrence of SOREMPS for 5 studies of dispersed and nap sleep (Carskadon and Dement, 1975; Globus, 1966; Moses et al., 1975; Nakagaua, 1980: Weitzman et al., 1974) is rather

substantial at 39% of all REM periods recorded (range 2.8% 72%). Since such events are almost never observed in normal nocturnal sleep, SOREMPS have been considered to be indications of the overall disruption of normal sleep structure associated with day sleep episodes and

dispersed sleep regimens (Weitzmian et al ., 1974; Mloses et al., 1975).

In the present study, of 153 REM episodes recorded over the bedre1st period, only 3 were initiated at the beginning of a sleep period (i.e., before the appearance of 2 min of stage 2). The infrequency of such occurrences suggests that the dispersed sleep of the present study may be of a different sort than typically observed nap sleep or experimentally dispersed sleep.

Suich a view is further supported by the regularity with which sleep stages occurred across the bedrest period. Seventy percent of all- sleep

periods recorded during BP were 'full blown' sleep episodes (61% of day sleep episodes); that is, all sleep stages were present. By comparison, only 16% of the 550 opportunities for sleep in the Weitzman et a]. study contained all stages. Of the 80 naps examined by Moses et al. (1975), 27.5% contained both SWS and REM sleep. REM sleep occurred in 90% of

all sleep episodes during BP (86% of day sleep episodes), SWS occurred





47


in 85% of the sleep periods in this study (80% of day sleep periods),

Again by comparison, the naps of subjects studied by Moses et al. contained REM episodes 43% of the time; 35% of the sleep opportunities examined by Weitzman et al. contained REM and 48% contained SWS.

Seven of the 9 sleep periods which contained no REM sleep occurred between 2PA and 10PM. Such a finding is consistent with previous ones that REM sleep typically increases in amount from midnight to morning and decreases toward evening (Moses et al., 1975; Nakagawa., 1980; Wei'tzrnan et al. 1974). It is likely, then, that the absense of REM during some sleep episodes in the present study were, at leas In part,, due to circadian effects.

A similar possibility also exists relative to 'missing' periods of SIAS (specifically, stage 4). Several examinations of extended

sleep (see for example, Gagnon and Dekoninck, 1981; Webb, Campbell and Hendlin, 1981) have noted the re-occurrence of stage 4 (in relatively large amounts) during the late morning. In the present study, sleep episodes characterized by the absence of stage 4 occurred across the entire day (6:40AM to 10:45 PM). It is interesting to note, however, that an obvious 'gap' occurred between 10:30AM and 2:20PM. That is, the sleep episodes initiated during those four hours always contained

stage 4 sleep. Sleep periods initiated after about 11PM,. and throughout the night, also contained stage 4. Thus, it might be suggested that the absence of stage 4 in 15% of the sleep episodes in this study, was the manifestation of the biphasic cycling of stage 4 sleep, rather than a reflection of the disruption of normal sleep processes. A further indication of the possible bipolar nature of stage 4 is the observation that total stage 4 time for BP was evenly distributed between day and night (716' min vs 714 min).





48


Clearly, the differences in the appearance and sequencing of sleep stages between this study and the 2 studies of nap sleep with which it has just been compared, are attributable in large part to the difference in mean sleep episode length (3 hrs. vs 1 hr.) This i~s espccially true of REM sleep, if sleep dependency is assumed. Nevertheless, the persistence arid regularity with which sleep stages occurred across the bedrest period is noteworthy. There were no significant differences in the proportions of sleep stages between bedrest sleep and Night 2 sleep. Nor, were there significant changes in sleep stage percentages between the first and second 24 hours of bedrest. Finally, only REM sleep percent varied significantly relative to day or night onset times.

Weitzman et al. (1974) noted that the sleep episodes which they observed were "clearly not minatures of the normal 8 hr. sleep pattern." Such a view has been echoed by most others who have studied dispersed sleep and naps. Most of these studies have differed from the present one, in that the sleep periods were typically experimentally circumscribed, rather than being placed on a 'self-demand' schedule. Such a methodological difference may account for the dissimilarity of results reported here.

Yet, Nakagawa (1980) also reported significant differences in day episodes versus baseline (night) sleep, despite the-fact

that the onset and termination of sleep periods was not regulated. Significant increases in stage 1, significant decreases in stages 4

and REM and the presence of SOREMPS led Nakagawa to conclude that "daytime sleep itself might be characterized as a special sleep at the onset of the diurnal sleep-waking cycle," (p). 534).






49



In that study, Nakagawa noted a sleep/waking cycle of approxi.mately 4 hours, and suggested that a significant relationship seemed to exist between this cycle and the scheduling of meals. The quadriphasic rhythm observed during the present study did not appear to be the result of feeding schedules, since less than half (42%) of waking periods were used for meals. Further, one subject (#932) fasted

for the duration of the experiment, yet maintained the least variable sleep/waking cycle of the group.

Uniphasic or Ultradian? The results of the present study seem to suggest, then, that the daytime sleep observed was not a "special" sleep, but rather was the continuation of an ultradian sleep/waking rhythm, which nevertheless maintained a definite tendency to be weighted toward a nocturnal placement, relative to total sleep time.

The regularity with which sleep episodes were initiated across the bedrest period also seems to support the notion- of an ultradian pattern of sleep. The mean within-subject variability of the sleep/ waking cycles was 2.66 h (=SD). While the between-subject variability was similar (SD =2.73), most of the variability was contributed by a single subject. This subject (#946, see Table II) exhibited a sleep/waking cycle (13.59 h) which more than doubled the mean sleep/ waking cycle length of the other 8 subjects (5.85 h). Likewise, the

variability of the sleep/waking cycle of this subject (OD = 4.52) exceeded the group mean by almost twice (SD m2.43). Excluding this

subject's sleep/waking cycle length from the group mean reduces the

variability of the quadriphasic rhythm to less than an hour (SD 57 min).




50


Such a pattern of variability is similar to that A so rved for the REM sleep cycle. The 90 to 100 minute REM cycle is of limited value in predicting the occurrence of REM episodes within the sleep record of a single subject. Yet for a population, the predictability of a 90-100 minute cycle is quite good.

A further indication of the regularity of the quadriphasic sleep/ waking cycle is the observation that the overall cycle length did not change significantly from one 24 hour period to the next. On the other hand, the length of the sleep/waking cycle did differ as a function of day and night, with the night cycle being substantially longer (7.8 hrs. vs 5.2 hrs.). 'It might be speculated that this difference was the consequence of subjects having spent the majority of their lives entrained to a monophasic sleep system, and that given time, the difference in cycle lengths might diminish. The present data do not support such a notion, however since the difference between night sleep/waking cycle length and day sleep/waking cycle length did not lessen significantly in the second 24 hours of bedrest.

The polyphasic distribution of the sleep of numerous animals under experimental conditions is well documented. Berger (1972) has noted that "if we wish to study comparative aspects of normal sleep, we shouldn't impose wakefulness at any time on any species, including the human, but should take 24-hour recordings under equivalent conditions." The relatively basal activity levels, absence of novel perceptual stimulation and the satisfaction of bodily functions

characteristic of the present experimental situation are comparable





51


to typical conditions under which animal sleep is studied.- That a polyphasic sleep pattern was exhibited under 'such conditions, therefore, was not completely unexpected. Indeed, in 1972 Snyder speculated that "even in the human we would probably find polyphasic sleep if we used the same conditions that we use for animals," (p. 45),

While the monotony of the situation was quite likely instrumental in the increase in total sleep time over the bedrest period, it seems less probable that such "neurological boredom" would be expressed in a cyclic manner. Yet, the possibility remains, as noted by Weitzman (1972) with regard to the sleep cycle in cats. "In your experimental situation, you are measuring in part the shift o~f pattern which is taking place from a previous long-established light-dark pattern, inl addition to adaptation to a strange new environment. Are you measuring the effect of that shift as well as the intrinsic rhythmicity in the cat? If the measured rhythmicity were constant over, say, several weeks, thenit would support the concept that a basic rhythmicity was beinj measured, However, if it tended to change over 1, 2, or 3 weeks that would suggest that you are at least partially measuring an adaptation effect," (p. 201). Obviously, the question relative to human sleep cannot be resolved by results of the' present experiment.

It is suggested, though, that the conditions of the present study

represented a further advancement along a continuum of slackened behavioral controls over the sleep process, the results of which was a

clearer manifestation of an inherent ultradian sleep/waking rhythm, rather than the adaptation of the sleep response to a "strange new environment." It is further suggested that such a rhythm, albeit in




52


highly dampened form, may also be reflected in the occasional naps taken by normal adults (especially on weekends), the increased frequency of naps associated with the loosened schedules of geriatric and college populations, the sleep patterns of infants, and the polyphasic sleep patterning noted in unstructured environments (i.e., Arctic expeditions).

And, I would suggest one additional manifestation of the ultradian nature of the sleep response. Webb (1969) has noted that "(no one yet has exploited a foundation to study the 'siesta' patterns of the tropi-cal and subtropical cultures, although at various odd moments, I am most tempted),' (p. 58). Results of the present study may suggest that such a proposal, to examine cross-cultural diversity in the cycling of rest and activity, .may deserve a better fate than parenthetical consideration.














APPENDIX 1
SLEEP QUESTIONNAIRE


I. How many hours, on the average, do You usually sleep per night?
less, 5, 51, 6, 61, 7, 7i5 8, 81, 9, 95, 10, more

2 How many hours of sleep per night do you think that you need to
feel well Ad function adequately?

3, How many hours of sleep per night do you prefer to have when You
have a chance?

4. In general, the amount of sleep that you usually get is:
not enough about enough too much

5. On the average, about how many naps do you usually take each week?
0, 1, 2, 3, 4, 5, 6, 7, 8, or more

6. About how many total hours do you usually spend each week taking
naps?

7. About how many times on the average do you wake up each night?
0, 1, 2, 3, 4, 5, 6, 7, 8, or more

8. About how much total time do you spend awake after going to sleep?.
none, less than 5 min., 5-15, 16-30, 31-60, more

9. About how many minutes does it typically take for you to go to
sleep? less than 5, 5-15, 16-30, 31-60, more than 1 hour


10. How often do you have trouble getting to sleep as quickly as you
would like? Almost never, occasionally, often, almost always

11. How much do you enjoy sleep?
not at all, a little, moderately, much

12. How regular are your bedtimes?
very regular, somewhat regular, somewhat irregular, very
irregular

13. How regular are your wake-up times?
very regular, somewhat regular, somewhat irregular, very
irregular

14. What is your average bedtime during weekdays? TO nearestl min-T


53





54



15. What is your average bedtime on weekends?
o nearest 15 m5j
16. What is your average wake up time on weekdays? (to nearest ITAT'n.7


17, What is your average wake up time on weekends?
To nearest 15 min.

IT How well do you usually sleep at night?
Very well, satisfactorily, some problems, poorly.

19. Do you usually feel well rested when you wake up, or soon thereafter?
Almost never, occasionally, often, almost always

20. On the average, how many days per week do you usually go to bed
acre than one (1) hour earlier or later than your average bedtime?
0, 1, 2, 3, 4, 5, 6, 7

21. On the average, how many d ays per week do you usually wake up Tfiore
than one (1) hour earlier or later than your usual wake up time?
0, 1, 2, 3, 4, 5, 6, 7 .

22. Do you consider your sleep very light or very deep? Check below.
very light-- -. very deep

23. Do you feel tired during the day because you have slept poorly?
Never sometimes often always

24. How much alcohol do you drink?
Never sometimes everyday

25. Do you wake up before you intend to?
never
once or twice a month
---once or twice a week
nearly every day

26. Do you use sleeping pills to get to sleep?
never
about once a month
_--about once a week
nearly nightly

27. What, if any, sleeping pills do you use?

28. What, if any, prescription drugs are you presently taking?














REFERENCES


Adey, W., Bors, E. and Porter, R. EEG Sleep Patterns After High
Cervical Lesions in Man. Arch. Neurol., 1968, 19, 377-383.

Agnew, H. and Webb, W. Sleep Stage Scoring. J. Suppl. Abstr.
Serv., Am. Psychol. Assoc., Ms., 1972a, No. 293.
Agnew, H. and Webb, W. Measurement of Sleep Onset by EEG Criteria.
Amer. J. EEG Technol., 1972b, 12(3), 127-134.

Aschoff, J. Circadian Rhythms in Man. Science, 1965, 148, 1427-1432.

Aschoff, J. Features of Circadian Rhythms Relevant for the Design
of Shift Schedules. Er onomics, 1978, 21(10), 739-754.

Aschoff, J. and Wever, R. Human Circadian Rhythms: A Multioscillatory System. Federation Proceedin9s, 1976, 35(12),
2326-2332.

Aserinsky, E. The Maximal Capacity for Sleep: Rapid Eye Movement
Density as an Index of Sleep Satiety. Biological Psychiatry,
1969, 1, 147-159.

Berger, R. Discussion in Chase, M. H. The Sleeping Brain, PerspecStives in the Brain Sciences, Brain Information Service/Brain
Research Institute, Vol. 1, p. 45, Los Angeles: UCLA, 1972.

Carskadon, M. and Dement, W. Sleep Studies on a 90-minute Day.
Electroenceph Clin. Neurophysiol., 1975, 39, 145-155.

Colquhoun, W. P. Circadian Variations in Mental Efficiency. In
W. Colquhoun (Ed.), Biological Rhythms and Human Performance,
pp. 39-107, New York: Academic Press, 1971.

Ellingson, R. Ontogenesis of Sleep in the Human. In G. Lairy and
P. Salzarulo (Eds.), The Experimental_ Study of Human Sleep:
Methodological Problems, pp. 129-150, New York: Elsevier Pub.,
1975.

Flanigan, W. Phylogeny and Sleep Behavior: A Review. Unpublished
manuscript, 1971.

Flanigan, W. Behavioral States and Electroencephaograms of Reptiles.
In M. Chase (Ed.), The Sleeping Brain, Brain Information Service/
Brain Research Institute, Los Angeles: UCLA, 1972.

55




56


Froberg, J., Karisson, C., Levi, L. and Lidberg, L. Psychobiological
Circadian Rhythms During a 72 Hour Vigil. Forsvarsmedicin, 1975,
11, 192-201.

Gagnon, P., and De Koninck. Reappareance of EEG Slow Waves in Extended
Sleep. Sleep Research, 1981 (in press).

Gesell, A., and Amatruda, C. The Embr'ology of Behavior, New York:
Harper, 1945, pp. 144-159.

Globus, G. Rapid Eye Movement Cycle in Real Time. Arch. Gen.
Psychiat., 1966, 15, 654-659.

Heron, W. The Pathology of Boredom. Scientific Amer., January,
1957, 1, 430-436.

Johns, M., Gay, T., Goodyear, M. and Masterton, J. Sleep Habits
of Healthy Young Adults: Use of a Sleep Questionnaire. Brit.
J. Prey. Soc. Med., 1971, 25, 236-241.

Kleitman, N. Sleep and Wakefulness, Chicago: University of
Chicago Press, 1963.

Kleitman, N., and Engelmann, T. Sleep Characteristics of Infants.
J. Appied Physiology, 1953, 6, 269-282.

Kleitman, N., Mullin, F., Cooperman, N. and Titelbaum, S. Sleep
Characteristics, Chicago: University of Chicago Press, 1937.

Lewis, H. Sleep Patterns on Polar Expeditions. In G. Woistenholme
and M. O'Connor (Eds.), CIBA Foundation Symposium on the Nature
of Sleep, pp. 322-328, London: J. and A. Churchill, 1961.

Lewis, H., and Masterton, J. Sleep and Wakefulness in the Arctic.
Lancet,.1957, 1, 1262-1266.

Maron, L., Rechtschaffen, D. and Wolpert, B. Sleep Cycle During
Napping. Arch. Gen. Psychiat., 1964, 11, 503-508.

Moses, J., Hord, D., Lubin, A., Johnson, L. and Naitoh, P. Dynamics
of Nap Sleep During a 40 Hour Period. Electoenceph Clin.
Neuropjhysiol., 1975, 39, 627-633.

Moses, J., Lubin, A., Johnson, L. and Naitoh, P. Rapid Eye Movement
Cycle is a Sleep-Dependent Rhythm. Nature, 1977, 265(5592),
360-361.

Murry, E., Williams, H. and Lubin, A. Body Temperature and Physchological Ratings During Sleep Deprivation. J. Exp. Psychology,
1958, 56(3), 271-274.




57


Nakagawa, Y. Continuous Observation of EEG Patterns at Night and
in Daytime of Normal Subjects Under Restrained Conditions I.
Quiescent State When Lying Down. Electroenceph. Clin. Neurophysiot, 1980, 49(5-6), 524-537.

Parmellee, A., Wenner, W. and Schulz, H. Infant Sleep Patterns
From Birth to 16 Weeks of Age. J. Pediat., 1964, 65, 576-582.

Potter, W., and Heron, W. Sleep During Perceptual Deprivation.
Brain Research, 1972, 40, 534-539.

Richter, C. R. Biological Clocks in Medicine and Psychiatry, Springfield, Illinois: C. C. Thomas, 1965.

Ryback, R., and Lewis, 0. Effects of Prolonged Bed Rest on EEG
Sleep Patterns in Young, Healthy Volunteers. EEG and Clinical
Neurophysiol., 1971, 31, 395-399.

Ryback, R., Lewis, 0. and Lessard, C. Psychobiologic Effects of
Prolonged Bed Rest (Weightless) in Young, Healthy Volunteers
(Study 2). Aerospace Medicine, 1971, 42(5), 529-535.

Salzarulo, P., Fagioloi, I., Salomon, F., Ricour, C., Raimbault,
G., Ambrosi, S., Cicchi, V., Duhamel, J. and Rigoard, M. Sleep
Patterns in Infants Under Continuous Feeding From Birth. EEG
and Clin. Neurophysiol., 1980, 49, 330-336.

Sander, L., Stechler, G., Burns, P. and Julia, H. Early MotherInfant Interaction and 24-Hour Patterns of Activity and Sleep.
J. Am. Acad. Child. Psychiat., 1970, 9, 103-123.

Snyder, F. Discussion in Chase, M. H., The Sleep Brain. Perspec* tives in the Brain Sciences, Brain Information Service/Brain Research Institute, Vol. 1, p. 45, Los Angeles: UCLA, 1972.

Sterman, M. Ontogeny of Sleep: Implications for Function. In
Drucker-Colin, Shkurovich and Sterman (Eds.). The Functions
of Sleep, pp. 207-232, New York: Academic Press, 1979.

Symposium on Quantitative Biology, Vol. XXV, Bioloqical Clocks.
Cold Springs Harbor, Long Island Cold Springs Harbor Symp.,.
Quant. Biol., 1960, 25.

Verdone, P. Sleep Satiation-Extended Sleep in Normal Subjects.
Electroenceph. Clin. Neurophysiol., 1968, 24(5), 417-423.

Webb, W. Twenty-Four Hour Sleep Cycling. In Kales, A. H. (Ed.),
Sleejp: Physiology and Pathology, pp. 53-65, Philadelphia:
J. B. Lippincott Co., 1969.




58


Webb, W. Sleep Behavior as a Biorhythm. In W. P. Colquhoun (Ed.),
Biological Rhythms and Human Performance, pp. 149-178, New York:
Academic Press, 1971.

Webb, W. Sleep and Naps, Speculations in Science and Tech., 1978,
1(4), 313-318.

Webb, W. Sleep Patterns in Older Males and Females. Research Comm.
in Psychiatry, Psycholoy and Biology, 1981 (in press).

Webb, W., and Agnew, H. Stage 4 Sleep: Influence of Time Course
Variables. Science, 1971, 124, 1354-1356.

Webb, W., and Agnew, H. Sleep and Waking in a Time-Free Environment.
Aerospace Med., 1974a, 45(6), 617-622.

Webb, W., and Agnew, H. Regularity in the Control of the FreeRunning Sleep-Wakefulness Rhythm. Aerospace Med., 1974b,
415(7), 701-.704 .

Webb, W.,and Agnew, H. Are We Chronically Sleep Deprived? Bulletin
of thePcooic Suciety, 1975, 6(1), 47-48.

Webb, W., and Agnew, H. Sleep Efficiency for Sleep-Wake Cycles of
Varied Length. Psychophysio., 1975a, 6(1), 47-48.

Webb, W.,and Agnew, H. The Effects on Sbbsequent Sleep of an Acute
Restriction of Sleep Length. Psychophysiol., 1975b, 12.(4),
367-370.

Webb, W., and Agnew, H. Analysis of the Sleep Stages in Sleep-Wakefulness Regimens of Varied Length. Psychophsioloqy, 1977,
14_(5), 445-450.

Webb, W., and Agnew, H. Effects of Rapidly Rotating Shifts on Sleep
Patterns and Sleep Structure. Aviat. Space and Environ. Med.,
1978, 49(2), 384-389.

Webb, W., Agnew, H. and Sternthal, H. Sleep During the Early
Morning. Psychonom. Sci., 1966, 6, 277-278.

Webb, W., Agnew, H. and Williams, R. Effect on Sleep of a Sleep
Period Time Displacement. Aerospace Med., 1971, 42(2), 152155.

Webb, W., Campbell, S. and Hendlin, R. The Termination of Extended
Sleep. Biological Psychology, 1980, 11, 45-48.





59



Weitzman, E. Periodicity in Sleep and Waking States in Chase M. H.
(Ed.), The Sleeping Brain, Perspectives in the Brain Sciences, Vol. 1, p. 45, Brain Information Service/Brain Research Institute, Los Angeles: UCLA, 1972.

Weitzman, E., Nogeire, C., Perlow, M., Fukushima, P., Sassin, J.,
McGregor, P., Gallagher, T. and Hellman, L. Effects of a Prolonged 3-Hour Sleep-Wake Cycle on Sleep Stages, Plasms Cortisol.
Gowth Homone and Body Temperature in Man. J. Clin. Endocrinol.
Metab., 1974, 38, 1018-1029.

Wever, R. Internal Phase-Angle Differences in Human Circadian
Rhythms: Causes for Changes and Problems of Determinations.
Int. J. Chronobiology, 1973, 1, 371-390.

White, R. Sleep Length and Variability: Measurement and Interrelationships. PhD Dissertation, University of Florida, 1975.

Williams, R.F, Agnew, H. and Webb, W. Sleep Patterns in Young Adults:
An EEG Study. Electroenceph. Clin. Neurophysiol., 1964,
17, 376-381.

Williams, R., Agnew, H. and Webb, W. Sleep Patterns in Young Adult
Females: An EEG Study. Electroenceph. Clin. Neurophysiol.,
1966, 20, 264-266.















BIOGRAPHICAL SKETCH



Scott Searcy Campbell was born February 25, 1952. HeI received his B.A. degree, with a major in psychology, from Florida State Univers ity in 0974. From September, 1972, to June, 1978, he attended Montana State University from which he received the M.S. degree in experimental psychology. In September, 1978, he continued his graduate studies, under the supervision of Wilse B. Webb, at the University of Florida. Upon completion of the Ph.D. degree, he accepted a postdoctoral fellowship at Harvard Medical School.





























60










I certify that i have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.


lot,__Wise B. Webb, Chairman Graduate Research Professor of Psychology


1 certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.-------William K. Berg
Professor of Psychology


I certif y that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.



Donald A. Dews bury
Professor of Psychology


I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deg~ree-'of Doctor of Philosophy.



Charles M. Levy
Professor of Psychology










I certify that I have read this study and that in my,; opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.




W/Uack R. Smith
Professor of Electrical
Engineering


This dissertation w~as submitted to the Graduate Faculty of the Department of Psychology in the College of Liberal Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of: the requirements for the degree of Doctor of Philosophy.



Augus 1979Dean for Graduate Studies and Research




Full Text
27
Of 84 total sleep episodes recorded for the group across the
bedrest period, 59 (70.2%) occurred during the day and 25 (29.8%)
occurred at night. However, over half of the total sleep time (54.7%)
during BP was obtained during episodes initiated between 11 PM and 7
AM.
Comparisons between day sleep periods (DS) and night sleep
periods (NS) of temporal measures and sleep stage percentages are
shown in Table IV. The most notable difference in measures of DS and
NS was with regard to mean sleep period duration. NS episodes con
tinued for a mean duration 5.52 hours (SD = 2.65 h) compared to an
average length of 1.93 hours (SD = ,98 h) for sleep episodes
initiated during the day (t (82) = 6.58, p<.001). Conversely, the
mean duration of waking episodes was 1.47 hours (SD = 1.04 h) for
nighttime events and 3.12 hours (SD = 2.4 h) for those initiated
during the day (t (82) = 4.29, p<.001). Consequently, the sleep/
waking cycle length (defined as the interval between the onset of
one sleep episode and the onset of the next) increased significantly
from day (5.25 h, SD = 1.74) to night (7^81 h, SD = 3,26) t (73) =
2.89, pc.01.
Figure 3 presents, by hours, the number of subjects sleeping
across the bedrest period. The group tendency toward nocturnal
sleep placement is evident. Of 144 total subject-hours available for
sleep during the 2 night periods (11 PM to 7 AM), 104 (72.3%) were
used for that purpose. In contrast, of the 279 subject hours com
prising the first and second day periods (.8 AM to II PM and 7 AM to
11 PM) less than half (45.2%) were spent sleeping.


41
The decrease in REM percent during the bedrest period, compared
to baseline, also appears to have been the consequence of the circadian
placement of REM sleep. While total REM time was differentially
obtained within a relatively circumscribed time block, total sleep
time was influenced somewhat less by such contingencies. In addition,
opportunities for day sleep, when REM was less likely to occur, occupied
a larger percentage of the bedrest period than did night sleep time. As
a result, over the bedrest period, absolute sleep time was augmented to
a greater extent than was REM time, thereby lessening the proportion of
TST spent in REM sleep.
The decreased proportion of REM sleep recorded over the bedrest
period may also be attributable, in part, to the decrease in mean sleep
period length, relative to baseline, associated with BP. It has been
shown that the amount of TST spent in REM sleep becomes proportionately
less with the reduction of sleep length (Webb and Agnew, 1977). Since
mean sleep period length during the bedrest period was almost two-thirds
less -than that for the baseline night (2.99 hrs. vs 8.75 hrs.), a corre
sponding decline in the proportion of REM sleep for the majority of sleep
periods during BP would also be expected.
The observed decrease in REM percent, however, may also be a
reflection of the decreased mobility and/or lack of novel perceptual
stimulation characteristic of the bedrest situation. Several investi
gations of sleep patterns during prolonged bedrest have reported similar
declines in REM sleep (Aseririsky, 1969; Nakagawa, 1980; Ryback and Lewis,
1971). In addition, Adey, Bors and Porter (1968) reported decreased
porportions of REM sleep in quadraplegic patients, relative to normal
amounts. The authors proposed that the diminished REM sleep was related


44
proportions of SWS sleep. Inconsistent with a prior wakefulness
hypothesis, however, is the observation by several investigators that
there is a SWS 'kick' in the records of extended sleep. That is,
extended sleep is often characterized by the reappearance of substantial
SWS episodes in the late morning. The possible significance of such a
finding is considered later in the discussion.
Results of the examination of the influence of prior sleep on the
REM sleep cycle are best described as equivocal. Two examinations
were made, one to test the notion that the REM cycle is sleep-dependent,
the other to examine the possibility that the REM cycle is, instead, a
sleep-independent rhythm based on elapsed clock time, regardless of state.
Support for both views may be found in the literature, For example,
Globus (1966) found that the cycling of REM sleep demonstrated "at least
in part" a tendency to be a time-locked, rather than a sleep-dependent,
rhythm in the nap sleep of 2 subjects recorded for a total of 107 REM
periods. This conclusion was based primarily on the observation that
correlations between sleep onset and REM onset decreased at certain
times in the afternoon (between noon and 3 PM). That is, during this
interval sleep onset got progressively later while REM onset remained
constant relative to real time. Yet, prior sleep was not considered
and, therefore, the interpretation of such findings is problematic.
In contrast, Moses et al. (1975) and Moses, Lubin, Johnson, and
Naitoh (1977) reported findings which suggest that the REM sleep cycle
is a sleep-dependent phenomenon. In the 1977 study, the nap sleep of
25 subjects was 'compressed,' by subtracting all intervening waking
periods, and then examined as a single sleep episode. The resulting
mean REM cycle lengths did not differ significantly from those recorded


58
Webb, W. Sleep Behavior as a Biorhythm. In W. P, Colquhoun (Ed.),
Biological Rhythms and Human Performance, pp. 149-178, New York
Academic Press, 1971.
Webb, W. Sleep and Maps, Speculations in Science and Tech., 1978,
1(4), 313-318.
Webb, W. Sleep Patterns in Older Males and Females. Research Comm.
in Psychiatry, Psychology and Biology, 1981 (in press).
Webb, W., and Agnew, H. Stage 4 Sleep: Influence of Time Course
Variables. Science, 1971. 124, 1354-1356.
Webb, W., and Agnew, H. Sleep and Waking in a Time-Free Environment'.
Aerospace.Med., 1974a, 45(6), 617-622.
Webb, W., and Agnew, H. Regularity in the Control of the Free-
Running Sleep-Wakefulness Rhythm. Aerospace Med., 1974b,
45(7), 701-704.
Webb, W.,and Agnew, H. Are We Chronically Sleep Deprived? Bulletin
of the Psychonotnic Society, 1975, 6(1 ), 47-48.
Webb, W., and Agnew, H. Sleep Efficiency for Sleep-Wake Cycles of
Varied Length. Psychophysiol., 1975a, 60), 47-48.
Webb, W.,and Agnew, H. The Effects on Subsequent Sleep of an Acute
Restriction of Sleep Length. Psychophysiol., 1975b, 12(4),
367-370.
Webb, W., and Agnew, H. Analysis of the Sleep Stages in Sleep-Wake
fulness Regimens of Varied Length. Psychophsioloqy, 1977,
14(5), 445-450.
Webb, W., and Agnew, H. Effects of Rapidly Rotating Shifts on Sleep
Patterns and Sleep Structure. Aviat. Space and Environ. Med.,
1978, 49(2), 384-389.
Webb, W., Agnew, H. and Sternthal, H. Sleep During the Early
Morning. Psychonom. Sci., 1966, 6, 277-278.
Webb, W., Agnew, H. and Williams, R. Effect on Sleep of a Sleep
Period Time Displacement. Aerospace Med., 1971, 42(2), 152-
155.
Webb, W., Campbell, S. and Hendlin, R. The Termination of Extended
Sleep. Biological Psychology, 1980, H, 45-48.


22
detail, the sequencing of sleep and wakefulness of one subject (932)
across the bedrest period. Individual and group statistics for the
entire bedrest period are presented in Table II. BP was character
ized by an alternation of waking periods with an average length of
2.7 hours (range = .4 to 10.2 hours) and sleep periods with a mean
duration of 2.99 hours (range = .5 to 11.6 hours). Subjects spent
almost half of BP asleep (47.6%) with a mean total sleep time of
28 hours (range = 20 to 37.3 hours), Sleep was dispersed across an
average of 9.3 episodes during the bedrest period (range = 5 to 13).
These results are in contrast to what would be expected under
the assumption of a uniphasic model of sleep. Under such an assump
tion, one would predict 2 discrete sleep episodes across the 60 hours,
placed within the normal nocturnal periods (approximately 11 PM to
7 AM), with a mean TST for the bedrest period of about 16 hours
(about 27% of BP).
Sleep periods during BP could be divided into four general
categories based on the presence or absence of stage 4 and REM sleep,
as shown in Table III. Seventy percent (59) of all sleep periods
recorded over the bedrest period were 'full blown' sleep episodes,
in that all stages of sleep viere observed. Of the remaining 25
sleep periods, 3 (all day episodes) showed neither REM nor stage 4
sleep. Nine sleep periods (8 days, 1 night) exhibited an absence of
REM sleep and 13 (12 days, 1 night) showed no stage 4 sleep.
The mean period length of those episodes with neither REM nor
stage 4 was 61 minutes. The mean duration of 'noREM' sleep periods
was 65 minutes, and the 'noStg4'- sleep episodes averaged 84 minutes


4
phenomenons subject both to the dictates of physiological maturation
and the influences of environment. Such factors as geography and
social custom entrain the physiological substrates of the sleep-waking
cycle to determine the behavioral patterns which will come to
characterize these states in a given culture."
Indeed, the very fact that adult humans maintain a 24-hour
sleep/waking periodicity, in spite of an inherent rhythm with a
frequency longer than 24 hours, is clear evidence of the entrain-
ability of the rhythm.
It appears then, that the oscillation between sleep and wake
fulness, characteristic of adult human subjects, meets the criteria
necessary for its designation as a biological rhythm. Sleep is an
inborn, independently regulated, exogenously entrainable, regularly
occurring response.
Uni phasic or Ultradian? While sleep and wakefulness clearly
comprise a physiological system which may be classified as a bio
logical rhythm, determination of the rhythm's inherent periodicity
remains less clear, since few studies have been conducted which
adequately address this point. It is generally assumed that the
sleep/waking cycle is an uniphasic one. In addition to the obvious
support for such a notion, i.e., that adult humans typically obtain
sleep in a single, nocturnal block, results from studies in which the
normal sleep period has been dispersed or eliminated have also been
cited as evidence for this view. In such experiments, performance
measures are usually lowest and subjective reports of sleepiness are


40
contributed somewhat to the observed tendency for extended sleep times,
as wel 1.
Intrasleep Structure. The mean sleep stage proportions of the
combined sleep periods during BP did not differ significantly from those
of the baseline night (Night 2), and were within normal limits reported
for young adult populations, although REM and slow wave sleep (SWS) were
significantly reduced when compared to the normative sample used here.
Such findings are in agreement with those reported by Webb and Agnew
(1977) relative to the distribution of sleep stages in experimentally
controlled, polyphasic sleep regimens. That is, despite influences from
changes in sleep length, prior wakefulness and sleep onset times assoc
iated with various schedules of dispersed sleep, "when the total sleep
obtained during these altered regimens is compared with baseline sleep
on a 24-hour regimen, the basic structure of sleep persists" (p. 448).
Also, in accord with previous reports of dispersed and nap sleep
(Marn, Rechtschaffen and Wolpert, 1964; Moses, Hord, Lubin, Johnson,
and Naitoh, 1 975; Nakagawa, 1980; Webb and Agnew, 1 977 ; Weitzman et al.,
1974) the occurrence of REM sleep maintained a circadian pattern across
the bedrest period. Sixty percent of total REM time occurred during
sleep episodes initiated between 11 PM and 7AM. Further, while only 9
of the 84 sleep periods recorded during BP contained no REM sleep, 8 of
those sleep episodes occurred during the day, and 7 occurred between 2PM
and 10PM. Finally, Figure 4 graphically points out the tendency of
REM sleep to occur differentially at intervals across the 24-hour day,
with maxima and minima of the circadian oscillation occurring between
4 and 8AM and 4 and 8PM, respectively. Such data seem to underscore
the tendency for REM sleep to occur differentially across the 24 hours.


3
After observing the development of an initial 25-hour periodicity in
an infant maintained on a self-demand sleep-wakefulness and feeding
schedules Kleitman and Engelmann (1953) concluded that "there may be
a 'natural' rhythm which only has to be adjusted to the 24 hour
alternation of night and day" (p. 280). Other researchers have also
observed the development of adult-like, primarily monophasic,
nocturnally placed sleep under similar conditions of self-demand
feeding (Parmelee, Wenner, and Schulz, 1964; Salzarulo, Fagioli,
Salomon, Ricour, Raimbault, Ambrosi, Cicchi, Duhamel, and Rigoard,
1980).
While the sleep/waking rhythm clearly appears to be inborn and
independently regulated, it may, at the same time be entrained by
exogenous Zeitgebers, to frequencies other than those characteristic
of the free-running environment. It has been demonstrated that
human subjects are capable of maintaining sleep/waking cycles of
various lengths, quite efficiently within a range of about 20 to 27
hours, as the result of entrainment to appropriate 1 ight/dark regimens.
It has been further shown that the sleep/waking rhythm characteristic
of a free-running environment may be regulated by "rigid control of
the sleep portion of the cycle . particularly of the wakeup time"
(Webb and Agnew, 1974b; p. 701).
Several investigators also point to the developmental literature
as a source of evidence for the entrairiability of the sleep/waking
rhythm (Ellingson, 1975; Gesell and Armatruda, 1945; Kleitman, 1963;
Sander, Stechler, Burns, and Julia, 1970; Sterman, 1979). As
expressed by Sterman (p. 213), "the development of sleep is a postnatal


34
nocturnal placement of total sleep time. Over half (55%) of the total
sleep time recorded during BP was obtained during sleep periods initi
ated between 11PM and 7AM, despite the fact that 'night' comprised only
27% of the total bedrest period.
These findings are in accord with those reported by Weitzman et al.
(1974) relative to the placement of the sleep of subjects maintained on
a 3-hour 'day'. In that study, the sleep of 7 subjects exhibited a
strong tendency to occur during the scheduled sleep periods beginning
at 3AM, 6AM, 9AM, and noon. While the authors note the shift away from
typical sleep placement (i.e., 11PM 7AM) during the 10 days of ultra-
dian cycling, they emphasize the "highly resistant nature" of the
circadian loading of the sleep response.
Webb and Agnew (1974a) also noted a tendency for subjects to
maintain a circadian, primarily nocturnal, placement of total sleep
time in a free-running environment. The authors cite three aspects of
their findings which suggest a continuation of "the entrained influence
of the original 24-hour cycle." First, several subjects showed sleep
onset times during the first 3 or 4 days which were significantly
earlier than would be predicted by a pure free-running hypothesis.
Secondly, throughout the 14 days of the study, there were tendencies
on the parts of 5 of the 7 subjects for sleep onset times to 'regress'
periodically, thereby increasing the probability that sleep onset
would occur during the nocturnal phase (between midnight and 7AM).
Finally, there were two occasions in which subjects skipped sleep
periods, the predicted onsets of which would have occurred during the
day phase (specifically, noon and 4PM), and instead delayed initiation
of the next sleep periods to nocturnal placements (11PM and 3AM,
respectively).


25
TABLE IV, Comparison of sleep variables as a function of night
(11 PM 7 AM) and day (7 AM 11 PM).
Overal1
Might
Day
Mean TST
27,99 h
15,34 h
12,65 h
(5.39)
(3.30)
(4.91)
Mean Sleep
2.99 h
5.52 h **
1.93 h
Period Length
(2.35)
(2.65)
(.98)
Total # Sleep
84
25
59
Episodes
Sleep/Wake
6.10 h
7.81 h *
5.25 h
Cycle Length
(2.68)
(3.26)
(1.74)
% Stage 0
3.59
3.50
3.64
(2.28)
(2.26)
(3.58)
% Stage 1
3.28
3.88
2.81
(1.87)
(2.79)
(2.06)
% Stage 2
58.41
56.37
60.80
(4.94)
(5.76)
(6.22)
% Stage 3
4.33
4.13
4.94
0.24)
(1.58)
(1.75)
% Stage 4
9.86
8.89
11.01
(4.54)
(5.37)
(4.28)
% Stage REM
20.47
23.13 **
17.54
(2.94)
(4.01)
(3.87)
Mean
** p <.001
t test, two-tailed
(SD)
* p < 01


SUBJECT NUMBER NUMBER OF
SUBJECTS
Night 2 12 24 36 48 60
Termination
HOURS OF BEDREST
Figure 1. Placement of sleep/waking episodes across 60 hours for individuals and group. Dark bars
represent sleep periods.


11
relinquish their typical uniphasic sleep placement in favor of a
more dispersed, polyphasic sleep system.
The present study attempted to address the hypothesis that
behavioral, cultural and environmental factors serve to entrain an
inherently ultradian sleep/waking rhythm to a uniphasic, 24-hour
framework by imposing upon the organism a series of behavioral
controls. In the experimental environment, such controls are
frequently mimicked by circumscribed sleep/waking schedules, feeding
regimens and testing procedures. These external controls "may create
incompatible responses which suppress the expression of (the)
rhythmic response or create such a 'noise' background as to not
permit measurement" of the hypothesized ultradian rhythm (Webb, 1378
p. 315). Therefore, by minimizing "incompatible responses" and "noise
background" the present study sought to "emancipate" the hypothesized
ultradian rhythm from control by a self-imposed and/or sociophysical-
ly mediated circadian framework. By so doing, it was proposed that
the presence of such a rhythm may be more readily observed and
measured. Specifically, subjects were required to maintain relative
ly static, basal levels of behavior over a period of almost three
days, during which time their sleep/waking behavior was the primary
variable of interest.


48
Clearly, the differences in the appearance and sequencing of
sleep stages between this study and the 2 studies of nap sleep with
which it has just been compared, are attributable in large part to
the difference in mean sleep episode length (3 hrs. vs 1 hr.). This is
especially true of REM sleep, if sleep dependency is assumed. Never
theless, the persistence and regularity with which sleep stages occurred
across the bedrest period is noteworthy. There were no significant
differences in the proportions of sleep stages between bedrest sleep
and Might 2 sleep. Nor, were there significant changes in sleep stage
percentages between the first and second 24 hours of bedrest. Finally,
only REM sleep percent varied significantly relative to day or night
onset times.
Weitzman et al. (1974) noted that the sleep episodes which they
observed were "clearly not minatures of the normal 8 hr. sleep pat
tern." Such a view has been echoed by most others who have studied
dispersed sleep and naps. Most of these studies have differed from
the present one, in that the sleep periods were typically experi
mentally circumscribed, rather than being placed on a 'self-demand1
schedule. Such a methodological difference may account for the
dissimilarity of results reported here.
Yet, Nakagawa (1980) also reported significant differences in
day episodes versus baseline (night) sleep, despite the fact
that the onset and termination of sleep periods was not regulated.
Significant increases in stage 1, significant decreases in stages 4
and REM and the presence of SOREMPS led Nakagawa to conclude that
"daytime sleep itself might be characterized as a special sleep at
the onset of the diurnal sleep-waking cycle," (p. 534).


31
TABLE VI, ~ Comparison of sleep variables, between first and second 24
hours, of bedrest period.
Overall
1st 24 HRS
2nd 24 HRS
Mean 1ST
27.99 h
14.46 h
** 10.97 h
(5.39)
(2.28)
(2.36)
Mean Sleep
2.99 h
3.54
2.82 h
Period Length
(2.35)
(1.92)
(1.44)
Total # Sleep
Episodes
84
39
33
Sleep/Wake
6.10 h
5.19 h
5,64 h
Cycle Length
(2.68)
(2.10)
(2.04)
% Stage 0
3.59 .
2.94
4.48
(2.28)
(2.23)
(4.25)
% Stage 1
3.28
3.03
3.74
(1.87)
(1.93)
(2.66)
% Stage 2
58.41
59.69
56.68
(4.94)
(4.03)
(8.50)
% Stage 3
4.33
4.15
4.66
(1.24)
(1.40)
(1.58)
% Stage 4
9.86
8.39
11.39
(4.54)
'(3.60)
(6.88)
% Stage REM
20.47
21.77
19.02
(2.94)
(2.56)
(4.82)
Mean
** p < .01
(SD)
t test5
two-tailed


32
In addition to day/night differences, there was als.o disparity
in temporal measures as a function of elapsed time within the bedrest
period. Table VI compares 1ST, sleep period duration, sleep/wakirig
cycle length, and sleep stage percentages recorded during the first
24 hours of bedrest with those recorded during the second 24 hours.
Total sleep time for the first 24 hours (14.5 h) was significantly
longer than TST for the second 24 hours (10,9 h) (t (16) = 3.19, p<
.01). While the decrease in TST was, in part attributable to a
reduction in the total number of sleep episodes during the second 24
hours (39 vs 33), a decline in the mean sleep period length across BP
(3.54 h vs 2.82 h) also contributed to the difference.
Despite the decrease in overall sleep period duration, the
sleep/waking cycle length remained essentially unchanged throughout
the bedrest period. For purposes of comparison, only those sleep/
waking cycles initiated and completed within one or the other 24 hour
periods were considered. The mean duration of sleep/waking cycles
occurring during the first 24 hours was 5.2 hours (SD = 2.1 h), while
the mean cycle length of those completed during the second 24 hours
was 5.6 hours (SD = 2.4 h).
There were no significant differences in sleep stage proportions
between the 24-hour intervals. However, trends were observed relative
to stages 4 and REM. Stage 4 increased (from 8.4 to 11.4'/) from the
first 24 hours to the next; this despite the fact that there were
more sleep episodes containing stage 4 in the first 24 hours (32 vs
27). Conversely, REM sleep decreased by over 10% from the first to
the second 24 hours.


BIOGRAPHICAL SKETCH
Scott Searcy Campbell was born February 25, 1952. He received
his B.A. degree, with a major in psychology, from Florida State
University in 1374. From September, 1972, to June, 1978, he attended
Montana State University from which he received the M.S. degree in
experimental psychology. In September, 1978, he continued his
graduate studies, under the supervision of Wilse B. Webb, at the
University of Florida. Upon completion of the Ph.D. degree, he
accepted a postdoctoral fellowship at Harvard Medical School.
60


CHAPTER ONE
INTRODUCTION
Sleep as a Biological Rhythm. That the sleep response is one of
a vast array of physiological functions which maintain a reliable
tendency to repeat at regular intervals has been demonstrated on
numerous occasions over the past two decades (see for example,
Aschoff, 1965; Richter, 1965; Symposia on Quantitative Biology,
1960; Webb, 1971; Webb and Agnew, !S74a). That the modulation of
the sleep/waking cycle is controlled by innate and autonomous
mechanisms seems, also, to have been well established.
In a free-running (time free) environment, subjects display the
continuation of a sleep/waking rhythm which deviates slightly from
the standard 24 hours (usually longer). The only plausible explana
tion for such a finding, according to Aschoff (1965, p. 1427) is
"that this rhythm is not imposed on the organism by the environment,
but is truly endogenous. We are dealing with a self-sustaining
oscillation which is free-running under constant conditions and has
its own inherent frequency."
The autonomous nature of the sleep/waking rhythm may also be
demonstrated in the free-running environment, relative to other
rhythms. The vast majority of rhythms which are in phase with the
sleep/waking cycle in the 24-hour environment become desynchronized
upon introduction into the time free environment (Aschoff, 1978;
Aschoff and Wever, 1976; Wever, 1973). An example of such
1


6
efficiency for subjects attempting a 3-hour cycle compared with 90%
efficiency for the baseline condition.
It has also been noted in dispersed, as well as displaced sleep,
that the intrasleep structure of the "new" episodes differs in several
details from that seen in normally occurring nocturnal sleep. For
example, Weitzinan et al. (1974) observed "early" REM periods (preceded
by less than 10 minutes stage 2 at sleep onset) and "sleep onset" REM
periods (preceded only by wakefulness or stage 1) in over half of all
REM episodes recorded over the 10 days of 3-hour sleep/wake scheduling.
Results of the displacement of subjects' typically nocturnal sleep into
day sleep include significant increases in the amount of stage 1 and
reductions in the amounts of stages 3 and 4 (Webb and Agnew, 1978).
Such findings lend further support to the view that these sleep epi
sodes differ substantially from endogenously modulated sleep periods.
It may be argued, however, that studies of sleep dispersion and
and displacement do not adequately address the question of the in
herent frequency of the sleep/waking cycle. Rather, such studies
more specifically test the entrainability of the presumed uni phasic
rhythm. Failure to effectively impose a shortened regimen upon a
hypothesized uniphasic rhythm does not necessarily rule out the exis
tence of an ultradian sleep/waking rhythm. Yet, data supporting such
a possibility are not extensive.
One type of evidence used to support the suggestion that the
human sleep/waking rhythm is inherently ultradian is provided by the
analysis of sleep patterns of individuals participating in Polar
expeditions. Lewis (1961; Lewis and Masterton, 1957) has suggested


5
usually greatest during the 'interval in which the subject would
normally be sleeping (Colquhoun, 1971; Weitzman et al., 1974).
It has further been reported (Weitzman et al., 1974) that "a
striking persistence of a circadian pattern of total sleep time"
(p. 1020) was exhibited throughout 10 days of entrainment to a 3-hour
day. In this study, subjects were encouraged to sleep 1 hour out of
each 3 hours. The authors reported that the 4 sleep periods beginning
at 3 AM, 6 AM,. 9 AM, and noon. accounted for 70% of the total sleep
time.
The decreased sleep efficiency associated with dispersed and
displaced sleep schedules has also been seen as lending support to
the notion of an inborn, nocturnally placed, uniphasic system of
sleep and wakefulness. Studies in which the normal, nocturnal sleep
period has been shifted to a time typically employed for work or other
waking activities, indicate that such sleep episodes are characterized
by significantly greater proportions of wakefulness (stage 0) after
initial sleep onset, as compared to baseline (normal) sleep (Webb,
Agnew, and Williams, 1971). Such findings suggest that the main
tenance of sleep is more difficult when the sleep period is shifted
from its typical, nocturnal placement. Further, as experimentally
circumscribed 'days' deviate from the normal 24 hours, sleep
efficiency (amount of time in bed spent asleep), becomes less.
Webb and Agnew (1975) found that subjects maintaining a sleep/waking
schedule of 3 hours sleep/6 hours wakefulness only achieved 80%
sleep efficiency when compared to baseline conditions (8 hours sleep/
16 hours wakefulness). Weitzman et al. (1974) reported 56% sleep


CHAPTER THREE
RESULTS
One of the '10 subjects terminated participation prior to the
end of the bedrest period. Since only 12 hours of bedrest were re
corded for this subject, the results presented here are those of the
9 subjects who completed the session.
The criterion used here to define a !lsleep period1' (i.e., 30
minutes of sleep, uninterrupted by more than 20 minutes of wakeful
ness) accounted for 91.3% of all sleep recorded. Eight sleep episodes
had durations of less than 30 minutes (mean = 7.25 min., range = 2
to 18 min.), and therefore were not considered "sleep periods" in the
following analyses.
Initial Sleep Period. The mean total sleep time (TST) for the
sleep period immediately preceding the onset of the bedrest period
(Night 2) was 8.72 hours, with a range of from 5.82 hours to 10.43
hours. Sleep stage percentages for the period were generally within
normal limits. However, as shown in Table 1, there was a significant
ly greater proportion of Stage 0, t (df = 39) = 2,32, p.<05 (all
levels of significance are based on a two-tailed test), and a
smaller percentage of stage 1, t(39) = 5.04, p.<001, and stage 3,
t(39) 2.20, p.<05, in this sample as compared with a normative
group of 16 males and 16 females of approximately the same age (mean =
24 years) previously recorded in the same laboratory (Williams, Agnew
and Webb, 1964; 1966). "Chopping" each record to an artifical maximum
15


TABLE II.
Individual and group statistics for the bedrest period (BP).
Bedrest Sleep Period Sleep/Wake
Subject #
Period
TST
Length
# Cycle Length
%Stg0
%Stgl
%Stg2
?Stg3
%Stg4
%REM
928
62.22h
37.31h
2.87h
13
4.70h
2.09
5.34
60.72
3.83
3.74
24.25
929
60.23
23.87
2.17
11
4.84
4.17
3.41
63.92
4.38
8.56
15.52
932
60.14
33.12
2.76
12
4.81
4.02
2.06
63.75
3.06
6.98
20.11
936
60.37
32.30
3.23
10
6.01
8.23
5.76
52.88
5.55
10.81
16.93
937
60.57
19.92
2.49
8
6.89
3.67
4.68
53.59
4.93
9.61
23.32
940
60.27
28.40
3.55
8
7.04
1.76
3.70
60.28
3.17
9.69
21.38
941
53.08
25.28
3.16
8
6.18
1.57
.13
50.29
6.78
20.60 .
20.60
945
57.90
26.82
2.98
9
6.34
5.64
1.05
61.01
3.28
9.86
19.11
946
54.25
24.90
4.98
5
13.59
1.20
3.60
59.29
3.94
8.89
23.06
Mean
58.80
27.99-
2.99
9.3
6.10
3.59
3.28
58.41
4.33
9.86
20.47
(SD)
(3.11)
(5.39)
(2.35)
(2.4)
(2.68)
2.28
(1.37)
(4.94)
(1-24)
(4.54)
(2.94)
ro
o


CHAPTER FOUR
DISCUSSION
The results of this study substantiate and further detail pre
vious reports that a reduction in the control of the sleep response
results in a 'loosening' of the uniphasic, nocturnally placed sleep
pattern typically observed in adult humans. At the same time, how
ever, the results underscore the tendency of sleep to be differen
tially weighted toward a definite and persistent nocturnal placement.
On the one hand, the data suggest a quadriphasic sleep pattern,
with the onset of sleep episodes occurring at approximately 6 hour
intervals throughout the bedrest period. While such a rhythmicity
proved to be highly variable within subjects, from one sleep period
to the next, the relatively small between-subjects variability em
phasizes the regularity of such a rhythm within the group. That the
number of sleep periods was proportionately placed, relative to day
and night, further accentuates the uniform distribution of sleep epi
sodes across the 24-hour day. The 'day' phase (7 AM to 11 PM) com
prised approximately 73% of the total bedrest period. Correspondingly,
70% (59) of the total number of sleep episodes recorded-across BP were
initiated during the day.
On the other hand, comparison of the mean duration of those day
sleep periods with the average length of sleep periods initiated at
night clearly illustrates the substantial orientation toward the
33


38
this view have speculated that an important function of the sensory and
perceptual systems is to maintain a waking level of arousal in the
reticular activating system. In studies comparable to the present one,
where conditions are maintained at a relatively constant level, and
sensory stimulation is not only static but also depressed relative to
normal levels, input from peripheral afferent pathways may be insuffi
cient to furnish the appropriate level of RAS activation. "In general,
appropriate sensory stimulation is needed to maintain an arousal state
in animals, but repetitive monotonous stimulation habituates the arousal
reaction . (Nakaguawa, 1980, p. 532). In short the, according to
this view, extended sleep is the result of neurological "boredom" and/or
habituation. Implied in such a view is the notion that the extended
sleep time observed under unrestricted or bedrest conditions is not
representative of natural sleep tendency (or requirement), but rather,
is simply one consequence of the organism's decreased requirement to
process environmental stimuli and respond accordingly (Heron, 1957).
This is analagous to Flanigan's (1971) suggestion that the extended
sleep observed in nonhuman subjects maintained under controlled and
highly favorable conditions (i.e., limited activity, plenty of food, no
predators) may reflect the upper limits of sleep that are organ!smically
possible, rather than that which is typical in the animal's natural
habitats.
In contrast, it has been suggested (see for example Webb and Agnew,
1975a) that the extended sleep lengths characteristic of unrestricted
sleep conditions are, in fact, "normal" and that typically reported
sleep durations of 7 to 8 hours may actually be the product of chronic
sleep deprivation. In support of such a notion is the finding that


51
to typical conditions under which animal sleep is studied. That a
polyphasic sleep pattern was exhibited under such conditions, therefore,
was not completely unexpected. Indeed, in 1972 Snyder speculated that
"even in the human we would probably find polyphasic sleep if we used
the same conditions that we use for animals," (p. 45).
While the monotony of the situation was quite likely instrumental
in the increase in total sleep time over the bedrest period, it seems
less probable that such "neurological boredom" would be expressed in
a cyclic manner. Yet, the possibility remains, as noted by Weitzman
(1972) with regard to the sleep cycle in cats. "In your experimental
situation, .... you are measuring in part the shift of pattern which
is taking place from a previous long-established light-dark pattern, in
addition to adaptation to a strange new environment. Are you measuring
the effect of that shift as well as the intrinsic rhythmicity in the
cat? If the measured rhythmicity were constant over, say, several
weeks, then it would support the concept that a basic rhythmicity was
being measured. However, if it tended to change over 1, 2, or 3 weeks
that would suggest that you are at least partially measuring an adapta
tion effect," (p. 201). Obviously, the question relative to human
sleep cannot be resolved by results of the present experiment.
It is suggested, though, that the conditions of the present study
represented a further advancement along a continuum of slackened
behavioral controls over the sleep process, the results of which was a
clearer manifestation of an inherent ultradian sleep/waking rhythm,
rather than the adaptation of the sleep response to a "strange new
environment." It is further suggested that such a rhythm, albeit in


24
superimposed across each subject's 'compressed1 bedrest period, in
which waking intervals were removed to obtain a continuous sleep per-
iod, to test the possible sleep dependency of the REM sleep cycle.
The initial onset of the superimposed cycle corresponded to the onset
of the last REM sleep period of Might 2.
The relationship between the superimposed REM cycle and the oc
currence of actual REM episodes was examined by calculating the mean
difference between each onset point of the hypothesized cycle and the
closest REM epoch within that sleep period. The average difference
between each onset of the superimposed cycle and the appearance of ari
actual REM epoch, across the intact bedrest period, was 15.8 minutes
(_SD = 3.7 min). The mean difference between the values for the 'com
pressed' BP was 14.5 minutes (SD = 4.7 min). The difference between
the two conditions was not significant.
To further test the sleep dependency of the cycle, the mean REM
cycle length of the 'compressed' BP, for the group, was determined.
Again, beginning the cycle with the onset of the last REM period of
Night 2, mean cycle length for the 'compressed' BP was 96.84 minutes
(SD 10.42 min). This is compared with a mean REM cycle length of
96.56 minutes (SD = 12.76 min) for Night 2.
Circadian Considerations. There were considerable differences
in several measures as a function of time of day. For comparative
purposes, waking and sleep periods were considered daytime events
if they were initiated between 7 AM and 11 PM. Nighttime events
were defined as those episodes with onsets between 11 PM and 7 AM.


CHAPTER TWO
METHODS
The subjects for this study were 10 young adults (5 males and
5 females) ranging in age from 18 to 25 years (mean = 20.4). Sub
jects were selected after questionnaire (Appendix 1) responses
revealed normal sleep patterns and the absence of chronic health
problems and/or acute illness.
Each subject slept for 2 nights in the laboratory immediately
prior to the beginning of 60 consecutive hours of bedrest. Between
the 2 nights in the laboratory subjects were permitted to maintain
their normal daily schedules with the stipulation that no naps be
taken and no drugs (including alcohol and caffeine) be ingested
during that period.
For each session, 2 subjects reported to the laboratory approxi
mately Vs hours prior to their normal bedtimes (about 11 PM) for
electrode placement. Five millimeter gold disc electordes were
applied over Cambridge electrode jelly and held in place by coll odian
saturated gauze strips. EEC and electrooculogram (EOG) were contin
uously monitored for each subject using a Grass Model VI electro-
physiograph. The EEG of each subject was recorded on three channels
between electrode sites Fpl F7, PI T5, and C3 A2. The EOG
was monitored on a separate channel by electrodes placed on the
external canthus of each eye.
12


17
of 8 hours for Night 2 (i, e., scoring no.more than the first 460
minutes of each record) did not significantly change the mean pro
portions of sleep stages as compared to 'intact' Night 2, but such a
procedure did reduce the proportions of stages 0 and 3 so that they
were not significantly different relative to the normative group.
Bedrest Period. Although timing of the bedrest period (BP) be
gan with the spontaneous awakening from Night 2 and the subject's
acknowledgement that (s)he felt awake and rested, later analysis of
the records of three subjects indicated that the 'spontaneous awaken
ings1 were less than 20 minutes in length (7, 9, and 11 minutes) and
were followed by substantial sleep episodes (mean = 2.11 hours). These
episodes, therefore, were included in the initial sleep period (Night
2), and as a result, the bedrest periods of the 3 subjects were some
what under 60 hours (mean = 55.32 hours). The mean duration of BP
for the entire group was 58.8 hours (range = 53.1 to 62.2 hours)/
Referring to Table 1 it can be seen that the sleep stage per
centages averaged from sleep periods occurring during BP differed
only slightly from those recorded during Night 2. There were non
significant increases in the proportions of stages 0, 1, and 2, while
nonsignificant decreases in SWS and REM sleep were observed as com
pared to Might 2. Similar trends (all significant, however) were
observed relative to the normative group, with the exception of
stage 1, which comprised a smaller percentage of TST across BP than
in the normative sample.
The distribution of sleep and wakefulness across Bp, for all
nine subjects, is presented in Figure 1, Figure 2 presents., in


49
In that study, Nakagawa noted a sleep/waking cycle of approxi
mately 4 hours, and suggested that a significant relationship seemed
to exist between this cycle and the scheduling of meals. The quadri-
phasic rhythm observed during the present study did not appear to be
the result of feeding schedules, since less than half (42%) of waking
periods were used for meals. Further, one subject (#932) fasted
for the duration of the experiment, yet maintained the least variable
sleep/waking cycle of the group.
Uni phasic or Ultradian? The results of the present study seem
to suggest, then, that the daytime sleep observed was not a "special"
sleep, but rather was the continuation of an ultradian sleep/waking
rhythm, which nevertheless maintained a definite tendency to be
weighted toward a nocturnal placement, relative to total sleep time.
The regularity with which sleep episodes were initiated across
the bedrest period also seems to support the notion, of an ultradian
pattern of sleep. The mean within-subject variability of the sleep/
waking cycles was 2.66 h (^SD). While the between-subject variability
was similar (SD = 2.73), most of the variability was contributed by
a single subject. This subject (#946, see Table II) exhibited a
sleep/waking cycle (13.59 h) which more than doubled the mean sleep/
waking cycle length of the other 8 subjects (5.85 h). Likewise, the
variability of the sleep/waking cycle of this subject (SD = 4.52)
exceeded the group mean by almost twice (SD = 2.43). Excluding this
subject's sleep/waking cycle length from the group mean reduces the
variability of the quadriphasic rhythm to less than an hour (SD =
57 min).


45
during baseline sleep. The same procedure was carried out in the present
study, and similar results were obtained. Using the-onset of the last
REM period of Night 2 as a starting point, the mean REM cycle length was
determined for subjects' 'compressed' bedrest periods, and compared to
the mean cycle length for Night 2. The mean cycle length for the
compressed period differed from the cycle length for Night 2 by less
than 1 minute (96.84 min vs 96.56 min). Such evidence is strongly
suggestive of a-sleep-dependent cycle.
In a second test, each subject's mean REM cycle length for Night 2
was superimposed on both his/her 'intact' bedrest period (i.e., includ
ing waking periods) and 'compressed' bedrest period. The mean deviation
of an actual REM occurrence, from the onset of the hypothesized cycle,
was used as a measure of goodness of fit.' The mean deviation across
the 'compressed' bedrest period was 14.5 minutes. In other words, the
actual occurrence of REM periods during BP differed by an average of
about 15 minutes, as compared to expected occurrences, when intervening
waking episodes were disregarded. Again, such a finding is supportive
of a sleep-dependent REM sleep cycle. However, a similar finding that
the mean deviation of actual REM sleep from hypothesized REM periods
across the 'intact' bedrest period (15.8 min), precludes the acceptance
of the sleep-dependence of REM sleep at the absolute exclusion of the
'real time' model.
Numerous studies of napping and dispersed sleep (Carskadon and
Dement, 1975: Globus, 1966; Moses et al., 1975; Nakagawa, 1980; Webb
et al., 1966) have reported perturbations in the mode of appearance of
REM sleep episodes, characterized by the onset of REM sleep within
minutes of the initiation of a sleep period. The reported frequencies


59
Weitzman, E. Periodicity in Sleep and Waking States in Chase M. H.
(Ed.), The Sleeping Brain, Perspectives in the Brain Sciences,
V'o'l. 1, p. 45, Brain Information Service/Brain Research Insti
tute, Los Angeles: UCLA, 1972.
Weitzman, E., Nogeire, C., Perlow, M., Fukushima, P., Sassin, J.,
McGregor, P., Gallagher, T. and Heilman, L. Effects of a Pro
longed. 3-Hour Sleep-Wake Cycle on Sleep Stages, Plasms Cortisol
Gowth Homone and Body Temperature in Man. J. Clin. Endocrinol.
Metab., 1974, 38, 1018-1029.
Wever, R. Internal Phase-Angle Differences in Human Circadian
Rhythms: Causes for Changes and Problems of Determinations.
Int. J. Chronobiology, 1973, J[, 371 -390.
White, R. Sleep Length and Variability: Measurement and Inter
relationships. PhD Dissertation, University of Florida, 1975.
Williams, R., Agnew, H. and Webb, W. Sleep Patterns in Young Adults
An EEG Study. Electroenceph. Clin. Neurophysiol., 1964,
17, 376-381'.
Williams, R., Agnew, H. and Webb, W. Sleep Patterns in Young Adult
Females: An EEG Study. Electroenceph. Clin. Neurophysiol.,
1966, 20, 264-266.


10
While studies employing a bedrest design may clearly eliminate
some behavioral controls left in tact by the free-running design
(most notably physical exercise), this procedure is also typically
characterized by experimentally circumscribed sleep periods as a
consequence of full schedules of psychological, psychophysiological
and performance testing (see for example, Ryback & Lewis, 1971;
Ryback, Lewis & Lessard, 1971).
One recently published study (Nakagawa, 1980) however, does
provide some insight into sleep/waking cycles in the absence of
cultural mediation and behavioral controls. In this study, subjects
were required to remain lying quietly in bed for 10 to 12 hours
immediately following a full night's sleep. The original aim of
the study was to "elucidate changes in daytime states of conscious
ness" by observing "the EEG patterns of (awake) subjects in restrained
postures under minimally changing experimental conditions" (p. 524).
Despite explicit instructions to remain awake during the 10 to 12
hour period, however, subjects experienced "an uncontrollable
desire to fall asleep," resulting in sleep/waking cycles of approxi
mately 4 hours across the experimental period. Again, it appears
that the uniphasic placement of the sleep process loosened in response
to a reduction in behavioral controls.
In summary, most designs employed in the investigations of
the human sleep/waking rhythm are inadequate for that purpose since
they include strict controls over waking and sleeping as a necessary
part of the procedure. However, results from studies in which
such controls are relaxed indicate that subjects have a tendency to


45
of occurrence of such sleep onset REM periods (SOREMPS) have ranged
from as few as 3 out of 107 REM episodes (Globus 1966) to as many as
21 of 34 REM episodes (Moses et al,, 1975). Marn et al. (1964) reported
the occurrence of no REM onset periods in the examination of 18 subjects
who napped in either the afternoon (about 1:30PM) or evening (about
7:30PM). Yet, the overall rate of occurrence of SOREMPS for 5 studies
of dispersed and nap sleep (Carskadon and Dement, 1975; Globus, 1966;
Moses et al,, 1975: Nakaga.ua, 1980: Weitzman et al., 1974) is rather
substantial at 39% of all REM periods recorded (range 2.8% 72%).
Since such events are almost never observed in normal nocturnal sleep,
SOREMPS have been considered to be indications of the overall disruption
of normal sleep structure associated with day sleep episodes and
dispersed sleep regimens (Weitzman et al., 1974; Moses et al., 1975).
In the present study, of 153 REM episodes recorded over the bedrest
period, only 3 were initiated at the beginning of a sleep period (i.e.,
before the appearance of 2 min of stage 2). The infrequency of such
occurrences suggests that the dispersed sleep of the present study may
be of a different sort than typically observed nap sleep or experimen
tally dispersed sleep.
Such a view is further supported by the regularity with which sleep
stages occurred across the bedrest period. Seventy percent of all- sleep
periods recorded during BP were full blown' sleep episodes (61% of day
sleep episodes); that is, all sleep stages were present. By comparison,
only 16% of the 560 opportunities for sleep in the Weitzman et al. study
contained all stages. Of the 80 naps examined by Moses et al. (1975),
27.5% contained both SWS and REM sleep. REM sleep occurred in 90% of
all sleep episodes during BP (86% of day sleep episodes), SWS occurred


THE EFFECTS OF EXTENDED BED REST ON
HUMAN SLEEP PATTERNS
BY
SCOTT S. CAMPBELL.
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY
OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1981


30
The possible presence of a free-running sleep/wake rhythm was.
examined by comparing the mean sleep onset time of Night 2 (baseline)
and the mean onset times of the longest sleep periods of Nights 3 and
4. The mean sleep onset times for the group exhibited a tendency to.
occur later on each succeeding nightsiightly under 1 hour later
on Night 3 (as compared to baseline) and 42 minutes later on Might
4 (as compared to Night 3). However, as shown in Table V only 2
subjects (929 and 937) demonstrated consistent free-running patterns
across the three nights. Four additional subjects delayed the onset
of their major nocturnal sleep onset on Night 4 (relative to Night 3).
Sleep stage percentages were not affected substantially as
a function of time of day. While there was slightly less stage 1
(2.8 vs 3.9%) and slightly more stage 2 (60.8 vs 56,4%) and stage 4
(11 vs 8.9%) during the daytime sleep periods, only the difference
in REM sleep percent was significant, with REM sleep comprising just-
over 23% of NS episodes and only 17.5% of DS periods (t (82) = 5.91,
pc.001). The difference in REM percent appeared to be a consequence
not only of the difference in sleep period length but also of the main
tenance of a circadian pattern of REMS across the the bedrest period,
as shown in Figure 4, The largest proportion of TST spent in REM
sleep occurred during those sleep episodes initiated between 4 AM and
8 AM. The proportion of REM sleep decreased across the next three 4-
hour time blocks until reaching a low of 11.57% of TST for sleep
episodes initiated between 4 PM and 8 PM. The percentage of REM
sleep then exhibited a steady increase across the subsequent 4-hour
time blocks (8 PM to 12 MID and 12 M to 4 AM).


52
highly dampened form, may also be reflected in the occasional naps
taken by normal adults (especially on weekends), the increased frequency
of naps associated with the loosened schedules of geriatric and college
populations, the sleep patterns of infants, and the polyphasic sleep
patterning noted in unstructured environments (i.e., Arctic expeditions).
And, I would suggest one additional manifestation of the ultradian
nature of the sleep response. Webb (1969) has noted that "(no one yet
has exploited, a foundation to study the 'siesta' patterns of the tropi
cal and subtropical cultures, although at various odd moments, I am
most tempted)," (p. 53). Results of the present study may suggest
that such a proposal, to examine cross-cultural diversity in the
cycling of rest and activity,.may deserve a better fate than paren
thetical consideration.


TABLE I. Comparisons of sleep stage characteristics for baseline (Night 2), bedrest period, and
normative group.
TST
%STG 0
%STG 1
%STG 2
%STG 3
%STG 4
%STG REM
kk
Normative
7.52 h
.95
5.62
48.37
7.32
14.70
23.02
(N=32)
(1.35)
(3.74)
(7.02)
(3.25)
(5.30)
(3-71)
Night 2
7.61 h
2.54
1.63+++
52.56
5.97
14.72
22.41
(1 chopped')
(-80)
(2.68)
(1.32)
(5.95)
(1.85)
(5.18)
(3.72)
Night 2
8.72 h
2.99+
1.78+++
53.05
5.47+
13.42
23.55
(1.58)
(2.54)
(1.14)
(6.52)
(1.83)
(5.08)
(3.55)
Bedrest
27.99 h
3.59++
3.28+
58.41+++
4.33+++
9.8 6+
r 20.47+
Period
(5.39)
(2.28)
(1-87)
(4.94)
(1.24)
(4.54)
(2.94)
Mean
(SD)
significance levels (
t test, two tailed):
different
from Normative
+
p < .05
. ** from 2
studies by
Williams et al. (1964; 1966)
++
+++
p < .01
p< .001
O


I certify that. I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality, as a dissertation for
the degree of Doctor of Philosophy.
%)
I Vj
v'6
y.-
NT
U/
Wilse B. Webb, Chairman
Graduate Research Professor of
Psychology
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality, as a dissertation for
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality
the degree of Doctor of Philosophy.
as a dissertation for
Donald A. Dewsbury
Professor of Psychology
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality, as a dissertation for
the degree *of Doctor of Philosophy.
Charles M. Levy
Professor' of Psychology


36
1974a). The mean sleep length for Night 2 is also in agreement with
self-reported extensions in weekend sleep lengths (about 8.5 hrs).
The mean total sleep time reported here is, however, almost an
hour less than that of young adults (9.6 hrs) studied by Verdone (1968)
and Webb and Agnew (1975a). It seems likely that the mean total sleep
time for Night 2, like typically reported weekend sleep, was a conse
quence of the relaxation of self-imposed controls on sleep length
(e.g., alarm clocks), rather than an expression of the complete satis
faction of sleep 'need1. Several subjects in the present study were
rather impatient to terminate Night 2 and to initiate the bedrest
period. This is illustrated by the case of one subject who reported,
during a brief (3 min) awakening, that she felt rested and had "had
enough sleep." The report came after approximately 2 hours had elapsed
in Night 2. Clearly, such eagerness to 'get on with it' may have
resulted in the termination of sleep prior to the spontaneous ending of
sleep due to satiety.
A more accurate measure of maximum sleep length may be derived
from the examination, across the bedrest period, of sleep durations per
24 hours. Beginning with the termination of Night 2 (typically around
8AM) the first 24-hour period was characterized by an average of 14.5
hours of sleep, the second 24 hours contained an average of almost
11 hours of sleep. Such figures are consistent with previously re
ported 24-hour sleep amounts. For example, Aserinsky (1969) instructed
subjects to "make every attempt to sleep" while confined to bed for
30 hours (interrupted only twice for meals). During the first 24 hours
(midnight to midnight) subjects obtained an average of 14.4 hours of
sleep. During the 24-hour period beginning at 7:30AM, after a full


43
However, it seems somewhat unparsiinonious to have to employ both
versions to satisfactorily explain the data in terms, of immobilization.
In the present study, SWS comprised a smaller proportion of TST during
BP than during the baseline night. The hypothesis proposed by Adey
et al. is useful in explaining such results. Yet, within the bedrest
period, there was an increase in the percentage of SWS from the first
to the second 24 hours. These findings clearly require the Ryback and
Lewis version.
Several observations seem to support an alternative explanation
for the increase in SWS across the bedrest period, however. It has
been shown (webb and Agnew- 1971, p. 1354) that "longer periods of
wakefulness before sleep result in greater amounts of stage 4 sleep
in the first 3 hours of sleep." In the present study, the mean inter
val between sleep episodes increased by some 70% (from 1.65 hrs. to
2.82 hrs.) from the first to the second 24-hour period. At the same
time, the proportion of SWS increased almost 25% (from 12.9% to 16%).
By the same token, periods of wakefulness between sleep episodes
averaged about 1 hour longer during the day than at night (3.32 hrs. vs
2.29 hrs.). Correspondingly, SWS comprised a slightly greater propor
tion of day sleep than of night sleep episodes (16.2% vs 13.2%). The
relationship between prior wakefulness and subsequent amounts of SWS
seems apparent.
Such an association would also account for the decline of SWS
from Night 2 to the bedrest period. The amount of wakefulness preceed-
ing the baseline night (14 to 17 hrs.) far exceeded the duration of
waking intervals characteristic of BP. Consequently, the subsequent
sleep periods reflected this difference by exhibiting differential


THE EFFECTS OF EXTENDED BED REST ON
HUMAN SLEEP PATTERNS
BY
SCOTT S. CAMPBELL.
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY
OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1981

Copyright 1981
by
Scott S. Campbell

ACKNOWLEDGEMENTS
The author would like to thank the members of his supervisory
committee for their assistance and criticism during the preparation
of this dissertation. The author would especially like to express
his deep appreciation to Dr. Wilse B. Webb whose encouragement,
guidance, time, and patience provided a constant source of motivation
and inspiration. Thanks likewise go to Briggs, Lewis, Alton, and
Gregg whose technical assistance and friendship were invaluable.
And final thanks to my parents for their support, both moral and
financial, without which the completion of this dissertation would
have been impossible.

TABLE OF CONTENTS
CHAPTER PAGE
ACKNOWLEDGEMENTS .......... iii
ABSTRACT vii
ONE INTRODUCTION 1
Sleep as a Biological Rhythm ........ 1
Uniphasic or Ultradian ? 4
TWO METHODS 12
THREE RESULTS 15
Initial Sleep Period .... 15
Bedrest Period 17
Circadian Considerations 24
FOUR DISCUSSION
Total Sleep Time 35
Uniphasic or Ultradian? 49
APPENDIX
SLEEP QUESTIONNAIRE 53
REFERENCES 55
BIOGRAPHICAL SKETCH
60

LIST OF TABLES
TABLE
PAGE
I
Comparisons of sleep stage characteristics for
baseline (Night 2), bedrest period, and norma
tive group.
16
II
Individual and group statistics for the bedrest
period (BP)
20
III
Distribution of bedrest sleep periods by the
presence or absence of REM and Stage 4 sleep.
21
IV
Comparison of sleep variables as a function of
night (11 PM 7 AM) and day (7 AM 11 PM).
25
V
Sleep onset times of longest sleep episodes for
Night 2 (baseline), 3, and 4.
28
VI
Comparison of sleep variables between first and
second 24 hours of bedrest period.
31
v

LIST OF FIGURES
FIGURE PAGE
"1 Placement of sleep/waking episode across 60 "18
hours for individuals and group.
2 Example of sleep across BP (#932). 19
3 Total number of subjects asleep, for each hour 26
across the bedrest period.
4. Proportions of TST spent in REM sleep, as a 29
function of time of day.
vi

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
THE EFFECTS OF EXTENDED BED REST ON
HUMAN SLEEP PATTERNS
By
Scott S. Campbell
August 1581
Chairman: Wilse B. Webb
Major Department: Department of Psychology
That the sleep response is one of numerous physiological functions
which may be classified as a biological rhythm is well-established.
Less well-founded, despite its popularity, is the notion that the
sleep of adult humans maintains an inherent tendency to occur mono-
phasically within each 24 hours, typically during the nocturnal por
tion of the 24-hour day.
This d issertation addresses the hypothesis that self-imposed and/'
or sociophysically mediated behavioral controls have a significant
influence on the manifestation of such a uni phasic sleep pattern,
and that attenuation of behavioral controls may result in a "loosening11
of the circadian framework of typical adult sleep.
Each of 9 young adults was electrographically recorded over a
period of 60 hours, during which behavior was restricted primarily to
lying quitely in bed; i.e., subjects were not permitted to read, write,
vn

listen to music, watch TV, exercise, etc. Heals were served at random
intervals, and subjects had no knowledge of time of day throughout, the
bedrest period (BP).
The data indicate that a reduction in behavioral controls, which
typically influence the sleep process, results in substantial altera
tions in the placement and periodicity of sleep. Sleep episodes became
evenly dispersed across the 24-hour day, resulting in an approximate 6
hour sleep/waking cycle.
While the average sleep period duration decreased, mean total
sleep time for the bedrest period (28 h) exceeded, by more than 40%,
that which would be expected under the assumption of a uni phasic sleep/
waking rhythm. The continued tendency toward a nocturnal weighting of
total sleep was clear, however, since 55% of TST for BP occurred during
the night (11PM to 7AM).
Within sleep structure (e.g., sleep stage percentages, mode of
appearance) was affected only slightly by the "loosening'-' of the sleep
response. REM sleep retained its tendency to occur differentially
across the 24-hour day; stage 4 sleep appeared to maintain its rela
tionship to prior wakefulness.
It is suggested that the sleep/waking patterns observed in this
study may be further manifestations of an inborn (albeit, highly,
dampened) ultradian sleep/waking rhythm.
vii i

CHAPTER ONE
INTRODUCTION
Sleep as a Biological Rhythm. That the sleep response is one of
a vast array of physiological functions which maintain a reliable
tendency to repeat at regular intervals has been demonstrated on
numerous occasions over the past two decades (see for example,
Aschoff, 1965; Richter, 1965; Symposia on Quantitative Biology,
1960; Webb, 1971; Webb and Agnew, !S74a). That the modulation of
the sleep/waking cycle is controlled by innate and autonomous
mechanisms seems, also, to have been well established.
In a free-running (time free) environment, subjects display the
continuation of a sleep/waking rhythm which deviates slightly from
the standard 24 hours (usually longer). The only plausible explana
tion for such a finding, according to Aschoff (1965, p. 1427) is
"that this rhythm is not imposed on the organism by the environment,
but is truly endogenous. We are dealing with a self-sustaining
oscillation which is free-running under constant conditions and has
its own inherent frequency."
The autonomous nature of the sleep/waking rhythm may also be
demonstrated in the free-running environment, relative to other
rhythms. The vast majority of rhythms which are in phase with the
sleep/waking cycle in the 24-hour environment become desynchronized
upon introduction into the time free environment (Aschoff, 1978;
Aschoff and Wever, 1976; Wever, 1973). An example of such
1

2
desynchronization is that which occurs between the temperature and
sleep/waking rhythms. When both rhythms are entrained to 24 hours,
the maximum temperature typically occurs just prior to retiring, and
the minimum temperature occurs toward the end of the sleep period.
However, under free-running conditions, maxima and minima of rectal
temperatures become independent of sleep onset and termination.
In addition, the sleep/waking rhythm may be entrained to arti
ficial light/dark regimens while typically in-phase rhythms maintain
their free-running periodicities. For example, throughout 10 days
of entrainment to a 3 hour sleep/waking cycle, subjects' temperatures
continued to exhibit characteristic circadian periodicities (Weitzman,
Nogeire, Perlow, Fukushima, Sassin, McGregor, Gallagher, and
Heilman, 1974). Such data "contradict any hypothesis of a simple
'cause-and-effect' relationship between the activity-rest cycle and
the rhythm of body temperature, and suggest a system of two coupled
oscillations . with a mutual phase relationship which depends on
the conditions" (Aschoff, 1978; p. 740).
Further evidence of the sleep/waking cycle's autonomy relative
to other biological rhythms may be seen in the results of sleep
deprivation studies. In the absence of sleep (for more than 70 hours)
measures of performance, temperature and endocrine function all
showed continuation of endogenous ciradian rhythms (Murry, Williams,
and Lubin, 1958; Froberg, Karlsson, Levi, and Lidberg, 1975, as
cited by Aschoff, 1978).
Examinations of the maturation of the human sleep/waking rhythm
have also provided evidence for its innate and autonomous nature.

3
After observing the development of an initial 25-hour periodicity in
an infant maintained on a self-demand sleep-wakefulness and feeding
schedules Kleitman and Engelmann (1953) concluded that "there may be
a 'natural' rhythm which only has to be adjusted to the 24 hour
alternation of night and day" (p. 280). Other researchers have also
observed the development of adult-like, primarily monophasic,
nocturnally placed sleep under similar conditions of self-demand
feeding (Parmelee, Wenner, and Schulz, 1964; Salzarulo, Fagioli,
Salomon, Ricour, Raimbault, Ambrosi, Cicchi, Duhamel, and Rigoard,
1980).
While the sleep/waking rhythm clearly appears to be inborn and
independently regulated, it may, at the same time be entrained by
exogenous Zeitgebers, to frequencies other than those characteristic
of the free-running environment. It has been demonstrated that
human subjects are capable of maintaining sleep/waking cycles of
various lengths, quite efficiently within a range of about 20 to 27
hours, as the result of entrainment to appropriate 1 ight/dark regimens.
It has been further shown that the sleep/waking rhythm characteristic
of a free-running environment may be regulated by "rigid control of
the sleep portion of the cycle . particularly of the wakeup time"
(Webb and Agnew, 1974b; p. 701).
Several investigators also point to the developmental literature
as a source of evidence for the entrairiability of the sleep/waking
rhythm (Ellingson, 1975; Gesell and Armatruda, 1945; Kleitman, 1963;
Sander, Stechler, Burns, and Julia, 1970; Sterman, 1979). As
expressed by Sterman (p. 213), "the development of sleep is a postnatal

4
phenomenons subject both to the dictates of physiological maturation
and the influences of environment. Such factors as geography and
social custom entrain the physiological substrates of the sleep-waking
cycle to determine the behavioral patterns which will come to
characterize these states in a given culture."
Indeed, the very fact that adult humans maintain a 24-hour
sleep/waking periodicity, in spite of an inherent rhythm with a
frequency longer than 24 hours, is clear evidence of the entrain-
ability of the rhythm.
It appears then, that the oscillation between sleep and wake
fulness, characteristic of adult human subjects, meets the criteria
necessary for its designation as a biological rhythm. Sleep is an
inborn, independently regulated, exogenously entrainable, regularly
occurring response.
Uni phasic or Ultradian? While sleep and wakefulness clearly
comprise a physiological system which may be classified as a bio
logical rhythm, determination of the rhythm's inherent periodicity
remains less clear, since few studies have been conducted which
adequately address this point. It is generally assumed that the
sleep/waking cycle is an uniphasic one. In addition to the obvious
support for such a notion, i.e., that adult humans typically obtain
sleep in a single, nocturnal block, results from studies in which the
normal sleep period has been dispersed or eliminated have also been
cited as evidence for this view. In such experiments, performance
measures are usually lowest and subjective reports of sleepiness are

5
usually greatest during the 'interval in which the subject would
normally be sleeping (Colquhoun, 1971; Weitzman et al., 1974).
It has further been reported (Weitzman et al., 1974) that "a
striking persistence of a circadian pattern of total sleep time"
(p. 1020) was exhibited throughout 10 days of entrainment to a 3-hour
day. In this study, subjects were encouraged to sleep 1 hour out of
each 3 hours. The authors reported that the 4 sleep periods beginning
at 3 AM, 6 AM,. 9 AM, and noon. accounted for 70% of the total sleep
time.
The decreased sleep efficiency associated with dispersed and
displaced sleep schedules has also been seen as lending support to
the notion of an inborn, nocturnally placed, uniphasic system of
sleep and wakefulness. Studies in which the normal, nocturnal sleep
period has been shifted to a time typically employed for work or other
waking activities, indicate that such sleep episodes are characterized
by significantly greater proportions of wakefulness (stage 0) after
initial sleep onset, as compared to baseline (normal) sleep (Webb,
Agnew, and Williams, 1971). Such findings suggest that the main
tenance of sleep is more difficult when the sleep period is shifted
from its typical, nocturnal placement. Further, as experimentally
circumscribed 'days' deviate from the normal 24 hours, sleep
efficiency (amount of time in bed spent asleep), becomes less.
Webb and Agnew (1975) found that subjects maintaining a sleep/waking
schedule of 3 hours sleep/6 hours wakefulness only achieved 80%
sleep efficiency when compared to baseline conditions (8 hours sleep/
16 hours wakefulness). Weitzman et al. (1974) reported 56% sleep

6
efficiency for subjects attempting a 3-hour cycle compared with 90%
efficiency for the baseline condition.
It has also been noted in dispersed, as well as displaced sleep,
that the intrasleep structure of the "new" episodes differs in several
details from that seen in normally occurring nocturnal sleep. For
example, Weitzinan et al. (1974) observed "early" REM periods (preceded
by less than 10 minutes stage 2 at sleep onset) and "sleep onset" REM
periods (preceded only by wakefulness or stage 1) in over half of all
REM episodes recorded over the 10 days of 3-hour sleep/wake scheduling.
Results of the displacement of subjects' typically nocturnal sleep into
day sleep include significant increases in the amount of stage 1 and
reductions in the amounts of stages 3 and 4 (Webb and Agnew, 1978).
Such findings lend further support to the view that these sleep epi
sodes differ substantially from endogenously modulated sleep periods.
It may be argued, however, that studies of sleep dispersion and
and displacement do not adequately address the question of the in
herent frequency of the sleep/waking cycle. Rather, such studies
more specifically test the entrainability of the presumed uni phasic
rhythm. Failure to effectively impose a shortened regimen upon a
hypothesized uniphasic rhythm does not necessarily rule out the exis
tence of an ultradian sleep/waking rhythm. Yet, data supporting such
a possibility are not extensive.
One type of evidence used to support the suggestion that the
human sleep/waking rhythm is inherently ultradian is provided by the
analysis of sleep patterns of individuals participating in Polar
expeditions. Lewis (1961; Lewis and Masterton, 1957) has suggested

7
that given a situation in which adherence to customary circadian
scheduling is less stringently controlled, a more accurate reflection
of the inborn sleep/waking rhythm may emerge.
In the Polar regions the photoperiod is not diurnal, but seasonal
Further, expeditions into these regions typically maintain little
resemblance to an organized community, in that there are few circum
scribed daily routines, no specific feeding schedules, few timetables
(Lewis, 1961). Sleep records from 5 Polar campaigns suggest that the
organization of sleep/waking patterns was a direct function of the
stringency with which the "community" was structured. For example,
members of the British North Greenland Expedition (1952-54) were
allowed to sleep as long as they liked, at virtually any time they had
the inclination to do so. Under such conditions, sleep durations per
24 hours did not increase substantially over baseline amounts. Yet.
the length and placement of individual sleep episodes showed con
siderable deviations from members' normal sleep habits recorded in
England. The tendency toward polyphasic sleep placement was most
extreme during the three months of continuous light or darkness.
"During these periods . men used any of the 24 clock-hours for
sleep. The winter months were characterized by many interruptions to
sleep and by the taking of naps" (Lewis, 1961, p, 325). Although
no data were presented relative to the r'nythmicity of these multiple
sleep episodes, it appears that the uni phasic "nature" of the sleep
process loosened somewhat under conditions of limited environmental
cues and negligible behavioral and cultural restraints.

8
Within the more structured hours of daily life, it has been
suggested that the ultradian nature of sleep and wakefulness may be
manifest in the form of naps (Webb, 1978). Webb's proposal that naps
are part of an inborn sleep/waking rhythm is based on the contention
that certain characteristics of these intermittent sleep episodes
match features which are common to biological rhythms in general.
Specifically, it is maintained that naps are: 1) temporally repeti
tive, 2) species specific, 3) often developmental, 4) innate and
unlearned, 5) endogenous, and 6) adaptive. While evidence seems to
support the first four characteristics, and only post hoc specula
tion can address the sixth, Webb points out that experimental
support for the claim that naps are endogenously mediated is
extremely limited, and that further data are "likely to be slow
forthcoming." This is due primarily to the fact that virtually
all procedures utilized for the examination of presumed biological
rhythms have prohibited other-than-circumscribed sleep episodes
as a consequence of experimental design considerations. Such experi
ments typically focus on the interactions of major components of
sleep, waking and performance. As such, "the very nature of naps
in these designs are uncontrolled and interfering sources of experi
mental variance" (Webb, 1978 p. 316). Two primary designs of this
type are the free-running and bedrest procedures.
It has been noted that the sleep/waking rhythms maintained by
subjects in free-running environments provide excellent support for
the notion that such rhythms are innately and independently controll
ed. Under such conditions, time cues are removed from the

9
environment, but other behavioral controls are typically left
unaltered. That is, subjects are asked to lead a "normal" life,
including eating three meals a day in normal sequence, having the
opportunity for exercise and other activities such as reading,
writing, watching TV, etc., and maintaining circumscribed sleep
schedules (no naps). As a result, these subjects have a tendency
to carry on day-to-day life in a manner to which they have grown
accustomed over the years. Clearly, such circumstances may strongly
bias behavior toward a circadian schedule, perhaps entraining an
inherently ultradian rhythm to an environmentally and culturally
modulated circadian framework. These factors notwithstanding,
the sleep patterns displayed by subjects in time free environments
have been seen as providing evidence not only for the biological
basis of the sleep/waking rhythm, but for its circadian (uniphasic)
nature, as well.
The latter inference seems tenuous, at best. For, when controls
on the number and placement of sleep episodes are reduced in free-
running environments, evidence seems to indicate that the "normal
sleep pattern" (single sleep period of from 7 to 10 hours with
occasional napping) is modified. Webb and Agnew (1974a) reported
that under such circumstances "a significant increase-in both long
and short periods of sleep, compared to the 24-hour environment" (p.
620) resulted. While neither the placement nor the periodicity of
the increased episodes was analyzed, the authors concluded that such
changes in the sleep response showed the extent to which the
uniphasic sleep pattern is disrupted by reduced controls on sleep
placement.

10
While studies employing a bedrest design may clearly eliminate
some behavioral controls left in tact by the free-running design
(most notably physical exercise), this procedure is also typically
characterized by experimentally circumscribed sleep periods as a
consequence of full schedules of psychological, psychophysiological
and performance testing (see for example, Ryback & Lewis, 1971;
Ryback, Lewis & Lessard, 1971).
One recently published study (Nakagawa, 1980) however, does
provide some insight into sleep/waking cycles in the absence of
cultural mediation and behavioral controls. In this study, subjects
were required to remain lying quietly in bed for 10 to 12 hours
immediately following a full night's sleep. The original aim of
the study was to "elucidate changes in daytime states of conscious
ness" by observing "the EEG patterns of (awake) subjects in restrained
postures under minimally changing experimental conditions" (p. 524).
Despite explicit instructions to remain awake during the 10 to 12
hour period, however, subjects experienced "an uncontrollable
desire to fall asleep," resulting in sleep/waking cycles of approxi
mately 4 hours across the experimental period. Again, it appears
that the uniphasic placement of the sleep process loosened in response
to a reduction in behavioral controls.
In summary, most designs employed in the investigations of
the human sleep/waking rhythm are inadequate for that purpose since
they include strict controls over waking and sleeping as a necessary
part of the procedure. However, results from studies in which
such controls are relaxed indicate that subjects have a tendency to

11
relinquish their typical uniphasic sleep placement in favor of a
more dispersed, polyphasic sleep system.
The present study attempted to address the hypothesis that
behavioral, cultural and environmental factors serve to entrain an
inherently ultradian sleep/waking rhythm to a uniphasic, 24-hour
framework by imposing upon the organism a series of behavioral
controls. In the experimental environment, such controls are
frequently mimicked by circumscribed sleep/waking schedules, feeding
regimens and testing procedures. These external controls "may create
incompatible responses which suppress the expression of (the)
rhythmic response or create such a 'noise' background as to not
permit measurement" of the hypothesized ultradian rhythm (Webb, 1378
p. 315). Therefore, by minimizing "incompatible responses" and "noise
background" the present study sought to "emancipate" the hypothesized
ultradian rhythm from control by a self-imposed and/or sociophysical-
ly mediated circadian framework. By so doing, it was proposed that
the presence of such a rhythm may be more readily observed and
measured. Specifically, subjects were required to maintain relative
ly static, basal levels of behavior over a period of almost three
days, during which time their sleep/waking behavior was the primary
variable of interest.

CHAPTER TWO
METHODS
The subjects for this study were 10 young adults (5 males and
5 females) ranging in age from 18 to 25 years (mean = 20.4). Sub
jects were selected after questionnaire (Appendix 1) responses
revealed normal sleep patterns and the absence of chronic health
problems and/or acute illness.
Each subject slept for 2 nights in the laboratory immediately
prior to the beginning of 60 consecutive hours of bedrest. Between
the 2 nights in the laboratory subjects were permitted to maintain
their normal daily schedules with the stipulation that no naps be
taken and no drugs (including alcohol and caffeine) be ingested
during that period.
For each session, 2 subjects reported to the laboratory approxi
mately Vs hours prior to their normal bedtimes (about 11 PM) for
electrode placement. Five millimeter gold disc electordes were
applied over Cambridge electrode jelly and held in place by coll odian
saturated gauze strips. EEC and electrooculogram (EOG) were contin
uously monitored for each subject using a Grass Model VI electro-
physiograph. The EEG of each subject was recorded on three channels
between electrode sites Fpl F7, PI T5, and C3 A2. The EOG
was monitored on a separate channel by electrodes placed on the
external canthus of each eye.
12

13
At the completion of electrode placement each subject retired
to a private, electrically shielded, sound attenuated, temperature
controlled room. Each subject's record was scored in one-minute
epochs, following the methods described by Agnew and Webb (1972a).
Cross-scoring of random 2-3 hour segments of the sleep portions of
each record revealed a between-scorer reliability of greater than
90%.
Sleep onset was determined by the appearance of the first
spindle or K-complex (Agnew & Webb, 1972b). A sleep period was
considered as such when sleep continued, uninterrupted by more than
20 minutes of wakefulness, for at least 30 minutes. A waking period
was defined as Stage 0 EEG for at least 20 minutes.
Timing of the 60 hours of bedrest began with the spontaneous
awakening of each subject from his/her second night of laboratory
sleep and the subject's verbal report that he or she felt rested
and had had enough sleep. At this time subjects were given an op
portunity to eat breakfast and use the bathroom. No intentional
time cues were given the subjects upon awakening nor at any other
time throughout the bedrest period, with the exception that each
subject was told when he or she was "past the halfway point" at a
random time after 30 hours had past. However, uncontrollable, gen
eral time cues from exogenous sources (eg., nearby building construc
tion, general building noise during working hours, homecoming fes
tivities during one session (subjects 945 & 946)) were present in
varying degrees from session to session.
During the 60 hours of bedrest, subjects were allowed minimal
exogenous stimulation. Activities such as reading, writing, listening

14
to music or watching television were prohibited. Conversations'
during the period were limited to brief dialogues with the experi
menter at meal times. No instructions were given relative to when
or when not to sleep. Subjects were asked to lie as quietly as pos
sible, but were permitted to change positions in bed when inclined
to do so. Closed circuit TV monitors were employed to insure adher
ence to the experimental instructions. Subjects were permitted to
go to the bathroom (in room) at their convenience and were allowed
to sit up in bed during meals.
Meals were served unsystematically (always during an ongoing
waking period, as determined by the EEC) and consisted of a choice
of several types of sandwiches, vegetable dishes, desserts and non-
caffeinated beverages. A maximum of ^ hour was permitted for the
completion of each meal. Several types of fruit and a pitcher of
ice water were also available to each subject ad libitum.
Illumination in each room was provided by a 60 watt lamp placed
on a table by the subject's bed, and control of light and control of
light and darkness was at the discretion of each subject.

CHAPTER THREE
RESULTS
One of the '10 subjects terminated participation prior to the
end of the bedrest period. Since only 12 hours of bedrest were re
corded for this subject, the results presented here are those of the
9 subjects who completed the session.
The criterion used here to define a !lsleep period1' (i.e., 30
minutes of sleep, uninterrupted by more than 20 minutes of wakeful
ness) accounted for 91.3% of all sleep recorded. Eight sleep episodes
had durations of less than 30 minutes (mean = 7.25 min., range = 2
to 18 min.), and therefore were not considered "sleep periods" in the
following analyses.
Initial Sleep Period. The mean total sleep time (TST) for the
sleep period immediately preceding the onset of the bedrest period
(Night 2) was 8.72 hours, with a range of from 5.82 hours to 10.43
hours. Sleep stage percentages for the period were generally within
normal limits. However, as shown in Table 1, there was a significant
ly greater proportion of Stage 0, t (df = 39) = 2,32, p.<05 (all
levels of significance are based on a two-tailed test), and a
smaller percentage of stage 1, t(39) = 5.04, p.<001, and stage 3,
t(39) 2.20, p.<05, in this sample as compared with a normative
group of 16 males and 16 females of approximately the same age (mean =
24 years) previously recorded in the same laboratory (Williams, Agnew
and Webb, 1964; 1966). "Chopping" each record to an artifical maximum
15

TABLE I. Comparisons of sleep stage characteristics for baseline (Night 2), bedrest period, and
normative group.
TST
%STG 0
%STG 1
%STG 2
%STG 3
%STG 4
%STG REM
kk
Normative
7.52 h
.95
5.62
48.37
7.32
14.70
23.02
(N=32)
(1.35)
(3.74)
(7.02)
(3.25)
(5.30)
(3-71)
Night 2
7.61 h
2.54
1.63+++
52.56
5.97
14.72
22.41
(1 chopped')
(-80)
(2.68)
(1.32)
(5.95)
(1.85)
(5.18)
(3.72)
Night 2
8.72 h
2.99+
1.78+++
53.05
5.47+
13.42
23.55
(1.58)
(2.54)
(1.14)
(6.52)
(1.83)
(5.08)
(3.55)
Bedrest
27.99 h
3.59++
3.28+
58.41+++
4.33+++
9.8 6+
r 20.47+
Period
(5.39)
(2.28)
(1-87)
(4.94)
(1.24)
(4.54)
(2.94)
Mean
(SD)
significance levels (
t test, two tailed):
different
from Normative
+
p < .05
. ** from 2
studies by
Williams et al. (1964; 1966)
++
+++
p < .01
p< .001
O

17
of 8 hours for Night 2 (i, e., scoring no.more than the first 460
minutes of each record) did not significantly change the mean pro
portions of sleep stages as compared to 'intact' Night 2, but such a
procedure did reduce the proportions of stages 0 and 3 so that they
were not significantly different relative to the normative group.
Bedrest Period. Although timing of the bedrest period (BP) be
gan with the spontaneous awakening from Night 2 and the subject's
acknowledgement that (s)he felt awake and rested, later analysis of
the records of three subjects indicated that the 'spontaneous awaken
ings1 were less than 20 minutes in length (7, 9, and 11 minutes) and
were followed by substantial sleep episodes (mean = 2.11 hours). These
episodes, therefore, were included in the initial sleep period (Night
2), and as a result, the bedrest periods of the 3 subjects were some
what under 60 hours (mean = 55.32 hours). The mean duration of BP
for the entire group was 58.8 hours (range = 53.1 to 62.2 hours)/
Referring to Table 1 it can be seen that the sleep stage per
centages averaged from sleep periods occurring during BP differed
only slightly from those recorded during Night 2. There were non
significant increases in the proportions of stages 0, 1, and 2, while
nonsignificant decreases in SWS and REM sleep were observed as com
pared to Might 2. Similar trends (all significant, however) were
observed relative to the normative group, with the exception of
stage 1, which comprised a smaller percentage of TST across BP than
in the normative sample.
The distribution of sleep and wakefulness across Bp, for all
nine subjects, is presented in Figure 1, Figure 2 presents., in

SUBJECT NUMBER NUMBER OF
SUBJECTS
Night 2 12 24 36 48 60
Termination
HOURS OF BEDREST
Figure 1. Placement of sleep/waking episodes across 60 hours for individuals and group. Dark bars
represent sleep periods.

Sleep Slage
/
inI ^TJITLT
i u-r
I2M
6AM
f*
6PM
I2M
^ui
i r
pAjitl
r^i
UU
l:!r i
6AM
I2N
6 PM
12M
6AM
11
I2N
6PM
gure 2.
Example of sleep staging across BP (subject 932).
are not shown.
Awakenings of less than 20 minutes

TABLE II.
Individual and group statistics for the bedrest period (BP).
Bedrest Sleep Period Sleep/Wake
Subject #
Period
TST
Length
# Cycle Length
%Stg0
%Stgl
%Stg2
?Stg3
%Stg4
%REM
928
62.22h
37.31h
2.87h
13
4.70h
2.09
5.34
60.72
3.83
3.74
24.25
929
60.23
23.87
2.17
11
4.84
4.17
3.41
63.92
4.38
8.56
15.52
932
60.14
33.12
2.76
12
4.81
4.02
2.06
63.75
3.06
6.98
20.11
936
60.37
32.30
3.23
10
6.01
8.23
5.76
52.88
5.55
10.81
16.93
937
60.57
19.92
2.49
8
6.89
3.67
4.68
53.59
4.93
9.61
23.32
940
60.27
28.40
3.55
8
7.04
1.76
3.70
60.28
3.17
9.69
21.38
941
53.08
25.28
3.16
8
6.18
1.57
.13
50.29
6.78
20.60 .
20.60
945
57.90
26.82
2.98
9
6.34
5.64
1.05
61.01
3.28
9.86
19.11
946
54.25
24.90
4.98
5
13.59
1.20
3.60
59.29
3.94
8.89
23.06
Mean
58.80
27.99-
2.99
9.3
6.10
3.59
3.28
58.41
4.33
9.86
20.47
(SD)
(3.11)
(5.39)
(2.35)
(2.4)
(2.68)
2.28
(1.37)
(4.94)
(1-24)
(4.54)
(2.94)
ro
o

TABLE III. Distribution of bedrest sleep periods by the presence or absence of REM and Stage 4 sleep.
Total
#
Day
Night
Mean duration
(range)
iime of occurrence
All stages
present
59
36
23
3.75 hrs
(.87 to 11.56 hrs)
No stage 4
13
12
1
1.41 hrs
(.5 to 2.9 hrs)
6:40am, 7:40am, 8:00am, 8:30am,
10:00am ,10:30am, 2:20pm, 3:10pm,
3:30pm, 4:30pm, 6:20pm, 9:00pm,
10:45pm
Mo stage REM
9
8
1
1,08 hrs
(34 to 161 min)
2:40am, 10:20am, 2:05pm, 3:10pm,
4:00pm, 4:30pm, 6:05pm, 7:20pm,
9:55pm
Neither REM
nor stage 4
3
3
0
1.02 hrs
(38 to 85 min)
12:05pm, 2:40pm, 10:30pm

22
detail, the sequencing of sleep and wakefulness of one subject (932)
across the bedrest period. Individual and group statistics for the
entire bedrest period are presented in Table II. BP was character
ized by an alternation of waking periods with an average length of
2.7 hours (range = .4 to 10.2 hours) and sleep periods with a mean
duration of 2.99 hours (range = .5 to 11.6 hours). Subjects spent
almost half of BP asleep (47.6%) with a mean total sleep time of
28 hours (range = 20 to 37.3 hours), Sleep was dispersed across an
average of 9.3 episodes during the bedrest period (range = 5 to 13).
These results are in contrast to what would be expected under
the assumption of a uniphasic model of sleep. Under such an assump
tion, one would predict 2 discrete sleep episodes across the 60 hours,
placed within the normal nocturnal periods (approximately 11 PM to
7 AM), with a mean TST for the bedrest period of about 16 hours
(about 27% of BP).
Sleep periods during BP could be divided into four general
categories based on the presence or absence of stage 4 and REM sleep,
as shown in Table III. Seventy percent (59) of all sleep periods
recorded over the bedrest period were 'full blown' sleep episodes,
in that all stages of sleep viere observed. Of the remaining 25
sleep periods, 3 (all day episodes) showed neither REM nor stage 4
sleep. Nine sleep periods (8 days, 1 night) exhibited an absence of
REM sleep and 13 (12 days, 1 night) showed no stage 4 sleep.
The mean period length of those episodes with neither REM nor
stage 4 was 61 minutes. The mean duration of 'noREM' sleep periods
was 65 minutes, and the 'noStg4'- sleep episodes averaged 84 minutes

23
duration. This is compared to a mean duration of 225 minutes
(3.75 h) for those sleep episodes which contained both stage 4 and
REM.
The occurrence of these events in terms of sidereal time was
also examined. The three periods containing no REM and no stage 4
occurred after noon and before the onset of the night phase (11 PM).
All but 2 of the 'noREM' periods occurred between 2 PM and 10 PM.
The 1noStg41 episodes occurred across the entire day phase (7 AM to
11 PM) with only one such episode occurring at 'night' (6:40 AM).
Regarding the mode of appearance of REM sleep episodes, of 153
REM periods recorded over BP, only 3 were characterized by onsets at
the beginning of a sleep period. Sleep onset REM periods (SOREMPS)
were defined as those REM episodes occurring at the onset of sleep
which were not preceded by more than 1 minute of stage 2. Two sleep
episodes containing SOREMPS also were characterized by the absence of
stage 4 sleep. The SOREMPS were contributed by 3 subjects, had a
mean duration of 15.3 minutes (9, 10, and 27 minutes), and occurred
at approximately 6:30 AM, 6:30 PM, and 1:30 PM, respectively.
In an effort to further elucidate the factors involved in the
cycling of REM sleep, two examinations were made. In the first pro
cedure, the REM sleep cycle (defined as the interval between the on
set of one REM period and the onset of the next) was calculated for
each subject's baseline night (Night 2). The onset points of the
cycle were then superimposed across each subject's entire bedrest
period (including waking episodes) to test the hypothesis that the
REM cycle was independent of prior sleep time. Also, the cycle was

24
superimposed across each subject's 'compressed1 bedrest period, in
which waking intervals were removed to obtain a continuous sleep per-
iod, to test the possible sleep dependency of the REM sleep cycle.
The initial onset of the superimposed cycle corresponded to the onset
of the last REM sleep period of Might 2.
The relationship between the superimposed REM cycle and the oc
currence of actual REM episodes was examined by calculating the mean
difference between each onset point of the hypothesized cycle and the
closest REM epoch within that sleep period. The average difference
between each onset of the superimposed cycle and the appearance of ari
actual REM epoch, across the intact bedrest period, was 15.8 minutes
(_SD = 3.7 min). The mean difference between the values for the 'com
pressed' BP was 14.5 minutes (SD = 4.7 min). The difference between
the two conditions was not significant.
To further test the sleep dependency of the cycle, the mean REM
cycle length of the 'compressed' BP, for the group, was determined.
Again, beginning the cycle with the onset of the last REM period of
Night 2, mean cycle length for the 'compressed' BP was 96.84 minutes
(SD 10.42 min). This is compared with a mean REM cycle length of
96.56 minutes (SD = 12.76 min) for Night 2.
Circadian Considerations. There were considerable differences
in several measures as a function of time of day. For comparative
purposes, waking and sleep periods were considered daytime events
if they were initiated between 7 AM and 11 PM. Nighttime events
were defined as those episodes with onsets between 11 PM and 7 AM.

25
TABLE IV, Comparison of sleep variables as a function of night
(11 PM 7 AM) and day (7 AM 11 PM).
Overal1
Might
Day
Mean TST
27,99 h
15,34 h
12,65 h
(5.39)
(3.30)
(4.91)
Mean Sleep
2.99 h
5.52 h **
1.93 h
Period Length
(2.35)
(2.65)
(.98)
Total # Sleep
84
25
59
Episodes
Sleep/Wake
6.10 h
7.81 h *
5.25 h
Cycle Length
(2.68)
(3.26)
(1.74)
% Stage 0
3.59
3.50
3.64
(2.28)
(2.26)
(3.58)
% Stage 1
3.28
3.88
2.81
(1.87)
(2.79)
(2.06)
% Stage 2
58.41
56.37
60.80
(4.94)
(5.76)
(6.22)
% Stage 3
4.33
4.13
4.94
0.24)
(1.58)
(1.75)
% Stage 4
9.86
8.89
11.01
(4.54)
(5.37)
(4.28)
% Stage REM
20.47
23.13 **
17.54
(2.94)
(4.01)
(3.87)
Mean
** p <.001
t test, two-tailed
(SD)
* p < 01

Figure 3.-Total number of subjects asleep, for each hour across
the bedrest period.
no
cn

27
Of 84 total sleep episodes recorded for the group across the
bedrest period, 59 (70.2%) occurred during the day and 25 (29.8%)
occurred at night. However, over half of the total sleep time (54.7%)
during BP was obtained during episodes initiated between 11 PM and 7
AM.
Comparisons between day sleep periods (DS) and night sleep
periods (NS) of temporal measures and sleep stage percentages are
shown in Table IV. The most notable difference in measures of DS and
NS was with regard to mean sleep period duration. NS episodes con
tinued for a mean duration 5.52 hours (SD = 2.65 h) compared to an
average length of 1.93 hours (SD = ,98 h) for sleep episodes
initiated during the day (t (82) = 6.58, p<.001). Conversely, the
mean duration of waking episodes was 1.47 hours (SD = 1.04 h) for
nighttime events and 3.12 hours (SD = 2.4 h) for those initiated
during the day (t (82) = 4.29, p<.001). Consequently, the sleep/
waking cycle length (defined as the interval between the onset of
one sleep episode and the onset of the next) increased significantly
from day (5.25 h, SD = 1.74) to night (7^81 h, SD = 3,26) t (73) =
2.89, pc.01.
Figure 3 presents, by hours, the number of subjects sleeping
across the bedrest period. The group tendency toward nocturnal
sleep placement is evident. Of 144 total subject-hours available for
sleep during the 2 night periods (11 PM to 7 AM), 104 (72.3%) were
used for that purpose. In contrast, of the 279 subject hours com
prising the first and second day periods (.8 AM to II PM and 7 AM to
11 PM) less than half (45.2%) were spent sleeping.

TABLE V.
- Sleep
onset times
of longest
sleep
episodes
for Nights
2 (baseline), 3,
and 4.
Subject #
928
929
932
936
937
940
941
945
946
Mean
Night 2
11:48P
11:27
11 :54
11 :42
11:50
11:57
12:05A
12:15
12:18
11:55P
3
12:30A
12:34
3:30A
10:50
2:42
2:42
2:58
11:45
9:39
1:01A
4
12:07
2:54
12:04
2:15
4:33
2:28
11:30
11 :21
2:14
1:43A
IND
00

PERCENT REM
30
T\D
VC
Figure 4.-Proportions of TST spent in REM sleep as a Unction of time of
day. Percentages were obtained by dividing the total minutes
of REM sleep for the group by total minutes of sleep, for each
time block across BP.

30
The possible presence of a free-running sleep/wake rhythm was.
examined by comparing the mean sleep onset time of Night 2 (baseline)
and the mean onset times of the longest sleep periods of Nights 3 and
4. The mean sleep onset times for the group exhibited a tendency to.
occur later on each succeeding nightsiightly under 1 hour later
on Night 3 (as compared to baseline) and 42 minutes later on Might
4 (as compared to Night 3). However, as shown in Table V only 2
subjects (929 and 937) demonstrated consistent free-running patterns
across the three nights. Four additional subjects delayed the onset
of their major nocturnal sleep onset on Night 4 (relative to Night 3).
Sleep stage percentages were not affected substantially as
a function of time of day. While there was slightly less stage 1
(2.8 vs 3.9%) and slightly more stage 2 (60.8 vs 56,4%) and stage 4
(11 vs 8.9%) during the daytime sleep periods, only the difference
in REM sleep percent was significant, with REM sleep comprising just-
over 23% of NS episodes and only 17.5% of DS periods (t (82) = 5.91,
pc.001). The difference in REM percent appeared to be a consequence
not only of the difference in sleep period length but also of the main
tenance of a circadian pattern of REMS across the the bedrest period,
as shown in Figure 4, The largest proportion of TST spent in REM
sleep occurred during those sleep episodes initiated between 4 AM and
8 AM. The proportion of REM sleep decreased across the next three 4-
hour time blocks until reaching a low of 11.57% of TST for sleep
episodes initiated between 4 PM and 8 PM. The percentage of REM
sleep then exhibited a steady increase across the subsequent 4-hour
time blocks (8 PM to 12 MID and 12 M to 4 AM).

31
TABLE VI, ~ Comparison of sleep variables, between first and second 24
hours, of bedrest period.
Overall
1st 24 HRS
2nd 24 HRS
Mean 1ST
27.99 h
14.46 h
** 10.97 h
(5.39)
(2.28)
(2.36)
Mean Sleep
2.99 h
3.54
2.82 h
Period Length
(2.35)
(1.92)
(1.44)
Total # Sleep
Episodes
84
39
33
Sleep/Wake
6.10 h
5.19 h
5,64 h
Cycle Length
(2.68)
(2.10)
(2.04)
% Stage 0
3.59 .
2.94
4.48
(2.28)
(2.23)
(4.25)
% Stage 1
3.28
3.03
3.74
(1.87)
(1.93)
(2.66)
% Stage 2
58.41
59.69
56.68
(4.94)
(4.03)
(8.50)
% Stage 3
4.33
4.15
4.66
(1.24)
(1.40)
(1.58)
% Stage 4
9.86
8.39
11.39
(4.54)
'(3.60)
(6.88)
% Stage REM
20.47
21.77
19.02
(2.94)
(2.56)
(4.82)
Mean
** p < .01
(SD)
t test5
two-tailed

32
In addition to day/night differences, there was als.o disparity
in temporal measures as a function of elapsed time within the bedrest
period. Table VI compares 1ST, sleep period duration, sleep/wakirig
cycle length, and sleep stage percentages recorded during the first
24 hours of bedrest with those recorded during the second 24 hours.
Total sleep time for the first 24 hours (14.5 h) was significantly
longer than TST for the second 24 hours (10,9 h) (t (16) = 3.19, p<
.01). While the decrease in TST was, in part attributable to a
reduction in the total number of sleep episodes during the second 24
hours (39 vs 33), a decline in the mean sleep period length across BP
(3.54 h vs 2.82 h) also contributed to the difference.
Despite the decrease in overall sleep period duration, the
sleep/waking cycle length remained essentially unchanged throughout
the bedrest period. For purposes of comparison, only those sleep/
waking cycles initiated and completed within one or the other 24 hour
periods were considered. The mean duration of sleep/waking cycles
occurring during the first 24 hours was 5.2 hours (SD = 2.1 h), while
the mean cycle length of those completed during the second 24 hours
was 5.6 hours (SD = 2.4 h).
There were no significant differences in sleep stage proportions
between the 24-hour intervals. However, trends were observed relative
to stages 4 and REM. Stage 4 increased (from 8.4 to 11.4'/) from the
first 24 hours to the next; this despite the fact that there were
more sleep episodes containing stage 4 in the first 24 hours (32 vs
27). Conversely, REM sleep decreased by over 10% from the first to
the second 24 hours.

CHAPTER FOUR
DISCUSSION
The results of this study substantiate and further detail pre
vious reports that a reduction in the control of the sleep response
results in a 'loosening' of the uniphasic, nocturnally placed sleep
pattern typically observed in adult humans. At the same time, how
ever, the results underscore the tendency of sleep to be differen
tially weighted toward a definite and persistent nocturnal placement.
On the one hand, the data suggest a quadriphasic sleep pattern,
with the onset of sleep episodes occurring at approximately 6 hour
intervals throughout the bedrest period. While such a rhythmicity
proved to be highly variable within subjects, from one sleep period
to the next, the relatively small between-subjects variability em
phasizes the regularity of such a rhythm within the group. That the
number of sleep periods was proportionately placed, relative to day
and night, further accentuates the uniform distribution of sleep epi
sodes across the 24-hour day. The 'day' phase (7 AM to 11 PM) com
prised approximately 73% of the total bedrest period. Correspondingly,
70% (59) of the total number of sleep episodes recorded-across BP were
initiated during the day.
On the other hand, comparison of the mean duration of those day
sleep periods with the average length of sleep periods initiated at
night clearly illustrates the substantial orientation toward the
33

34
nocturnal placement of total sleep time. Over half (55%) of the total
sleep time recorded during BP was obtained during sleep periods initi
ated between 11PM and 7AM, despite the fact that 'night' comprised only
27% of the total bedrest period.
These findings are in accord with those reported by Weitzman et al.
(1974) relative to the placement of the sleep of subjects maintained on
a 3-hour 'day'. In that study, the sleep of 7 subjects exhibited a
strong tendency to occur during the scheduled sleep periods beginning
at 3AM, 6AM, 9AM, and noon. While the authors note the shift away from
typical sleep placement (i.e., 11PM 7AM) during the 10 days of ultra-
dian cycling, they emphasize the "highly resistant nature" of the
circadian loading of the sleep response.
Webb and Agnew (1974a) also noted a tendency for subjects to
maintain a circadian, primarily nocturnal, placement of total sleep
time in a free-running environment. The authors cite three aspects of
their findings which suggest a continuation of "the entrained influence
of the original 24-hour cycle." First, several subjects showed sleep
onset times during the first 3 or 4 days which were significantly
earlier than would be predicted by a pure free-running hypothesis.
Secondly, throughout the 14 days of the study, there were tendencies
on the parts of 5 of the 7 subjects for sleep onset times to 'regress'
periodically, thereby increasing the probability that sleep onset
would occur during the nocturnal phase (between midnight and 7AM).
Finally, there were two occasions in which subjects skipped sleep
periods, the predicted onsets of which would have occurred during the
day phase (specifically, noon and 4PM), and instead delayed initiation
of the next sleep periods to nocturnal placements (11PM and 3AM,
respectively).

35
The tendency toward a free-running sleep/waking cycle was not
apparent in the present study. While the mean sleep onset times of
the major nocturnal sleep episodes became successively later across
the three nights, only 2 individuals exhibited such a tendency. Other
subjects delayed sleep onset on Night 3, only to advance onset time on
Night 4. Still others exhibited a reverse trend. Such 'scalloping'
of the expected free-running rhythm in the present study may reflect the
continued inclination of some subjects to maintain their sleep in its
typical nocturnal placement.
Total Sleep Time. The results of the present study are also
consistent with previous findings relative to the amount of total sleep
obtained. In summing up the findings on studies of unrestricted sleep,
Webb and Agnew (1975a) noted that "it may now be stated with confidence
that permitting subjects to sleep ad lib under conditions which have
not increased the sleep need by prior deprivation or increased energy
expenditure will consistently result in increased sleep length," (p.. 369). In
addition, sleep logs maintained across several weeks reveal that average
weekday ('restricted') sleep and weekend (unrestricted1) sleep differs
by an hour or more (Johns, Gay, Goodyear, and Masterton, 1971; Webb,
1981; White, 1975).
The subjects in the present study were instructed,' on Night 2, to
sleep until they felt well-rested and no longer sleepy. The mean sleep
duration for Night 2 (8.7 hrs) was within minutes of the mean total
sleep time (8.8 hrs) reported for 4 subjects (10 nights each) per
mitted to sleep as long as they wished, following 10 nights of
'restricted' sleep which averaged 7.4 hrs per night (Webb and Agnew,

36
1974a). The mean sleep length for Night 2 is also in agreement with
self-reported extensions in weekend sleep lengths (about 8.5 hrs).
The mean total sleep time reported here is, however, almost an
hour less than that of young adults (9.6 hrs) studied by Verdone (1968)
and Webb and Agnew (1975a). It seems likely that the mean total sleep
time for Night 2, like typically reported weekend sleep, was a conse
quence of the relaxation of self-imposed controls on sleep length
(e.g., alarm clocks), rather than an expression of the complete satis
faction of sleep 'need1. Several subjects in the present study were
rather impatient to terminate Night 2 and to initiate the bedrest
period. This is illustrated by the case of one subject who reported,
during a brief (3 min) awakening, that she felt rested and had "had
enough sleep." The report came after approximately 2 hours had elapsed
in Night 2. Clearly, such eagerness to 'get on with it' may have
resulted in the termination of sleep prior to the spontaneous ending of
sleep due to satiety.
A more accurate measure of maximum sleep length may be derived
from the examination, across the bedrest period, of sleep durations per
24 hours. Beginning with the termination of Night 2 (typically around
8AM) the first 24-hour period was characterized by an average of 14.5
hours of sleep, the second 24 hours contained an average of almost
11 hours of sleep. Such figures are consistent with previously re
ported 24-hour sleep amounts. For example, Aserinsky (1969) instructed
subjects to "make every attempt to sleep" while confined to bed for
30 hours (interrupted only twice for meals). During the first 24 hours
(midnight to midnight) subjects obtained an average of 14.4 hours of
sleep. During the 24-hour period beginning at 7:30AM, after a full

37
night's sleep (mean = 7.04 hrs.), the mean sleep duration was 12.98 hrs.
This figure is in remarkable agreement with the average 1ST for each of
the first 2 24-hour periods, subsequent to Night 2, in the present
study (12.71 hrs.). In light of such findings, it appears that 12 to
13 hours of sleep per 24 hours may be considered a close approximation
of the maximum capacity for sleep (Aserinsky, 196S). However, at least
three lines of evidence suggest that substantially shorter sleep times
may represent the maximum point of sleep "need". First, it has been
proposed that REM density may be a reflection of "satisfaction of a
sleep need, or . the buildup of a pressure to awaken" (Aserinsky,
1969, p. 155). It was found in that study that REM density reached
maximum values after 7.5 to 10 hours of sleep. Secondly, under condi
tions of perceptual deprivation, subjects averaged about 12 hours of
sleep during the first 24 hours, but subsequently reduced sleep lengths
until they approached baseline levels (about 7.5 hrs.) by the third or
fourth day of deprivation (Potter and Heron, 1972). Also in the present
study, total sleep time during the second 24 hours of bedrest decreased
almost 25I when compared with that of the first 24 hours. It might be
speculated that the decline would have continued, if bedrest had continued.
Nevertheless, with the exception of the perceptual deprivation
study cited above, all values of sleep satiation or fulfillment of
sleep "need" substsntialy exceed the 7- to 8~hour sleep durations
typically reported for normal young adults. The interpretation of such
findings may be generally separated into two opposing views. It has
been suggested, on the one hand, that the extended sleep observed in
unrestricted sleep regimens is a consequence of decreased sensory
input (see for example, Nakagawa, 1980; Heron, 1957). Proponents of

38
this view have speculated that an important function of the sensory and
perceptual systems is to maintain a waking level of arousal in the
reticular activating system. In studies comparable to the present one,
where conditions are maintained at a relatively constant level, and
sensory stimulation is not only static but also depressed relative to
normal levels, input from peripheral afferent pathways may be insuffi
cient to furnish the appropriate level of RAS activation. "In general,
appropriate sensory stimulation is needed to maintain an arousal state
in animals, but repetitive monotonous stimulation habituates the arousal
reaction . (Nakaguawa, 1980, p. 532). In short the, according to
this view, extended sleep is the result of neurological "boredom" and/or
habituation. Implied in such a view is the notion that the extended
sleep time observed under unrestricted or bedrest conditions is not
representative of natural sleep tendency (or requirement), but rather,
is simply one consequence of the organism's decreased requirement to
process environmental stimuli and respond accordingly (Heron, 1957).
This is analagous to Flanigan's (1971) suggestion that the extended
sleep observed in nonhuman subjects maintained under controlled and
highly favorable conditions (i.e., limited activity, plenty of food, no
predators) may reflect the upper limits of sleep that are organ!smically
possible, rather than that which is typical in the animal's natural
habitats.
In contrast, it has been suggested (see for example Webb and Agnew,
1975a) that the extended sleep lengths characteristic of unrestricted
sleep conditions are, in fact, "normal" and that typically reported
sleep durations of 7 to 8 hours may actually be the product of chronic
sleep deprivation. In support of such a notion is the finding that

39
only one third of a group of individuals who completed 6-month sleep
diaries reported that they usually awakened spontaneously in the morn
ing. Further, half of the respondents in the study reported that they
were typically not well-rested upon awakening (Kleitman, Mull in, Cooper-
man, and Titelbaum, 1937). The extension of sleep on weekends reported
by most individuals also suggests a !need! for more than the typically
obtained weekday mean. Further, it has been shown that after 3 consecu
tive nights of sleep, restricted to a period between 11PM and 7AM (mean
= 7.5 hrs), subjects slept an average of 2 hours longer on the 4th night
when allowed to awaken spontaneously (Webb and Agnew, 1975a). The
subjects in that study were riot encouraged to extend sleep: indeed, they
were "unaware that they would be allowed to sleep longer."
Similarly, the subjects in the present study were given no in
structions relative to the maintenance of sleep and wakefulness during
the bedrest period, yet they also obtained substantially more sleep than
self-reported normal sleep durations. This seems to be suggestive of an
inclination toward, if not an actual 'need' for, more sleep than is
typically obtained. Of interest is the observation that the extended
sleep durations resulted not from a lengthening of sleep episodes
initiated at night, but rather from the addition of day sleep periods.
Nocturnal sleep (11PM to 7AM) per 24 hours did not exceed normal night
sleep (7.8 h). However, day sleep episodes contributed an average of
approximately 5 hours to each 24-hour total. It might be speculated
that the usual sacrifice of such day sleep periods is the source of our
chronic sleep deprivation. On the other hand, the bedrest environment
in the present study was highly monotonous, with novel sensory input
severely curtailed. It is, therefore, likely that such conditions

40
contributed somewhat to the observed tendency for extended sleep times,
as wel 1.
Intrasleep Structure. The mean sleep stage proportions of the
combined sleep periods during BP did not differ significantly from those
of the baseline night (Night 2), and were within normal limits reported
for young adult populations, although REM and slow wave sleep (SWS) were
significantly reduced when compared to the normative sample used here.
Such findings are in agreement with those reported by Webb and Agnew
(1977) relative to the distribution of sleep stages in experimentally
controlled, polyphasic sleep regimens. That is, despite influences from
changes in sleep length, prior wakefulness and sleep onset times assoc
iated with various schedules of dispersed sleep, "when the total sleep
obtained during these altered regimens is compared with baseline sleep
on a 24-hour regimen, the basic structure of sleep persists" (p. 448).
Also, in accord with previous reports of dispersed and nap sleep
(Marn, Rechtschaffen and Wolpert, 1964; Moses, Hord, Lubin, Johnson,
and Naitoh, 1 975; Nakagawa, 1980; Webb and Agnew, 1 977 ; Weitzman et al.,
1974) the occurrence of REM sleep maintained a circadian pattern across
the bedrest period. Sixty percent of total REM time occurred during
sleep episodes initiated between 11 PM and 7AM. Further, while only 9
of the 84 sleep periods recorded during BP contained no REM sleep, 8 of
those sleep episodes occurred during the day, and 7 occurred between 2PM
and 10PM. Finally, Figure 4 graphically points out the tendency of
REM sleep to occur differentially at intervals across the 24-hour day,
with maxima and minima of the circadian oscillation occurring between
4 and 8AM and 4 and 8PM, respectively. Such data seem to underscore
the tendency for REM sleep to occur differentially across the 24 hours.

41
The decrease in REM percent during the bedrest period, compared
to baseline, also appears to have been the consequence of the circadian
placement of REM sleep. While total REM time was differentially
obtained within a relatively circumscribed time block, total sleep
time was influenced somewhat less by such contingencies. In addition,
opportunities for day sleep, when REM was less likely to occur, occupied
a larger percentage of the bedrest period than did night sleep time. As
a result, over the bedrest period, absolute sleep time was augmented to
a greater extent than was REM time, thereby lessening the proportion of
TST spent in REM sleep.
The decreased proportion of REM sleep recorded over the bedrest
period may also be attributable, in part, to the decrease in mean sleep
period length, relative to baseline, associated with BP. It has been
shown that the amount of TST spent in REM sleep becomes proportionately
less with the reduction of sleep length (Webb and Agnew, 1977). Since
mean sleep period length during the bedrest period was almost two-thirds
less -than that for the baseline night (2.99 hrs. vs 8.75 hrs.), a corre
sponding decline in the proportion of REM sleep for the majority of sleep
periods during BP would also be expected.
The observed decrease in REM percent, however, may also be a
reflection of the decreased mobility and/or lack of novel perceptual
stimulation characteristic of the bedrest situation. Several investi
gations of sleep patterns during prolonged bedrest have reported similar
declines in REM sleep (Aseririsky, 1969; Nakagawa, 1980; Ryback and Lewis,
1971). In addition, Adey, Bors and Porter (1968) reported decreased
porportions of REM sleep in quadraplegic patients, relative to normal
amounts. The authors proposed that the diminished REM sleep was related

42
to Toss of most motor performance in the waking state," as well as the
"absence of highly focused attention in complex motor activity." The
finding by Ryback and Lewis (1971), that subjects who were permitted no
exercise while confined to bed exhibited decreased REM sleep, while
those who were allowed to exercise during bedrest did not exhibit such
a decline, also provides some support for- this notion. In the present
study, the finding that REMS percent decreased (albeit slightly) from
the first to the second 24 hours, may likewise be interpreted as
supporting an immobilization/sensory habituation effect on the amount
of REM sleep obtained.
A rather unexpected finding in the present study was that of an
increase (though not significant) in the proportion of SWS, as the
bedrest period progressed. In their examination of patients with high
crevical lesions, Adey et al. (1968), reported the diminution of SWS
as well as decreased REM percentage. As with REM sleep, the authors
concluded that such a decline was the consequence of immobility. Thus,
the findings of the present study appear contrary to what would be
expected by an immobilization hypothesis.
On the other hand, the results reported here are in accord with
the findings relative to SWS reported by Ryback and Lewis (1971).
The bedrest subjects in that study exhibited increases in SWS with the
nonexercise group showing the larger rise. These results were inter
preted by the authors as indicating that SWS may serve to "repair or
maintain the muscular system" in response to the effects of atrophy,
as well as hypertrophy. Thus, it would appear that a change in the
proportion of SWS, in either direction, may be explained by one or the
other version of an immobilization hypothesis.

43
However, it seems somewhat unparsiinonious to have to employ both
versions to satisfactorily explain the data in terms, of immobilization.
In the present study, SWS comprised a smaller proportion of TST during
BP than during the baseline night. The hypothesis proposed by Adey
et al. is useful in explaining such results. Yet, within the bedrest
period, there was an increase in the percentage of SWS from the first
to the second 24 hours. These findings clearly require the Ryback and
Lewis version.
Several observations seem to support an alternative explanation
for the increase in SWS across the bedrest period, however. It has
been shown (webb and Agnew- 1971, p. 1354) that "longer periods of
wakefulness before sleep result in greater amounts of stage 4 sleep
in the first 3 hours of sleep." In the present study, the mean inter
val between sleep episodes increased by some 70% (from 1.65 hrs. to
2.82 hrs.) from the first to the second 24-hour period. At the same
time, the proportion of SWS increased almost 25% (from 12.9% to 16%).
By the same token, periods of wakefulness between sleep episodes
averaged about 1 hour longer during the day than at night (3.32 hrs. vs
2.29 hrs.). Correspondingly, SWS comprised a slightly greater propor
tion of day sleep than of night sleep episodes (16.2% vs 13.2%). The
relationship between prior wakefulness and subsequent amounts of SWS
seems apparent.
Such an association would also account for the decline of SWS
from Night 2 to the bedrest period. The amount of wakefulness preceed-
ing the baseline night (14 to 17 hrs.) far exceeded the duration of
waking intervals characteristic of BP. Consequently, the subsequent
sleep periods reflected this difference by exhibiting differential

44
proportions of SWS sleep. Inconsistent with a prior wakefulness
hypothesis, however, is the observation by several investigators that
there is a SWS 'kick' in the records of extended sleep. That is,
extended sleep is often characterized by the reappearance of substantial
SWS episodes in the late morning. The possible significance of such a
finding is considered later in the discussion.
Results of the examination of the influence of prior sleep on the
REM sleep cycle are best described as equivocal. Two examinations
were made, one to test the notion that the REM cycle is sleep-dependent,
the other to examine the possibility that the REM cycle is, instead, a
sleep-independent rhythm based on elapsed clock time, regardless of state.
Support for both views may be found in the literature, For example,
Globus (1966) found that the cycling of REM sleep demonstrated "at least
in part" a tendency to be a time-locked, rather than a sleep-dependent,
rhythm in the nap sleep of 2 subjects recorded for a total of 107 REM
periods. This conclusion was based primarily on the observation that
correlations between sleep onset and REM onset decreased at certain
times in the afternoon (between noon and 3 PM). That is, during this
interval sleep onset got progressively later while REM onset remained
constant relative to real time. Yet, prior sleep was not considered
and, therefore, the interpretation of such findings is problematic.
In contrast, Moses et al. (1975) and Moses, Lubin, Johnson, and
Naitoh (1977) reported findings which suggest that the REM sleep cycle
is a sleep-dependent phenomenon. In the 1977 study, the nap sleep of
25 subjects was 'compressed,' by subtracting all intervening waking
periods, and then examined as a single sleep episode. The resulting
mean REM cycle lengths did not differ significantly from those recorded

45
during baseline sleep. The same procedure was carried out in the present
study, and similar results were obtained. Using the-onset of the last
REM period of Night 2 as a starting point, the mean REM cycle length was
determined for subjects' 'compressed' bedrest periods, and compared to
the mean cycle length for Night 2. The mean cycle length for the
compressed period differed from the cycle length for Night 2 by less
than 1 minute (96.84 min vs 96.56 min). Such evidence is strongly
suggestive of a-sleep-dependent cycle.
In a second test, each subject's mean REM cycle length for Night 2
was superimposed on both his/her 'intact' bedrest period (i.e., includ
ing waking periods) and 'compressed' bedrest period. The mean deviation
of an actual REM occurrence, from the onset of the hypothesized cycle,
was used as a measure of goodness of fit.' The mean deviation across
the 'compressed' bedrest period was 14.5 minutes. In other words, the
actual occurrence of REM periods during BP differed by an average of
about 15 minutes, as compared to expected occurrences, when intervening
waking episodes were disregarded. Again, such a finding is supportive
of a sleep-dependent REM sleep cycle. However, a similar finding that
the mean deviation of actual REM sleep from hypothesized REM periods
across the 'intact' bedrest period (15.8 min), precludes the acceptance
of the sleep-dependence of REM sleep at the absolute exclusion of the
'real time' model.
Numerous studies of napping and dispersed sleep (Carskadon and
Dement, 1975: Globus, 1966; Moses et al., 1975; Nakagawa, 1980; Webb
et al., 1966) have reported perturbations in the mode of appearance of
REM sleep episodes, characterized by the onset of REM sleep within
minutes of the initiation of a sleep period. The reported frequencies

45
of occurrence of such sleep onset REM periods (SOREMPS) have ranged
from as few as 3 out of 107 REM episodes (Globus 1966) to as many as
21 of 34 REM episodes (Moses et al,, 1975). Marn et al. (1964) reported
the occurrence of no REM onset periods in the examination of 18 subjects
who napped in either the afternoon (about 1:30PM) or evening (about
7:30PM). Yet, the overall rate of occurrence of SOREMPS for 5 studies
of dispersed and nap sleep (Carskadon and Dement, 1975; Globus, 1966;
Moses et al,, 1975: Nakaga.ua, 1980: Weitzman et al., 1974) is rather
substantial at 39% of all REM periods recorded (range 2.8% 72%).
Since such events are almost never observed in normal nocturnal sleep,
SOREMPS have been considered to be indications of the overall disruption
of normal sleep structure associated with day sleep episodes and
dispersed sleep regimens (Weitzman et al., 1974; Moses et al., 1975).
In the present study, of 153 REM episodes recorded over the bedrest
period, only 3 were initiated at the beginning of a sleep period (i.e.,
before the appearance of 2 min of stage 2). The infrequency of such
occurrences suggests that the dispersed sleep of the present study may
be of a different sort than typically observed nap sleep or experimen
tally dispersed sleep.
Such a view is further supported by the regularity with which sleep
stages occurred across the bedrest period. Seventy percent of all- sleep
periods recorded during BP were full blown' sleep episodes (61% of day
sleep episodes); that is, all sleep stages were present. By comparison,
only 16% of the 560 opportunities for sleep in the Weitzman et al. study
contained all stages. Of the 80 naps examined by Moses et al. (1975),
27.5% contained both SWS and REM sleep. REM sleep occurred in 90% of
all sleep episodes during BP (86% of day sleep episodes), SWS occurred

47
in 85% of the sleep periods in this study (80% of day sleep periods).
Again by comparison, the naps of subjects studied by Moses et al.
contained REM episodes 43% of the time; 35% of the sleep opportunities
examined by Weitzman et al. contained REM and 48% contained SWS.
Seven of the 9 sleep periods which contained no REM sleep occurred
between 2PM and 10PM. Such a finding is consistent with previous ones
that REM sleep typically increases in amount from midnight to morning
and decreases toward evening (Moses et al., 1975; Nakagawa, 1980;
Weitzman et al., 1974). It is likely, then, that the absense of REM
during some sleep episodes in the present study were, at least in part,
due to circadian effects.
A similar possibility also exists relative to 'missing' periods
of SWS (specifically, stage 4). Several examinations of extended
sleep (see for example, Gagnon and Dekoninck, 1981; Webb, Campbell and
Hend in, 1981) have noted the re-occurrence of stage 4 (in relatively
large amounts) during the late morning. In the present study, sleep
episodes characterized by the absence of stage 4 occurred across the
entire day (6:40AM to 10:45 PM). It is interesting to note, however,
that an obvious 'gap' occurred between 10:30AM and 2:20PM. That is,
the sleep episodes initiated during those four hours always contained
stage 4 sleep. Sleep periods initiated after about 11PM, and
throughout the night, also contained stage 4. Thus, it might be
suggested that the absence of stage 4 in 15% of the sleep episodes in
this study, was the manifestation of the biphasic cycling of stage 4
sleep, rather than a reflection of the disruption of normal sleep
processes. A further indication of the possible bipolar nature of
stage 4 is the observation that total stage 4 time for BP was evenly
distributed between day and night (716 min vs 714 min).

48
Clearly, the differences in the appearance and sequencing of
sleep stages between this study and the 2 studies of nap sleep with
which it has just been compared, are attributable in large part to
the difference in mean sleep episode length (3 hrs. vs 1 hr.). This is
especially true of REM sleep, if sleep dependency is assumed. Never
theless, the persistence and regularity with which sleep stages occurred
across the bedrest period is noteworthy. There were no significant
differences in the proportions of sleep stages between bedrest sleep
and Might 2 sleep. Nor, were there significant changes in sleep stage
percentages between the first and second 24 hours of bedrest. Finally,
only REM sleep percent varied significantly relative to day or night
onset times.
Weitzman et al. (1974) noted that the sleep episodes which they
observed were "clearly not minatures of the normal 8 hr. sleep pat
tern." Such a view has been echoed by most others who have studied
dispersed sleep and naps. Most of these studies have differed from
the present one, in that the sleep periods were typically experi
mentally circumscribed, rather than being placed on a 'self-demand1
schedule. Such a methodological difference may account for the
dissimilarity of results reported here.
Yet, Nakagawa (1980) also reported significant differences in
day episodes versus baseline (night) sleep, despite the fact
that the onset and termination of sleep periods was not regulated.
Significant increases in stage 1, significant decreases in stages 4
and REM and the presence of SOREMPS led Nakagawa to conclude that
"daytime sleep itself might be characterized as a special sleep at
the onset of the diurnal sleep-waking cycle," (p. 534).

49
In that study, Nakagawa noted a sleep/waking cycle of approxi
mately 4 hours, and suggested that a significant relationship seemed
to exist between this cycle and the scheduling of meals. The quadri-
phasic rhythm observed during the present study did not appear to be
the result of feeding schedules, since less than half (42%) of waking
periods were used for meals. Further, one subject (#932) fasted
for the duration of the experiment, yet maintained the least variable
sleep/waking cycle of the group.
Uni phasic or Ultradian? The results of the present study seem
to suggest, then, that the daytime sleep observed was not a "special"
sleep, but rather was the continuation of an ultradian sleep/waking
rhythm, which nevertheless maintained a definite tendency to be
weighted toward a nocturnal placement, relative to total sleep time.
The regularity with which sleep episodes were initiated across
the bedrest period also seems to support the notion, of an ultradian
pattern of sleep. The mean within-subject variability of the sleep/
waking cycles was 2.66 h (^SD). While the between-subject variability
was similar (SD = 2.73), most of the variability was contributed by
a single subject. This subject (#946, see Table II) exhibited a
sleep/waking cycle (13.59 h) which more than doubled the mean sleep/
waking cycle length of the other 8 subjects (5.85 h). Likewise, the
variability of the sleep/waking cycle of this subject (SD = 4.52)
exceeded the group mean by almost twice (SD = 2.43). Excluding this
subject's sleep/waking cycle length from the group mean reduces the
variability of the quadriphasic rhythm to less than an hour (SD =
57 min).

50
Such a pattern of variability is similar to that observed for
the REM sleep cycle. The 90 to 100 minute REM cycle is of limited
value in predicting the occurrence of REM episodes within the sleep
record of a single subject. Yet for a population, the predictability
of a 90-100 minute cycle is quite good.
A further indication of the regularity of the quadriphasic sleep/
waking cycle is the observation that the overall cycle length did
not change significantly from one 24 hour period to the next. On
the other hand, the length of the sleep/waking cycle did differ as a
function of day and night, with the night cycle being substantially
longer (7.8 hrs. vs 5.2 hrs.). It might be speculated that this
difference was the consequence of subjects having spent the majority
of their lives entrained to a monophasic sleep system, and that given
time, the difference in cycle lengths might diminish. The present
data do not support such a notion, however since the difference
between night sleep/waking cycle length and day sleep/waking cycle
length did not lessen significantly in the second 24 hours of bed
rest.
The polyphasic distribution of the sleep of numerous animals
under experimental conditions is well documented. Berger (1972)
has noted that "if we wish to study comparative aspects of normal
sleep, we shouldn't impose wakefulness at any time on any species,
including the human, but should take 24-hour recordings under equiva
lent conditions." The relatively basal activity levels, absence of
novel perceptual stimulation and the satisfaction of bodily functions
characteristic of the present experimental situation are comparable

51
to typical conditions under which animal sleep is studied. That a
polyphasic sleep pattern was exhibited under such conditions, therefore,
was not completely unexpected. Indeed, in 1972 Snyder speculated that
"even in the human we would probably find polyphasic sleep if we used
the same conditions that we use for animals," (p. 45).
While the monotony of the situation was quite likely instrumental
in the increase in total sleep time over the bedrest period, it seems
less probable that such "neurological boredom" would be expressed in
a cyclic manner. Yet, the possibility remains, as noted by Weitzman
(1972) with regard to the sleep cycle in cats. "In your experimental
situation, .... you are measuring in part the shift of pattern which
is taking place from a previous long-established light-dark pattern, in
addition to adaptation to a strange new environment. Are you measuring
the effect of that shift as well as the intrinsic rhythmicity in the
cat? If the measured rhythmicity were constant over, say, several
weeks, then it would support the concept that a basic rhythmicity was
being measured. However, if it tended to change over 1, 2, or 3 weeks
that would suggest that you are at least partially measuring an adapta
tion effect," (p. 201). Obviously, the question relative to human
sleep cannot be resolved by results of the present experiment.
It is suggested, though, that the conditions of the present study
represented a further advancement along a continuum of slackened
behavioral controls over the sleep process, the results of which was a
clearer manifestation of an inherent ultradian sleep/waking rhythm,
rather than the adaptation of the sleep response to a "strange new
environment." It is further suggested that such a rhythm, albeit in

52
highly dampened form, may also be reflected in the occasional naps
taken by normal adults (especially on weekends), the increased frequency
of naps associated with the loosened schedules of geriatric and college
populations, the sleep patterns of infants, and the polyphasic sleep
patterning noted in unstructured environments (i.e., Arctic expeditions).
And, I would suggest one additional manifestation of the ultradian
nature of the sleep response. Webb (1969) has noted that "(no one yet
has exploited, a foundation to study the 'siesta' patterns of the tropi
cal and subtropical cultures, although at various odd moments, I am
most tempted)," (p. 53). Results of the present study may suggest
that such a proposal, to examine cross-cultural diversity in the
cycling of rest and activity,.may deserve a better fate than paren
thetical consideration.

APPENDIX 1
SLEEP QUESTIONNAIRE
1. How many hours, on the average, do you usually sleep per night?
less, 5, 5%, 6, 6k, 7, Ik, 8, 8k, 9, 9k, 10, more
2. How many hours of sleep pen night do you think that you need to
feel well and function adequately?
3. How many hours of sleep per night do you prefer to have when you
have a chance?
4. In general, the amount of sleep that you usually get is:
not enough about enough too much
5. On the average, about how many naps do you usually take each week?
0, 1, 2, 3, 4, 5, 6, 7, 8, or more
6. About how many total hours do you usually spend each week taking
naps?
7. About how many times on the average do you wake up each night?
0, 1, 2, 3, 4, 5, 6, 7, 8, or more
8. About how much total time do you spend awake after going to sleep?,
none, less than 5 min., 5-15, 16-30, 31-60, more
9. About how many minutes does it typically take for you to go to
sleep? less than 5, 5-15, 16-30, 31-60, more than 1 hour
10. How often do you have trouble getting to sleep as quickly as you
would like? Almost never, occasionally, often, almost always
11. How much do you enjoy sleep?
not at all, a little, moderately, much
12. How regular are your bedtimes?
very regular, somewhat regular, somewhat irregular, very
irregular
13. How regular are your wake-up times?
very regular, somewhat regular, somewhat irregular, very
irregular
14. What is your average bedtime during weekdays?
Tto nearest 15 min.7
53

54
15. What is your average bedtime on weekends? ^
(to~~nearest 15 min.)
16. What is your average wake up time on weekdays?
(to nearest 15 minTj
17. What is your average wake up time on weekends?
(to nearest 15 min.j
18. How well do you usually sleep at night?
Very well, satisfactorily, some problems, poorly.
IS. Do you usually feel well rested when you wake up, or soon thereafter?
Almost never, occasionally, often, almost always
20. On the average, how many days per week do you usually go to bed
more than one (1) hour earlier or later than your average bedtime?
0, 1, 2, 3, 4, 5, 6, 7
21. On the average, how many days per week do you usually wake up more
than one (1) hour earlier or later than your usual wake up time?
0, 1, 2, 3, 4, 5, 6, 7
22. Do you consider your sleep very light or very deep? Check below,
very light very deep
23. Do you feel tired during the day because you have slept poorly?
Never sometimes often always
24. How much alcohol do you drink?
Never sometimes everyday
25. Do you wake up before you intend to?
never
once or twice a month
once or twice a week
nearly every day
26. Do you use sleeping pills to get to sleep?
never
about once a month
about once a week
nearly nightly
27. What, if any, sleeping pills do you use?
28. What, if any, prescription drugs are you presently taking?

REFERENCES
Adey, W., Bors, E. and Porter, R. EEG Sleep Patterns After High
Cervical Lesions in Man. Arch. Neurol., 1968, ]9_, 377-383.
Agnew, H. and Webb, W. Sleep Stage Scoring. J. Suppl. Abstr.
Serv., Am. Psychol. Assoc., Ms., 1972a, No. 293.
Agnew, H. and Webb, W. Measurement of Sleep Onset by EEG Criteria.
Amer. J. EEG Techno!., 1972b, 12(3), 127-134.
Aschoff, J. Circadian Rhythms in Man. Science, 1965, 148, 1427-1432.
Aschoff, J. Features of Circadian Rhythms Relevant for the Design
of Shift Schedules. Ergonomics, 1978, 21 (10), 739-754.
Aschoff, J. and Wever, R. Human Circadian Rhythms: A Multi-
oscillatory System. Federation Proceedings, 1976, 35(12),
2326-2332.
Aserinsky, E. The Maximal Capacity for Sleep: Rapid Eye Movement
Density as an Index of Sleep Satiety. Biological Psychiatry,
1969, 1, 147-159.
Berger, R. Discussion in Chase, M. H. The Sleeping Brain, Perspec-
' tives in the Brain Sciences, Brain Information Service/Brain
Research Institute, Vol. 1, p. 45, Los Angeles: UCLA, 1972.
Carskadon, M. and Dement, W. Sleep Studies on a 90-minute Day.
Electroenceph Clin. Neurophysiol., 1975, 3_9, 145-155.
Colquhoun, W. P. Circadian Variations in Mental Efficiency. In
W. Colquhoun (Ed.), Biological Rhythms and Human Performance,
pp. 39-107, New York: Academic Press, 1971.
Ellingson, R. Ontogenesis of Sleep in the Human. In G. Lairy and
P. Salzarulo (Eds.), The Experimental Study of Human Sleep:
Methodological Problems, pp. 129-150, New York: Elsevier Pub.,
1975.
Flanigan, W. Phylogeny and Sleep Behavior: A'Review. Unpublished
manuscript, 1971.
Flanigan, W. Behavioral States and Electroencephaograms of Reptiles.
In M. Chase (Ed.), The Sleeping Brain, Brain Information Service/
Brain Research Institute, Los Angeles: UCLA, 1972.
55

56
Froberg, J., Karlsson, C., Levi, L. and Lidberg, L. Psychobiological
Circadian Rhythms During a 72 Hour Vigil. Forsvarsmdicin, 1975,
VU 192-201.'
Gagnon, P., and De Koninck. Reappareance of EEG Slow Waves in Extended
S1eep. Sleep Research, 1981 (in press).
Gesell, A., and Amatruda, C. The Embryology of Behavior, New York:
Harper, 1945, pp. 144-159.
Globus, G. Rapid Eye Movement Cycle in Real Time. Arch. Gen,.
Psychiat., 1966, 1_5, 654-659.
Heron, W. The Pathology of Boredom. Scientific Amer., January,
1957, 1, 430-436.
Johns, M., Gay, T., Goodyear, M. and Masterton, J. Sleep Habits
of Healthy Young Adults: Use of a Sleep Questionnaire. Brit.
J. Prev. Soc. Med., 1971, 25, 236-241.
Kleitman, N. Sleep and Wakefulness, Chicago: University of
Chicago Press, 1963.
Kleitman, N., and Engelmann, T. Sleep Characteristics of Infants.
J. Applied Physiology, 1953, 6_, 269-282.
Kleitman, N., Mull in, F., Cooperman, N. and Titelbaum, S. Sleep
Characteristies, Chicago: University of Chicago Press, 1937.
Lewis, H. Sleep Patterns on Polar Expeditions. In G. Wolstenholme
and M. O'Connor (Eds.), CIBA Foundation Symposium on the Nature
of Sleep, pp. 322-328, London: J. and A. Churchill, 1951.
Lewis, H., and Masterton, J. Sleep and Wakefulness in the Arctic.
Lancet, 1957, 1, 1262-1266.
Marn, L., Rechtschaffen, D. and Wolpert, B. Sleep Cycle During
Napping. Arch. Gen. Psychiat., 1964, J_}, 503-508.
Moses, J., Herd, D., Lubin, A., Johnson, L. and Naitoh, P. Dynamics
of Nap Sleep During a 40 Hour Period. Electoenceph Clin.
Neurophysiol., 1975, 39, 627-633.
Moses, J., Lubin, A., Johnson, L. and Naitoh, P. Rapid Eye Movement
Cycle is a Sleep-Dependent Rhythm. Nature, 1977, 265(5592),
360-361.
Murry, E., Williams, H. and Lubin, A. Body Temperature and Physcho-
logical Ratings During Sleep Deprivation. J. Exp. Psychology,
1958, 56(3), 271-274.

57
Nakagawa, Y. Continuous Observation of EEG Patterns at Night and
in Daytime of Normal Subjects Under Restrained Conditions I,
Quiescent State When Lying Down. Electroenceph. Clin. Neuro-
physiot, 1980,49(5-6), 524-537.
Parmellee, A,, Wenner, W. and Schulz, H. Infant Sleep Patterns
From Birth to 16 Weeks of Age. J. Peeliat., 1964, 6_5, 576-582.
Potter, W. and Heron, W. Sleep During Perceptual Deprivation.
Brain Research, 1972, 40, 534-539.
Richter, C. R. Biological Clocks in Medicine and Psychiatry, Spring-
field, Illinois: C. C. Thomas, 1965.
Ryback, R., and Lewis, 0. Effects of Prolonged Bed Rest on EEG
Sleep Patterns in Young, Healthy Volunteers. EEG and Clinical
Neurophysiol, 1971, 3J_, 395-399.
Ryback, R., Lewis, 0. and Lessard, C. Psychobiologic Effect's of
Prolonged Bed Rest (Weightless) in Young, Healthy Volunteers
(Study 2). Aerospace Medicine, 1971, 42(5), 529-535.
Salzarulo, P., Fagioloi, I., Salomon, F., Ricour, C., Raimbault,
G., Ambrosi, S., Cicchi, V., Duhamel, J. and Rigoard, M. Sleep
Patterns in Infants Under Continuous Feeding From Birth. EEG
and Cl in. Neurophysiol., 1980, 49, 330-336.
Sander, L., Stechler, G., Burns, P. and Julia, H. Early Mother-
Infant Interaction and 24-Hour Patterns of Activity and Sleep.
J. Am. Acad. Child. Psychiat.., 1970, 9, 103-123.
Snyder, F. Discussion in Chase, M. H., The Sleep Brain. Perspec-
' fives in the Brain Sciences, Brain Information Service/Brain
Research Institute, Vol. 1, p. 45, Los Angeles: UCLA, 1972.
Sterman, M. Ontogeny of Sleep: Implications for Function. In
Drucker-Colin, Shkurovich and Sterman (Eds.). The Functions
of Sleep, pp. 207-232, New York: Academic Press, 1979.
Symposium on Quantitative Biology, Vol. XXV, Biological Clocks.
Cold Springs Harbor, Long Island Cold Springs Harbor Symp.,-
Quant. Biol., 1960, 25.
Verdone, P. Sleep Satiation-Extended Sleep in Normal Subjects.
Electroenceph. Clin. Neurophysiol., 1968, 24(5), 417-423.
Webb, W. Twenty-Four Hour Sleep Cycling. In Kales, A. H. (Ed.),
Sleep: Physiology and Pathology, pp. 53-65, Philadelphia:
J. B. Lippincott Co., 1969.

58
Webb, W. Sleep Behavior as a Biorhythm. In W. P, Colquhoun (Ed.),
Biological Rhythms and Human Performance, pp. 149-178, New York
Academic Press, 1971.
Webb, W. Sleep and Maps, Speculations in Science and Tech., 1978,
1(4), 313-318.
Webb, W. Sleep Patterns in Older Males and Females. Research Comm.
in Psychiatry, Psychology and Biology, 1981 (in press).
Webb, W., and Agnew, H. Stage 4 Sleep: Influence of Time Course
Variables. Science, 1971. 124, 1354-1356.
Webb, W., and Agnew, H. Sleep and Waking in a Time-Free Environment'.
Aerospace.Med., 1974a, 45(6), 617-622.
Webb, W., and Agnew, H. Regularity in the Control of the Free-
Running Sleep-Wakefulness Rhythm. Aerospace Med., 1974b,
45(7), 701-704.
Webb, W.,and Agnew, H. Are We Chronically Sleep Deprived? Bulletin
of the Psychonotnic Society, 1975, 6(1 ), 47-48.
Webb, W., and Agnew, H. Sleep Efficiency for Sleep-Wake Cycles of
Varied Length. Psychophysiol., 1975a, 60), 47-48.
Webb, W.,and Agnew, H. The Effects on Subsequent Sleep of an Acute
Restriction of Sleep Length. Psychophysiol., 1975b, 12(4),
367-370.
Webb, W., and Agnew, H. Analysis of the Sleep Stages in Sleep-Wake
fulness Regimens of Varied Length. Psychophsioloqy, 1977,
14(5), 445-450.
Webb, W., and Agnew, H. Effects of Rapidly Rotating Shifts on Sleep
Patterns and Sleep Structure. Aviat. Space and Environ. Med.,
1978, 49(2), 384-389.
Webb, W., Agnew, H. and Sternthal, H. Sleep During the Early
Morning. Psychonom. Sci., 1966, 6, 277-278.
Webb, W., Agnew, H. and Williams, R. Effect on Sleep of a Sleep
Period Time Displacement. Aerospace Med., 1971, 42(2), 152-
155.
Webb, W., Campbell, S. and Hendlin, R. The Termination of Extended
Sleep. Biological Psychology, 1980, H, 45-48.

59
Weitzman, E. Periodicity in Sleep and Waking States in Chase M. H.
(Ed.), The Sleeping Brain, Perspectives in the Brain Sciences,
V'o'l. 1, p. 45, Brain Information Service/Brain Research Insti
tute, Los Angeles: UCLA, 1972.
Weitzman, E., Nogeire, C., Perlow, M., Fukushima, P., Sassin, J.,
McGregor, P., Gallagher, T. and Heilman, L. Effects of a Pro
longed. 3-Hour Sleep-Wake Cycle on Sleep Stages, Plasms Cortisol
Gowth Homone and Body Temperature in Man. J. Clin. Endocrinol.
Metab., 1974, 38, 1018-1029.
Wever, R. Internal Phase-Angle Differences in Human Circadian
Rhythms: Causes for Changes and Problems of Determinations.
Int. J. Chronobiology, 1973, J[, 371 -390.
White, R. Sleep Length and Variability: Measurement and Inter
relationships. PhD Dissertation, University of Florida, 1975.
Williams, R., Agnew, H. and Webb, W. Sleep Patterns in Young Adults
An EEG Study. Electroenceph. Clin. Neurophysiol., 1964,
17, 376-381'.
Williams, R., Agnew, H. and Webb, W. Sleep Patterns in Young Adult
Females: An EEG Study. Electroenceph. Clin. Neurophysiol.,
1966, 20, 264-266.

BIOGRAPHICAL SKETCH
Scott Searcy Campbell was born February 25, 1952. He received
his B.A. degree, with a major in psychology, from Florida State
University in 1374. From September, 1972, to June, 1978, he attended
Montana State University from which he received the M.S. degree in
experimental psychology. In September, 1978, he continued his
graduate studies, under the supervision of Wilse B. Webb, at the
University of Florida. Upon completion of the Ph.D. degree, he
accepted a postdoctoral fellowship at Harvard Medical School.
60

I certify that. I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality, as a dissertation for
the degree of Doctor of Philosophy.
%)
I Vj
v'6
y.-
NT
U/
Wilse B. Webb, Chairman
Graduate Research Professor of
Psychology
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality, as a dissertation for
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality
the degree of Doctor of Philosophy.
as a dissertation for
Donald A. Dewsbury
Professor of Psychology
I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation
and is fully adequate, in scope and quality, as a dissertation for
the degree *of Doctor of Philosophy.
Charles M. Levy
Professor' of Psychology

I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
Professor of Electrical
Engineering
This dissertation was submitted to the Graduate Faculty of the
Department of Psychology in the College of Liberal Arts and
Sciences and to the Graduate Council, and was accepted as partial
fulfillment of the requirements for the degree of Doctor of
Philosophy.
August 1979
Dean for Graduate Studies and
Research



9
environment, but other behavioral controls are typically left
unaltered. That is, subjects are asked to lead a "normal" life,
including eating three meals a day in normal sequence, having the
opportunity for exercise and other activities such as reading,
writing, watching TV, etc., and maintaining circumscribed sleep
schedules (no naps). As a result, these subjects have a tendency
to carry on day-to-day life in a manner to which they have grown
accustomed over the years. Clearly, such circumstances may strongly
bias behavior toward a circadian schedule, perhaps entraining an
inherently ultradian rhythm to an environmentally and culturally
modulated circadian framework. These factors notwithstanding,
the sleep patterns displayed by subjects in time free environments
have been seen as providing evidence not only for the biological
basis of the sleep/waking rhythm, but for its circadian (uniphasic)
nature, as well.
The latter inference seems tenuous, at best. For, when controls
on the number and placement of sleep episodes are reduced in free-
running environments, evidence seems to indicate that the "normal
sleep pattern" (single sleep period of from 7 to 10 hours with
occasional napping) is modified. Webb and Agnew (1974a) reported
that under such circumstances "a significant increase-in both long
and short periods of sleep, compared to the 24-hour environment" (p.
620) resulted. While neither the placement nor the periodicity of
the increased episodes was analyzed, the authors concluded that such
changes in the sleep response showed the extent to which the
uniphasic sleep pattern is disrupted by reduced controls on sleep
placement.


35
The tendency toward a free-running sleep/waking cycle was not
apparent in the present study. While the mean sleep onset times of
the major nocturnal sleep episodes became successively later across
the three nights, only 2 individuals exhibited such a tendency. Other
subjects delayed sleep onset on Night 3, only to advance onset time on
Night 4. Still others exhibited a reverse trend. Such 'scalloping'
of the expected free-running rhythm in the present study may reflect the
continued inclination of some subjects to maintain their sleep in its
typical nocturnal placement.
Total Sleep Time. The results of the present study are also
consistent with previous findings relative to the amount of total sleep
obtained. In summing up the findings on studies of unrestricted sleep,
Webb and Agnew (1975a) noted that "it may now be stated with confidence
that permitting subjects to sleep ad lib under conditions which have
not increased the sleep need by prior deprivation or increased energy
expenditure will consistently result in increased sleep length," (p.. 369). In
addition, sleep logs maintained across several weeks reveal that average
weekday ('restricted') sleep and weekend (unrestricted1) sleep differs
by an hour or more (Johns, Gay, Goodyear, and Masterton, 1971; Webb,
1981; White, 1975).
The subjects in the present study were instructed,' on Night 2, to
sleep until they felt well-rested and no longer sleepy. The mean sleep
duration for Night 2 (8.7 hrs) was within minutes of the mean total
sleep time (8.8 hrs) reported for 4 subjects (10 nights each) per
mitted to sleep as long as they wished, following 10 nights of
'restricted' sleep which averaged 7.4 hrs per night (Webb and Agnew,


8
Within the more structured hours of daily life, it has been
suggested that the ultradian nature of sleep and wakefulness may be
manifest in the form of naps (Webb, 1978). Webb's proposal that naps
are part of an inborn sleep/waking rhythm is based on the contention
that certain characteristics of these intermittent sleep episodes
match features which are common to biological rhythms in general.
Specifically, it is maintained that naps are: 1) temporally repeti
tive, 2) species specific, 3) often developmental, 4) innate and
unlearned, 5) endogenous, and 6) adaptive. While evidence seems to
support the first four characteristics, and only post hoc specula
tion can address the sixth, Webb points out that experimental
support for the claim that naps are endogenously mediated is
extremely limited, and that further data are "likely to be slow
forthcoming." This is due primarily to the fact that virtually
all procedures utilized for the examination of presumed biological
rhythms have prohibited other-than-circumscribed sleep episodes
as a consequence of experimental design considerations. Such experi
ments typically focus on the interactions of major components of
sleep, waking and performance. As such, "the very nature of naps
in these designs are uncontrolled and interfering sources of experi
mental variance" (Webb, 1978 p. 316). Two primary designs of this
type are the free-running and bedrest procedures.
It has been noted that the sleep/waking rhythms maintained by
subjects in free-running environments provide excellent support for
the notion that such rhythms are innately and independently controll
ed. Under such conditions, time cues are removed from the


Figure 3.-Total number of subjects asleep, for each hour across
the bedrest period.
no
cn


LIST OF FIGURES
FIGURE PAGE
"1 Placement of sleep/waking episode across 60 "18
hours for individuals and group.
2 Example of sleep across BP (#932). 19
3 Total number of subjects asleep, for each hour 26
across the bedrest period.
4. Proportions of TST spent in REM sleep, as a 29
function of time of day.
vi


7
that given a situation in which adherence to customary circadian
scheduling is less stringently controlled, a more accurate reflection
of the inborn sleep/waking rhythm may emerge.
In the Polar regions the photoperiod is not diurnal, but seasonal
Further, expeditions into these regions typically maintain little
resemblance to an organized community, in that there are few circum
scribed daily routines, no specific feeding schedules, few timetables
(Lewis, 1961). Sleep records from 5 Polar campaigns suggest that the
organization of sleep/waking patterns was a direct function of the
stringency with which the "community" was structured. For example,
members of the British North Greenland Expedition (1952-54) were
allowed to sleep as long as they liked, at virtually any time they had
the inclination to do so. Under such conditions, sleep durations per
24 hours did not increase substantially over baseline amounts. Yet.
the length and placement of individual sleep episodes showed con
siderable deviations from members' normal sleep habits recorded in
England. The tendency toward polyphasic sleep placement was most
extreme during the three months of continuous light or darkness.
"During these periods . men used any of the 24 clock-hours for
sleep. The winter months were characterized by many interruptions to
sleep and by the taking of naps" (Lewis, 1961, p, 325). Although
no data were presented relative to the r'nythmicity of these multiple
sleep episodes, it appears that the uni phasic "nature" of the sleep
process loosened somewhat under conditions of limited environmental
cues and negligible behavioral and cultural restraints.


37
night's sleep (mean = 7.04 hrs.), the mean sleep duration was 12.98 hrs.
This figure is in remarkable agreement with the average 1ST for each of
the first 2 24-hour periods, subsequent to Night 2, in the present
study (12.71 hrs.). In light of such findings, it appears that 12 to
13 hours of sleep per 24 hours may be considered a close approximation
of the maximum capacity for sleep (Aserinsky, 196S). However, at least
three lines of evidence suggest that substantially shorter sleep times
may represent the maximum point of sleep "need". First, it has been
proposed that REM density may be a reflection of "satisfaction of a
sleep need, or . the buildup of a pressure to awaken" (Aserinsky,
1969, p. 155). It was found in that study that REM density reached
maximum values after 7.5 to 10 hours of sleep. Secondly, under condi
tions of perceptual deprivation, subjects averaged about 12 hours of
sleep during the first 24 hours, but subsequently reduced sleep lengths
until they approached baseline levels (about 7.5 hrs.) by the third or
fourth day of deprivation (Potter and Heron, 1972). Also in the present
study, total sleep time during the second 24 hours of bedrest decreased
almost 25I when compared with that of the first 24 hours. It might be
speculated that the decline would have continued, if bedrest had continued.
Nevertheless, with the exception of the perceptual deprivation
study cited above, all values of sleep satiation or fulfillment of
sleep "need" substsntialy exceed the 7- to 8~hour sleep durations
typically reported for normal young adults. The interpretation of such
findings may be generally separated into two opposing views. It has
been suggested, on the one hand, that the extended sleep observed in
unrestricted sleep regimens is a consequence of decreased sensory
input (see for example, Nakagawa, 1980; Heron, 1957). Proponents of


39
only one third of a group of individuals who completed 6-month sleep
diaries reported that they usually awakened spontaneously in the morn
ing. Further, half of the respondents in the study reported that they
were typically not well-rested upon awakening (Kleitman, Mull in, Cooper-
man, and Titelbaum, 1937). The extension of sleep on weekends reported
by most individuals also suggests a !need! for more than the typically
obtained weekday mean. Further, it has been shown that after 3 consecu
tive nights of sleep, restricted to a period between 11PM and 7AM (mean
= 7.5 hrs), subjects slept an average of 2 hours longer on the 4th night
when allowed to awaken spontaneously (Webb and Agnew, 1975a). The
subjects in that study were riot encouraged to extend sleep: indeed, they
were "unaware that they would be allowed to sleep longer."
Similarly, the subjects in the present study were given no in
structions relative to the maintenance of sleep and wakefulness during
the bedrest period, yet they also obtained substantially more sleep than
self-reported normal sleep durations. This seems to be suggestive of an
inclination toward, if not an actual 'need' for, more sleep than is
typically obtained. Of interest is the observation that the extended
sleep durations resulted not from a lengthening of sleep episodes
initiated at night, but rather from the addition of day sleep periods.
Nocturnal sleep (11PM to 7AM) per 24 hours did not exceed normal night
sleep (7.8 h). However, day sleep episodes contributed an average of
approximately 5 hours to each 24-hour total. It might be speculated
that the usual sacrifice of such day sleep periods is the source of our
chronic sleep deprivation. On the other hand, the bedrest environment
in the present study was highly monotonous, with novel sensory input
severely curtailed. It is, therefore, likely that such conditions


23
duration. This is compared to a mean duration of 225 minutes
(3.75 h) for those sleep episodes which contained both stage 4 and
REM.
The occurrence of these events in terms of sidereal time was
also examined. The three periods containing no REM and no stage 4
occurred after noon and before the onset of the night phase (11 PM).
All but 2 of the 'noREM' periods occurred between 2 PM and 10 PM.
The 1noStg41 episodes occurred across the entire day phase (7 AM to
11 PM) with only one such episode occurring at 'night' (6:40 AM).
Regarding the mode of appearance of REM sleep episodes, of 153
REM periods recorded over BP, only 3 were characterized by onsets at
the beginning of a sleep period. Sleep onset REM periods (SOREMPS)
were defined as those REM episodes occurring at the onset of sleep
which were not preceded by more than 1 minute of stage 2. Two sleep
episodes containing SOREMPS also were characterized by the absence of
stage 4 sleep. The SOREMPS were contributed by 3 subjects, had a
mean duration of 15.3 minutes (9, 10, and 27 minutes), and occurred
at approximately 6:30 AM, 6:30 PM, and 1:30 PM, respectively.
In an effort to further elucidate the factors involved in the
cycling of REM sleep, two examinations were made. In the first pro
cedure, the REM sleep cycle (defined as the interval between the on
set of one REM period and the onset of the next) was calculated for
each subject's baseline night (Night 2). The onset points of the
cycle were then superimposed across each subject's entire bedrest
period (including waking episodes) to test the hypothesis that the
REM cycle was independent of prior sleep time. Also, the cycle was


TABLE OF CONTENTS
CHAPTER PAGE
ACKNOWLEDGEMENTS .......... iii
ABSTRACT vii
ONE INTRODUCTION 1
Sleep as a Biological Rhythm ........ 1
Uniphasic or Ultradian ? 4
TWO METHODS 12
THREE RESULTS 15
Initial Sleep Period .... 15
Bedrest Period 17
Circadian Considerations 24
FOUR DISCUSSION
Total Sleep Time 35
Uniphasic or Ultradian? 49
APPENDIX
SLEEP QUESTIONNAIRE 53
REFERENCES 55
BIOGRAPHICAL SKETCH
60


47
in 85% of the sleep periods in this study (80% of day sleep periods).
Again by comparison, the naps of subjects studied by Moses et al.
contained REM episodes 43% of the time; 35% of the sleep opportunities
examined by Weitzman et al. contained REM and 48% contained SWS.
Seven of the 9 sleep periods which contained no REM sleep occurred
between 2PM and 10PM. Such a finding is consistent with previous ones
that REM sleep typically increases in amount from midnight to morning
and decreases toward evening (Moses et al., 1975; Nakagawa, 1980;
Weitzman et al., 1974). It is likely, then, that the absense of REM
during some sleep episodes in the present study were, at least in part,
due to circadian effects.
A similar possibility also exists relative to 'missing' periods
of SWS (specifically, stage 4). Several examinations of extended
sleep (see for example, Gagnon and Dekoninck, 1981; Webb, Campbell and
Hend in, 1981) have noted the re-occurrence of stage 4 (in relatively
large amounts) during the late morning. In the present study, sleep
episodes characterized by the absence of stage 4 occurred across the
entire day (6:40AM to 10:45 PM). It is interesting to note, however,
that an obvious 'gap' occurred between 10:30AM and 2:20PM. That is,
the sleep episodes initiated during those four hours always contained
stage 4 sleep. Sleep periods initiated after about 11PM, and
throughout the night, also contained stage 4. Thus, it might be
suggested that the absence of stage 4 in 15% of the sleep episodes in
this study, was the manifestation of the biphasic cycling of stage 4
sleep, rather than a reflection of the disruption of normal sleep
processes. A further indication of the possible bipolar nature of
stage 4 is the observation that total stage 4 time for BP was evenly
distributed between day and night (716 min vs 714 min).


PERCENT REM
30
T\D
VC
Figure 4.-Proportions of TST spent in REM sleep as a Unction of time of
day. Percentages were obtained by dividing the total minutes
of REM sleep for the group by total minutes of sleep, for each
time block across BP.


2
desynchronization is that which occurs between the temperature and
sleep/waking rhythms. When both rhythms are entrained to 24 hours,
the maximum temperature typically occurs just prior to retiring, and
the minimum temperature occurs toward the end of the sleep period.
However, under free-running conditions, maxima and minima of rectal
temperatures become independent of sleep onset and termination.
In addition, the sleep/waking rhythm may be entrained to arti
ficial light/dark regimens while typically in-phase rhythms maintain
their free-running periodicities. For example, throughout 10 days
of entrainment to a 3 hour sleep/waking cycle, subjects' temperatures
continued to exhibit characteristic circadian periodicities (Weitzman,
Nogeire, Perlow, Fukushima, Sassin, McGregor, Gallagher, and
Heilman, 1974). Such data "contradict any hypothesis of a simple
'cause-and-effect' relationship between the activity-rest cycle and
the rhythm of body temperature, and suggest a system of two coupled
oscillations . with a mutual phase relationship which depends on
the conditions" (Aschoff, 1978; p. 740).
Further evidence of the sleep/waking cycle's autonomy relative
to other biological rhythms may be seen in the results of sleep
deprivation studies. In the absence of sleep (for more than 70 hours)
measures of performance, temperature and endocrine function all
showed continuation of endogenous ciradian rhythms (Murry, Williams,
and Lubin, 1958; Froberg, Karlsson, Levi, and Lidberg, 1975, as
cited by Aschoff, 1978).
Examinations of the maturation of the human sleep/waking rhythm
have also provided evidence for its innate and autonomous nature.


Copyright 1981
by
Scott S. Campbell


54
15. What is your average bedtime on weekends? ^
(to~~nearest 15 min.)
16. What is your average wake up time on weekdays?
(to nearest 15 minTj
17. What is your average wake up time on weekends?
(to nearest 15 min.j
18. How well do you usually sleep at night?
Very well, satisfactorily, some problems, poorly.
IS. Do you usually feel well rested when you wake up, or soon thereafter?
Almost never, occasionally, often, almost always
20. On the average, how many days per week do you usually go to bed
more than one (1) hour earlier or later than your average bedtime?
0, 1, 2, 3, 4, 5, 6, 7
21. On the average, how many days per week do you usually wake up more
than one (1) hour earlier or later than your usual wake up time?
0, 1, 2, 3, 4, 5, 6, 7
22. Do you consider your sleep very light or very deep? Check below,
very light very deep
23. Do you feel tired during the day because you have slept poorly?
Never sometimes often always
24. How much alcohol do you drink?
Never sometimes everyday
25. Do you wake up before you intend to?
never
once or twice a month
once or twice a week
nearly every day
26. Do you use sleeping pills to get to sleep?
never
about once a month
about once a week
nearly nightly
27. What, if any, sleeping pills do you use?
28. What, if any, prescription drugs are you presently taking?


Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
THE EFFECTS OF EXTENDED BED REST ON
HUMAN SLEEP PATTERNS
By
Scott S. Campbell
August 1581
Chairman: Wilse B. Webb
Major Department: Department of Psychology
That the sleep response is one of numerous physiological functions
which may be classified as a biological rhythm is well-established.
Less well-founded, despite its popularity, is the notion that the
sleep of adult humans maintains an inherent tendency to occur mono-
phasically within each 24 hours, typically during the nocturnal por
tion of the 24-hour day.
This d issertation addresses the hypothesis that self-imposed and/'
or sociophysically mediated behavioral controls have a significant
influence on the manifestation of such a uni phasic sleep pattern,
and that attenuation of behavioral controls may result in a "loosening11
of the circadian framework of typical adult sleep.
Each of 9 young adults was electrographically recorded over a
period of 60 hours, during which behavior was restricted primarily to
lying quitely in bed; i.e., subjects were not permitted to read, write,
vn


TABLE III. Distribution of bedrest sleep periods by the presence or absence of REM and Stage 4 sleep.
Total
#
Day
Night
Mean duration
(range)
iime of occurrence
All stages
present
59
36
23
3.75 hrs
(.87 to 11.56 hrs)
No stage 4
13
12
1
1.41 hrs
(.5 to 2.9 hrs)
6:40am, 7:40am, 8:00am, 8:30am,
10:00am ,10:30am, 2:20pm, 3:10pm,
3:30pm, 4:30pm, 6:20pm, 9:00pm,
10:45pm
Mo stage REM
9
8
1
1,08 hrs
(34 to 161 min)
2:40am, 10:20am, 2:05pm, 3:10pm,
4:00pm, 4:30pm, 6:05pm, 7:20pm,
9:55pm
Neither REM
nor stage 4
3
3
0
1.02 hrs
(38 to 85 min)
12:05pm, 2:40pm, 10:30pm


LIST OF TABLES
TABLE
PAGE
I
Comparisons of sleep stage characteristics for
baseline (Night 2), bedrest period, and norma
tive group.
16
II
Individual and group statistics for the bedrest
period (BP)
20
III
Distribution of bedrest sleep periods by the
presence or absence of REM and Stage 4 sleep.
21
IV
Comparison of sleep variables as a function of
night (11 PM 7 AM) and day (7 AM 11 PM).
25
V
Sleep onset times of longest sleep episodes for
Night 2 (baseline), 3, and 4.
28
VI
Comparison of sleep variables between first and
second 24 hours of bedrest period.
31
v


57
Nakagawa, Y. Continuous Observation of EEG Patterns at Night and
in Daytime of Normal Subjects Under Restrained Conditions I,
Quiescent State When Lying Down. Electroenceph. Clin. Neuro-
physiot, 1980,49(5-6), 524-537.
Parmellee, A,, Wenner, W. and Schulz, H. Infant Sleep Patterns
From Birth to 16 Weeks of Age. J. Peeliat., 1964, 6_5, 576-582.
Potter, W. and Heron, W. Sleep During Perceptual Deprivation.
Brain Research, 1972, 40, 534-539.
Richter, C. R. Biological Clocks in Medicine and Psychiatry, Spring-
field, Illinois: C. C. Thomas, 1965.
Ryback, R., and Lewis, 0. Effects of Prolonged Bed Rest on EEG
Sleep Patterns in Young, Healthy Volunteers. EEG and Clinical
Neurophysiol, 1971, 3J_, 395-399.
Ryback, R., Lewis, 0. and Lessard, C. Psychobiologic Effect's of
Prolonged Bed Rest (Weightless) in Young, Healthy Volunteers
(Study 2). Aerospace Medicine, 1971, 42(5), 529-535.
Salzarulo, P., Fagioloi, I., Salomon, F., Ricour, C., Raimbault,
G., Ambrosi, S., Cicchi, V., Duhamel, J. and Rigoard, M. Sleep
Patterns in Infants Under Continuous Feeding From Birth. EEG
and Cl in. Neurophysiol., 1980, 49, 330-336.
Sander, L., Stechler, G., Burns, P. and Julia, H. Early Mother-
Infant Interaction and 24-Hour Patterns of Activity and Sleep.
J. Am. Acad. Child. Psychiat.., 1970, 9, 103-123.
Snyder, F. Discussion in Chase, M. H., The Sleep Brain. Perspec-
' fives in the Brain Sciences, Brain Information Service/Brain
Research Institute, Vol. 1, p. 45, Los Angeles: UCLA, 1972.
Sterman, M. Ontogeny of Sleep: Implications for Function. In
Drucker-Colin, Shkurovich and Sterman (Eds.). The Functions
of Sleep, pp. 207-232, New York: Academic Press, 1979.
Symposium on Quantitative Biology, Vol. XXV, Biological Clocks.
Cold Springs Harbor, Long Island Cold Springs Harbor Symp.,-
Quant. Biol., 1960, 25.
Verdone, P. Sleep Satiation-Extended Sleep in Normal Subjects.
Electroenceph. Clin. Neurophysiol., 1968, 24(5), 417-423.
Webb, W. Twenty-Four Hour Sleep Cycling. In Kales, A. H. (Ed.),
Sleep: Physiology and Pathology, pp. 53-65, Philadelphia:
J. B. Lippincott Co., 1969.


TABLE V.
- Sleep
onset times
of longest
sleep
episodes
for Nights
2 (baseline), 3,
and 4.
Subject #
928
929
932
936
937
940
941
945
946
Mean
Night 2
11:48P
11:27
11 :54
11 :42
11:50
11:57
12:05A
12:15
12:18
11:55P
3
12:30A
12:34
3:30A
10:50
2:42
2:42
2:58
11:45
9:39
1:01A
4
12:07
2:54
12:04
2:15
4:33
2:28
11:30
11 :21
2:14
1:43A
IND
00


ACKNOWLEDGEMENTS
The author would like to thank the members of his supervisory
committee for their assistance and criticism during the preparation
of this dissertation. The author would especially like to express
his deep appreciation to Dr. Wilse B. Webb whose encouragement,
guidance, time, and patience provided a constant source of motivation
and inspiration. Thanks likewise go to Briggs, Lewis, Alton, and
Gregg whose technical assistance and friendship were invaluable.
And final thanks to my parents for their support, both moral and
financial, without which the completion of this dissertation would
have been impossible.


50
Such a pattern of variability is similar to that observed for
the REM sleep cycle. The 90 to 100 minute REM cycle is of limited
value in predicting the occurrence of REM episodes within the sleep
record of a single subject. Yet for a population, the predictability
of a 90-100 minute cycle is quite good.
A further indication of the regularity of the quadriphasic sleep/
waking cycle is the observation that the overall cycle length did
not change significantly from one 24 hour period to the next. On
the other hand, the length of the sleep/waking cycle did differ as a
function of day and night, with the night cycle being substantially
longer (7.8 hrs. vs 5.2 hrs.). It might be speculated that this
difference was the consequence of subjects having spent the majority
of their lives entrained to a monophasic sleep system, and that given
time, the difference in cycle lengths might diminish. The present
data do not support such a notion, however since the difference
between night sleep/waking cycle length and day sleep/waking cycle
length did not lessen significantly in the second 24 hours of bed
rest.
The polyphasic distribution of the sleep of numerous animals
under experimental conditions is well documented. Berger (1972)
has noted that "if we wish to study comparative aspects of normal
sleep, we shouldn't impose wakefulness at any time on any species,
including the human, but should take 24-hour recordings under equiva
lent conditions." The relatively basal activity levels, absence of
novel perceptual stimulation and the satisfaction of bodily functions
characteristic of the present experimental situation are comparable


13
At the completion of electrode placement each subject retired
to a private, electrically shielded, sound attenuated, temperature
controlled room. Each subject's record was scored in one-minute
epochs, following the methods described by Agnew and Webb (1972a).
Cross-scoring of random 2-3 hour segments of the sleep portions of
each record revealed a between-scorer reliability of greater than
90%.
Sleep onset was determined by the appearance of the first
spindle or K-complex (Agnew & Webb, 1972b). A sleep period was
considered as such when sleep continued, uninterrupted by more than
20 minutes of wakefulness, for at least 30 minutes. A waking period
was defined as Stage 0 EEG for at least 20 minutes.
Timing of the 60 hours of bedrest began with the spontaneous
awakening of each subject from his/her second night of laboratory
sleep and the subject's verbal report that he or she felt rested
and had had enough sleep. At this time subjects were given an op
portunity to eat breakfast and use the bathroom. No intentional
time cues were given the subjects upon awakening nor at any other
time throughout the bedrest period, with the exception that each
subject was told when he or she was "past the halfway point" at a
random time after 30 hours had past. However, uncontrollable, gen
eral time cues from exogenous sources (eg., nearby building construc
tion, general building noise during working hours, homecoming fes
tivities during one session (subjects 945 & 946)) were present in
varying degrees from session to session.
During the 60 hours of bedrest, subjects were allowed minimal
exogenous stimulation. Activities such as reading, writing, listening


APPENDIX 1
SLEEP QUESTIONNAIRE
1. How many hours, on the average, do you usually sleep per night?
less, 5, 5%, 6, 6k, 7, Ik, 8, 8k, 9, 9k, 10, more
2. How many hours of sleep pen night do you think that you need to
feel well and function adequately?
3. How many hours of sleep per night do you prefer to have when you
have a chance?
4. In general, the amount of sleep that you usually get is:
not enough about enough too much
5. On the average, about how many naps do you usually take each week?
0, 1, 2, 3, 4, 5, 6, 7, 8, or more
6. About how many total hours do you usually spend each week taking
naps?
7. About how many times on the average do you wake up each night?
0, 1, 2, 3, 4, 5, 6, 7, 8, or more
8. About how much total time do you spend awake after going to sleep?,
none, less than 5 min., 5-15, 16-30, 31-60, more
9. About how many minutes does it typically take for you to go to
sleep? less than 5, 5-15, 16-30, 31-60, more than 1 hour
10. How often do you have trouble getting to sleep as quickly as you
would like? Almost never, occasionally, often, almost always
11. How much do you enjoy sleep?
not at all, a little, moderately, much
12. How regular are your bedtimes?
very regular, somewhat regular, somewhat irregular, very
irregular
13. How regular are your wake-up times?
very regular, somewhat regular, somewhat irregular, very
irregular
14. What is your average bedtime during weekdays?
Tto nearest 15 min.7
53


listen to music, watch TV, exercise, etc. Heals were served at random
intervals, and subjects had no knowledge of time of day throughout, the
bedrest period (BP).
The data indicate that a reduction in behavioral controls, which
typically influence the sleep process, results in substantial altera
tions in the placement and periodicity of sleep. Sleep episodes became
evenly dispersed across the 24-hour day, resulting in an approximate 6
hour sleep/waking cycle.
While the average sleep period duration decreased, mean total
sleep time for the bedrest period (28 h) exceeded, by more than 40%,
that which would be expected under the assumption of a uni phasic sleep/
waking rhythm. The continued tendency toward a nocturnal weighting of
total sleep was clear, however, since 55% of TST for BP occurred during
the night (11PM to 7AM).
Within sleep structure (e.g., sleep stage percentages, mode of
appearance) was affected only slightly by the "loosening'-' of the sleep
response. REM sleep retained its tendency to occur differentially
across the 24-hour day; stage 4 sleep appeared to maintain its rela
tionship to prior wakefulness.
It is suggested that the sleep/waking patterns observed in this
study may be further manifestations of an inborn (albeit, highly,
dampened) ultradian sleep/waking rhythm.
vii i


42
to Toss of most motor performance in the waking state," as well as the
"absence of highly focused attention in complex motor activity." The
finding by Ryback and Lewis (1971), that subjects who were permitted no
exercise while confined to bed exhibited decreased REM sleep, while
those who were allowed to exercise during bedrest did not exhibit such
a decline, also provides some support for- this notion. In the present
study, the finding that REMS percent decreased (albeit slightly) from
the first to the second 24 hours, may likewise be interpreted as
supporting an immobilization/sensory habituation effect on the amount
of REM sleep obtained.
A rather unexpected finding in the present study was that of an
increase (though not significant) in the proportion of SWS, as the
bedrest period progressed. In their examination of patients with high
crevical lesions, Adey et al. (1968), reported the diminution of SWS
as well as decreased REM percentage. As with REM sleep, the authors
concluded that such a decline was the consequence of immobility. Thus,
the findings of the present study appear contrary to what would be
expected by an immobilization hypothesis.
On the other hand, the results reported here are in accord with
the findings relative to SWS reported by Ryback and Lewis (1971).
The bedrest subjects in that study exhibited increases in SWS with the
nonexercise group showing the larger rise. These results were inter
preted by the authors as indicating that SWS may serve to "repair or
maintain the muscular system" in response to the effects of atrophy,
as well as hypertrophy. Thus, it would appear that a change in the
proportion of SWS, in either direction, may be explained by one or the
other version of an immobilization hypothesis.


Sleep Slage
/
inI ^TJITLT
i u-r
I2M
6AM
f*
6PM
I2M
^ui
i r
pAjitl
r^i
UU
l:!r i
6AM
I2N
6 PM
12M
6AM
11
I2N
6PM
gure 2.
Example of sleep staging across BP (subject 932).
are not shown.
Awakenings of less than 20 minutes


14
to music or watching television were prohibited. Conversations'
during the period were limited to brief dialogues with the experi
menter at meal times. No instructions were given relative to when
or when not to sleep. Subjects were asked to lie as quietly as pos
sible, but were permitted to change positions in bed when inclined
to do so. Closed circuit TV monitors were employed to insure adher
ence to the experimental instructions. Subjects were permitted to
go to the bathroom (in room) at their convenience and were allowed
to sit up in bed during meals.
Meals were served unsystematically (always during an ongoing
waking period, as determined by the EEC) and consisted of a choice
of several types of sandwiches, vegetable dishes, desserts and non-
caffeinated beverages. A maximum of ^ hour was permitted for the
completion of each meal. Several types of fruit and a pitcher of
ice water were also available to each subject ad libitum.
Illumination in each room was provided by a 60 watt lamp placed
on a table by the subject's bed, and control of light and control of
light and darkness was at the discretion of each subject.


56
Froberg, J., Karlsson, C., Levi, L. and Lidberg, L. Psychobiological
Circadian Rhythms During a 72 Hour Vigil. Forsvarsmdicin, 1975,
VU 192-201.'
Gagnon, P., and De Koninck. Reappareance of EEG Slow Waves in Extended
S1eep. Sleep Research, 1981 (in press).
Gesell, A., and Amatruda, C. The Embryology of Behavior, New York:
Harper, 1945, pp. 144-159.
Globus, G. Rapid Eye Movement Cycle in Real Time. Arch. Gen,.
Psychiat., 1966, 1_5, 654-659.
Heron, W. The Pathology of Boredom. Scientific Amer., January,
1957, 1, 430-436.
Johns, M., Gay, T., Goodyear, M. and Masterton, J. Sleep Habits
of Healthy Young Adults: Use of a Sleep Questionnaire. Brit.
J. Prev. Soc. Med., 1971, 25, 236-241.
Kleitman, N. Sleep and Wakefulness, Chicago: University of
Chicago Press, 1963.
Kleitman, N., and Engelmann, T. Sleep Characteristics of Infants.
J. Applied Physiology, 1953, 6_, 269-282.
Kleitman, N., Mull in, F., Cooperman, N. and Titelbaum, S. Sleep
Characteristies, Chicago: University of Chicago Press, 1937.
Lewis, H. Sleep Patterns on Polar Expeditions. In G. Wolstenholme
and M. O'Connor (Eds.), CIBA Foundation Symposium on the Nature
of Sleep, pp. 322-328, London: J. and A. Churchill, 1951.
Lewis, H., and Masterton, J. Sleep and Wakefulness in the Arctic.
Lancet, 1957, 1, 1262-1266.
Marn, L., Rechtschaffen, D. and Wolpert, B. Sleep Cycle During
Napping. Arch. Gen. Psychiat., 1964, J_}, 503-508.
Moses, J., Herd, D., Lubin, A., Johnson, L. and Naitoh, P. Dynamics
of Nap Sleep During a 40 Hour Period. Electoenceph Clin.
Neurophysiol., 1975, 39, 627-633.
Moses, J., Lubin, A., Johnson, L. and Naitoh, P. Rapid Eye Movement
Cycle is a Sleep-Dependent Rhythm. Nature, 1977, 265(5592),
360-361.
Murry, E., Williams, H. and Lubin, A. Body Temperature and Physcho-
logical Ratings During Sleep Deprivation. J. Exp. Psychology,
1958, 56(3), 271-274.


REFERENCES
Adey, W., Bors, E. and Porter, R. EEG Sleep Patterns After High
Cervical Lesions in Man. Arch. Neurol., 1968, ]9_, 377-383.
Agnew, H. and Webb, W. Sleep Stage Scoring. J. Suppl. Abstr.
Serv., Am. Psychol. Assoc., Ms., 1972a, No. 293.
Agnew, H. and Webb, W. Measurement of Sleep Onset by EEG Criteria.
Amer. J. EEG Techno!., 1972b, 12(3), 127-134.
Aschoff, J. Circadian Rhythms in Man. Science, 1965, 148, 1427-1432.
Aschoff, J. Features of Circadian Rhythms Relevant for the Design
of Shift Schedules. Ergonomics, 1978, 21 (10), 739-754.
Aschoff, J. and Wever, R. Human Circadian Rhythms: A Multi-
oscillatory System. Federation Proceedings, 1976, 35(12),
2326-2332.
Aserinsky, E. The Maximal Capacity for Sleep: Rapid Eye Movement
Density as an Index of Sleep Satiety. Biological Psychiatry,
1969, 1, 147-159.
Berger, R. Discussion in Chase, M. H. The Sleeping Brain, Perspec-
' tives in the Brain Sciences, Brain Information Service/Brain
Research Institute, Vol. 1, p. 45, Los Angeles: UCLA, 1972.
Carskadon, M. and Dement, W. Sleep Studies on a 90-minute Day.
Electroenceph Clin. Neurophysiol., 1975, 3_9, 145-155.
Colquhoun, W. P. Circadian Variations in Mental Efficiency. In
W. Colquhoun (Ed.), Biological Rhythms and Human Performance,
pp. 39-107, New York: Academic Press, 1971.
Ellingson, R. Ontogenesis of Sleep in the Human. In G. Lairy and
P. Salzarulo (Eds.), The Experimental Study of Human Sleep:
Methodological Problems, pp. 129-150, New York: Elsevier Pub.,
1975.
Flanigan, W. Phylogeny and Sleep Behavior: A'Review. Unpublished
manuscript, 1971.
Flanigan, W. Behavioral States and Electroencephaograms of Reptiles.
In M. Chase (Ed.), The Sleeping Brain, Brain Information Service/
Brain Research Institute, Los Angeles: UCLA, 1972.
55


I certify that I have read this study and that in my opinion
it conforms to acceptable standards of scholarly presentation and
is fully adequate, in scope and quality, as a dissertation for the
degree of Doctor of Philosophy.
Professor of Electrical
Engineering
This dissertation was submitted to the Graduate Faculty of the
Department of Psychology in the College of Liberal Arts and
Sciences and to the Graduate Council, and was accepted as partial
fulfillment of the requirements for the degree of Doctor of
Philosophy.
August 1979
Dean for Graduate Studies and
Research