Variation of sleep measures in infant monkeys

Material Information

Variation of sleep measures in infant monkeys an adaptation to rearing environment and diurnal events
Earle, Huberta Potter
Place of Publication:
Denver, CO
University of Colorado Denver
Publication Date:
Physical Description:
x, 81 leaves : illustrations ; 29 cm


Subjects / Keywords:
Monkeys -- Behavior ( lcsh )
Sleep ( lcsh )
Monkeys -- Behavior ( fast )
Sleep ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 73-81).
Submitted in partial fulfillment of the requirements for the degree, Master of Arts, Department of Integrative Biology
Statement of Responsibility:
by Huberta Potter Earle.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
16860260 ( OCLC )
LD1190.L45 1985m .E27 ( lcc )


This item has the following downloads:

Full Text


OF SLEEP MEASURES IN INFANT AN ADAPTATION TO REARING ENVIRONMENT AND DIURNAL EVENTS by Huberta Potter Earle ;::"' B.A., University of Colorado, 1978 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Arts Department of Biology 1985


This thesis for the Master of Arts degree by Huberta Potter Earle has been approved for the Department of Biology by Driscoll Da


Earle, Huberta Potter (M.A., Biology) Variation of Sleep Measures in Infant Monkeys: An Adaptation to Rearing Environment and Diurnal Events Thesis directed by Assistant Professor Gerald J. Audesirk Variability of nocturnal sleep measures was assessed in 11 group-living infant pigtailed macaques (Macaca nemestrina). Sleep measures examined were Sleep Latency, REM Latency, Number of Arousals, Number of REM Periods, Inter-REM-Interval, Mean REM Length, Number of Stage Changes, REM Efficiency, Sleep Efficiency, Total Sleep Time, and time spent in Awake, Drowsy and Stages Two, 3-4, and REM sleep. Sleep physiology, obtained using biotelemetry systems implanted when the animals were 5 to 8 months of age, was compared with behavioral categories representing the infants' behavior collected during the first month of life and at the time when physiological data were monitored. Behavioral categories represented mother-infant, distress, activity, and aggcessive/dominance behaviors as well as time infants' spent in proximity to other animals other than mother. Variability of the sleep patterns was evaluated, compared with other species, and examined as a function of the animals' early rearing environment and day to day behavior. Sleep variability was found to be greater between animals than within animals and species-specific


iv differences were observed. Consistent trends of relative variability of the sleep measures were observed within and between animals as well as across species. Age, maternal rank, and gender were not found to influence variability. Variability, within and between animals, of some sleep measures was observed to be partially predicted by behavior and environment during the first month of life. Interindividually, aggressive/dominance behaviors predicted 54% of the variability of REM Latency and other animal behaviors predicted 41% of Inter-REM-Interval. Intraindividually, distress and other-animal behaviors predicted 57% of the variability of Sleep distress behavior predicted 61% of REM and 70% of REM time was predicted by aggression/dominance behaviors. Daily behaviors and environment were not observed to have an effect on nightly sleep variation. These results suggest that variability of sleep may be a function of the early rearing environment in monkeys, and that normal daily behavior and environment may be less important. Variation of sleep patterns is evaluated as a functional adaptation to the biological niche that may increase the animal's chances for survival.


DEDICATION I dedicate this thesis to my mother Mary Pearson Mirkil.


vi ACKNOWLEDGEMENTS I would like to thank Gerald Audesirk Ph.D. for his help and support as my committee chairman and Janice Driscoll Ph.D. for her constructive criticism and statistical advise. I am most grateful to Martin Reite M.D. for his guidance and patience throughout the course of this project, and for the use of his laboratory. Most importantly, I would like to thank my husband, Michael, for his unquestionable faith that this project would be completed.


CONTENTS List of Tables ............................. i x List of Figures ............................. x CHAPTER INTRODUCTION . . . . . . . . . . . . . . . . 1 Theories of Sleep Function ... 3 Factors Influencing Sleep Patterns .. 4 Statement of Objectives .... 8 II. METHODS . . . . . . . . . . . . . . . . . . 1 0 Subjects and Housing ..... 10 Behavior Data Collection and Analysis 12 Physiology Data Collection and Analysis 19 Transmitter implantation surgery 19 The biotelemetry system, data acquisition, and transmission . 20 Collection and analysis of sleep physiology ............................ 22 Evaluation of Maternal Rank .. 27 Phase One ........................ 29 Phase Two .................... 30 I II. RESULTS . . . . . . . 31 Assessment of Sleep Variability: Phase One and Phase Two ... 31


viii Interindividual variation .......... 33 Intraindividual variation ........... 35 Phase One: Rearing Environment and Sleep Variability ............................. 39 Gender, group, maternal rank and age at run ............................ 39 Behavioral consolidation ............. 40 Early behavior and sleep variability 44 Phase Two: Daily Events and Sleep Variability ............................. 48 Behavioral consolidation .......... 48 Daily behavior and sleep physiology ... SO Cross-Species Comparisons ...... 55 IV. DISCUSSION . . . . . . . . . . . . . . . . 59 REFERENCES 7 3


ix TABLES Table 1. Description of Infants ................... 11 2. Behaviors and Definitions ................. 13 3. Sleep Measures and Definitions ............ 28 4a. Phase One: Mean Values and Standard Deviations of Sleep Measures ................. 32 4b. Phase Two: Mean Values and Standard Deviations of Sleep Measures ...................... 32 5. Interindividual and Intraindividual Variability Profiles for Phase One and Phase Two ..... 34 6. Phase One: Product-moment Correlations between Behavioral Categories and Individual Sleep Measure Means ...................... 46 7. Phase One: Product-moment Correlations between Behavioral Categories and Variation Coefficients of Sleep Measures .......... 48 8. Phase Two: Multiple Regressions of Behavior and Sleep Variables ................................ 53 9. Cross-Species Comparisons:Mean Values ........ 56 10. Variation Coefficients of Present Data and Bonnet Data ................... 58


X FIGURES Figure 1. Diagram of Monkey Pen ...................... 18 2. Transmitter and Battery before Implantation ................................... 21 3. An Example of Awake EEG 24 4. An Example of Drowsy EEG ........ 24 5. An Example of Stage Two Sleep .................. 25 6. An Example of Stage 3-4 Sleep .................. 25 7. An Example of REM Sleep ........................ 26 Sa. Phase One: Range of Intraindividual Variability Profiles of Sleep Measures by Animal ........... 37 8b. Phase Two: Range of Intraindividual Variability Profiles of Sleep Measures by Animal ........... 37 9. Variability Profiles from Both Phases: Within and Between Animals ..................... 38 10. Phase One: Behavioral Consolidation ............ 43 11. Phase Two: Behavioral Consolidation ......... 51 12. Variability Profiles of Different Species Repre-sented in Table 9 .............................. 58


CHAPTER I INTRODUCTION Sleep is a behavioral phenomenon that has received considerable investigation from the scientific community. Researchers have examined its architecture and rhythms, its phylogenie and ontogenic development, what disturbs it and what enhances it, and why it occurs at all. Nevertheless, many questions remain regarding this behavior that occupies one third of a human's lifetime. With the use of electroencephalographic technology, two general types of sleep have been observed to occur in all mammalian species that have been studied thus far (except for the primitive spiny anteater, which is a monotreme): non-rapid eye movement sleep (NREM) and rapid eye movement sleep (REM) (Allison and Goff 1968: Aserinsky and Kleitman 1955: Dement and Kleitman 1957a: Dement and Kleitman 1957b: Hauri 1982: Webb and Cartwright 1978). NREM is considered the quiet, slow-wave sleep (SWS), and consists of three stages, stages 1, 2, and delta sleep. REM is the active, fast sleep, also known as paradoxical sleep, or dream sleep. Actual sleep is composed of portions of each stage thus forming a


sleep pattern for a night, for an animal, or for a species. 2 Although REM and NREM sleep are present in the sleeping behavior of most mammals, there is a substantial amount of interspecific as well as intraspecific variation of the sleep pattern that is observed (Bert et al. Zepelin and Rechtschaffen Allison and Cicchetti 1976}. Further, nightly variation of sleep patterns is seen within individuals as well (Moses et al. Clausen et al. Reite et al. Reite and Short 1980}. That variability is present in sleep patterns is not surprising, as variation has been noted in virtually every aspect of living organisms since Darwin's work on variation in the 1800's. The study of variation in its own right, however, has generally been neglected in later research and instead has been accepted as a given character of nature. This is unfortunate to the extent that variation is thought to be a primary factor in the process of evolution and a better understanding of it may provide insight into the adaptive significance of the characteristic traits exhibited by living organisms (Simpson Yablokov 1974}. Further, if sleep variability is examined as a functional adaptation rather than a natural manifestation of living organisms, then we might better understand the purpose of this enigmatic behavior.


3 Theories of Sleep Function The adaptive significance, or function, of sleep is not well understood. Clearly there is a biological need for sleep, as can be seen by the disturbances and NREM and REM rebound resulting from studies of the effects of sleep deprivation (Agnew et al. 1967: Pearlman and Becker 1974: Tilley and Empson 1978: Lubin et al. 1974: Moldofsky and Scarisbrick 1976). However, studies have shown that these rebounds are not in a one-to-one relationship with actual sleep loss (Dement et al. 1970: Williams et al. 1964). Nevertheless, one of the most favored hypotheses is that sleep is a recovery period that the body uses to replenish itself from the fatigue of waking activity (Williams et al. 1973). Hauri (1982) suggests that SWS may be related to musculoskeletal recovery (although the support for this is conflicting, see below), and REM sleep may be related to psychologic recovery. Others have proposed that sleep provides restoration of the biochemical processes rather than the gross reparations of the skeletal and visceral systems (Hartmann 1973b: Moruzzi 1972: Oswold 1970). Another theory on the function of sleep suggests that sleep occurs in order to conserve energy and prevent non-adaptive, ineffective activity (Allison and Van Twyver 1970: Berger 1975: Kleitman 1963: Webb 1971: Zepelin and Rechtschaffen 1974). Webb suggests that inactivity may be as effectively connected to the


4 environment as food gathering and protection from pred-ators, thereby reserving energy stores and increasing survival. Whatever its function, it is clear that sleep is necessary for survival or it would not be evolutionarily favorable for the animal to spend so much time in such a vulnerable physiological state. Factors Influencing Sleep Patterns Bert (1972) has commented on the biological importance of the adaptability of sleep patterns in that "it allows the individual to cope with changes in the environment while still assuring the maximum of vigilance compatible with satisfaction of minimum sleep needs" (pp. 32-33). Researchers have noted certain aspects of sleep patterns tend to correlate with ecological influences. Predators tend to sleep more and adapt better to the laboratory environment than prey species. In addition, species that sleep in secure places sleep more than those who sleep in the open (Allison and Van Twyver 1970). Further, Allison and Cicchetti (1976) have found that interspecifically, both NREM and REM sleep seem to decrease as the danger of predation increases. This is not to say that environmental factors provide the only contributions to sleeping behavior. Constitutional factors, age, metabolic rate and other endogenous influences have also been associated with differences in sleep patterns (Allison and Cicchetti Balzamo et al. Balzamo et al. Williams et


5 al. 1974: Zepelin and Rechtschaffen 1974). It is clear that sleep is a modifiable entity, and the patterns that result are a reflection of many influential mechanisms. Bert and Pegram (1972) have proposed three general factors that influence sleep organization: a genetic factor, a learning factor, and an environmental factor. Clearly these important factors are intimately related and must all be considered in the overall model of the hows and whys of sleep behavior. Initially, however, in order to gain a better understanding of sleep, its patterns and variability, investigators must attempt to isolate each area as much as possible, and investigate it thoroughly. Intraindividual sleep patterns have been observed to vary with certain environmental influences. Generally these factors can fall into 2 categories: alterations of the sleeping environment and alterations in the daily environment and behavior. Changes in the sleeping environment include such factors as temperature, noise level and strange ments. Extremes in temperature have been observed to affect sleep patterns in cats and in man (Parmeggiani and Rabini 1970: Angus et al. 1979: Otto 1972). Noise levels during sleep have also been noted to disturb sleep patterns both short term and long term if the noise persists (Globus et al. 1974: Levere et al. 1972: Otto 1972: Roth et al. 1972). The first night effect, disturbances in the first night of sleep in the testing environment, though


not universal, is commonly found in subjects being recorded in the laboratory (Schmidt and Kaebling 1971: Webb and Campbell 1979: Browman and Cartwright 1980: 6 Williams et al. 1972). Differences in sleep patterns have been observed when subjects sleep alone or with their spouses (Monroe, 1969). Further, animals subjected to chair-restraints demonstrate an extensive restructuring of their sleep patterns (Bert et al. 1975: Bert 1972). Daily events and alterations in behavioral patterns, particularly those events that induce stress in the individual, have also been observed to have an influence on sleep patterns. Hartmann (1973a) found that sleepers needed more sleep when times were stressful than when times were going well. In a study examining insomniacs and good sleepers, Healey et al. (1981) found that insomniacs had increased stressful life events. Emde et al. (1971) examined sleep in circumsized neonates and found an increase in NREM time compared to uncircumsized neonates. Adams-Tucker (1982) found that sleep disturbance was one of ten chief complaints in sexually abused children. DeMonchy et al. (1980) found that disturbances in parent-child relationships are an important cause of sleep disturbances. Parental anxieties and inconsistent care of the infant have also been cited as affecting sleep patterns in young children (Anders et al. 1972: Moore et al. 1957: Hirschberg 1957). Further, cognitive presleep stress has been noted to affect sleep


7 patterns in insomniacs and in good sleepers, where sleep latency was observed to increase in the insomniacs and decrease in the good sleepers (Haynes et al. 1981). When examining the effect of stressful events on sleep, it is important to acknowledge that individual personalities, or temperament, may influence how the stress will affect the sleep patterns: an example of the interconnection of environmental and endogenous influences on sleep. Depending on their mental state, individuals may react differently to a similar stressful situation, as was seen in the study on cognitive stress (Haynes et al. 1981). It has been noted that anxiety, depression, and other manifestations of psychopathology affect sleep patterns (Monroe Weissbluth DeKoninck et al. Hartmann Rosa et al. Reynolds et al. Hartmann 1969). Exercise has received much investigation as having an effect on long term and short term sleep patterns (Montgomery et al. Trinder Bunnell et al. 1983). Primarily, the research has focused on the hypothesis that slow-wave sleep is fatigue-related and represents a recovery/restoration period following exercise. The results are conflicting, however, and may be due to the differing methodologies concerning the type of exercise, the fitness of the subjects and the time and duration of the activity (Torsvall Horne 1981).


8 If the factors that influence sleep patterns are thought of as "stressors to the system" in that stressor is defined as an alteration of the external environment which necessitates adaptation to restore homeostasis (Selye 1956), then one might think of the variability of sleep as the organism's attempt to restore homeostasis. When variability of sleep is examined as a function of these stressors, perhaps we might increase our understanding of its adaptive significance. Statement of Objectives Since the environment and behavior appear to have some influence on certain aspects of sleep, perhaps some of the night-to-night variability of sleep might be explained by events and actions that have occurred during the day. Further, some individuals seem to be more variable in their sleep patterns than others. Perhaps these individuals have acquired more flexibility in their sleep patterns as a result of conditions in their rearing environment. These are questions that are specifically addressed in this thesis. In an attempt to answer them, sleep physiology and behavior have been examined in group-living, unrestrained infant pigtailed macaques, Macaca nemestrina. With the use of implantable radio telemetry to collect sleep physiology, the animals in this study were able to live with their social groups in their "natural" undisturbed laboratory environment.


9 The rationale behind this study is three-fold. First, the variation of sleep patterns as represented by this population of primates is assessed. Second, this study examines whether variations of sleep patterns themselves vary with conditions of the early rearing environment. Third, night-to-night variability of sleep and its association with daily events and activities are explored. In order to simplify the presentation of results, the thesis is divided into two sections: Phase One and Phase Two. Phase One addresses the question regarding early rearing environment, and Phase Two deals with sleep variation as a function of the day to day environment. Assessment of sleep variability is examined in both.


CHAPTER II METHODS Methodology is the same for both Phase One and Phase Two except where indicated. Subjects and Housing The subjects studied were infant pigtailed macaques (Macaca nemestrina) born between July 1977 and March 1980 to feral mothers living in the University of Colorado Health Sciences Center Primate Laboratory in Denver, Colorado (Table 1). All infants were reared in 1 of 2 social harem groups, each consisting of 1 adult male (Beau or George), and 6-8 females, some with infants. The social groups were housed in pens 2.1 X 2.4 X 4.0 m in size with glazed cinderblock walls, cement floors, wiremesh ceilings, and one-way glass observation windows. Plastic and steel shelves, 0.5 X 1.2 m in size, and a 6.2 em diameter plastic pipe were positioned on one or more walls of the pen, 1.0 and 2.0 m from the floor respectively. Timer-controlled lights (-700 lux) illuminated the pens between 0700 and 2000 hours daily. Water was available ad lib and was supplied to each pen


11 Table 1. Description of Infants Animal Date of Age at Age at Ma Phase Number Birth Gender Group Implant Run l Rank 2 One Two (days) (days) 11.6 7/8/77 F Beau 133 150 1 X 23.3 7/31/77 F Beau 164 183 2 X 21.2 3/15/78 M George 197 211 2 X 8.6 4/11/78 M George 191 217 1 X 29.1 5/29/78 M Beau 228 245 3 X 23.4 8/25/78 F Beau 166 179 2 X 31.1 10/9/78 M Beau 135 168 3 X 8.7 6/20/79 F George 142 161 1 X 20.3 6/25/79 F George 164 212 2 X 11.8 11/12/79 F Beau 232 248 1 X 29.2 3/27/80 M Beau 111 159 3 X 1Age at Run is the animal's age on the first day physiological data were collected. 2Maternal rank by thirds: Top third= 1, middle third= 2, and bottom third= 3. X X X X X


12 by a 19 mm copper pipe with a self-plugging nozzle that the animals released by pressing with their tongues. Behavior Data Collection and Analysis The behavioral variables used in this study were derived from the behavioral taxonomy of Macaca nemestrina by Kaufman and Rosenblum (1966) and are presented and de-fined in Table 2. Behavior was collected which best represented aspects of the infant's behavior and environ-ment that were thought to be meaningful to the infant. These areas included the mother-infant bond, the infant's relationships with the other animals, aggressive behaviors, the infant's distress, and the infant's activity. Each infant was observed in random order by trained technicians who viewed the subjects through one-way glass windows on a regular basis. Reliability between the technicians was 95%. 1 Behaviors were recorded directly as they occurred using a Hazeltine Model 1500 computer terminal interfaced to a PDPll/23 minicomputer system. 1Reliability was determined by each technician observing the same behavior session. The resulting sessions were then analyzed as described in the behavior analysis section. The resulting analysis was then compared for behaviors observed, frequency, and duration. The overall result was a reliability of 95%.


13 Table 2. Behaviors and Definitional Behavior D/F2 Phase Mother-Infant Behavior Cradle-3 D 1,2 Enclose-D 1,2 Contact-D 2 Carry-D 1,2 Nipple D 1,2 Receive Groom-D 2 Definition Ventral-ventral clasping of infant by mother while using hands, actively pressing the infant to the ventral surface of her body. Ventral-ventral contact of infant to mother without active use of her hands. Mother's hands and feet are loosely flexed around infant's body. Any other physical touching of infant by mother. Infant is carried ventrally by mother while she is locomoting. Infant's mouth is on or around mother's nipple. Infant receives groom from mother. Grooming is defined as careful picking through and or brushing aside fur with one or both forepaws. 1Behaviors are modified from Kaufman and Rosenblum (1966). 2D/F designates whether the behavior is a Duration behavior or a Frequency behavior (see text). 3 A hyphen following a behavior indicates that the behavior is followed by a numerical modifier signifying another individual either mother, the adult male, an adult female, or another infant or juvenile.


Table 2 (continued). Receive Genital Exploration-Activity Behavior Locomotion Play-Auto-Play Initiate Play-Activity Count D D D D F F Other Animal Behavior Cradle-D Enclose-D Passive Support -D 2 1,2 1,2 2 2 1,2 1,2 1,2 1,2 14 Infant receives genital exploration from mother. Genital explor ation includes sniffing, picking or lick ing of genital area. Infant is walking, running or climbing around the pen. Infant is tumbling about, running or par taking in any behavior that is characterized by lightness, freedom of movement, a lack of tenseness, and an absence of stereotyped sequences. Infant plays with self. Infant initiates play with another animal. The number entered at the end of the session designating the number of sections that the infant entered during a session (see text). See mother-infant behaviors. See mother-infant behaviors. Infant resting against another animal without any active support from the animal.


Table 2 (continued). Contact Lady Carry-Receive Groom-Receive Genital Exploration-D 1,2 D 1,2 D 1, 2 D 1, 2 Behavior Fight Mother Receives Aggression-D/F F Receive Aggression-F Receive Punishment-F Receive Threat-F 1,2 1,2 1,2 1,2 1,2 Infant is physically touching an adult female other than mother. See behaviors. 15 Infant receives groom from another animal in the group. Infant receives genital exploration from another animal. Aggressive fighting in the pen by the group animals. Mother is bitten, grabbed, pulled, pushed, hit, or restrained by another animal. Infant receives ag gression from another animal. Infant receives a grab, bite, push, or is res trained by another animal. This is a brief encounter and is less severe than aggression discussed above. Infant receives a threatening gesture from another animal. A threat is a facial expression where one animal stares at another, raises its eyebrows, and drops its bottom jaw so that the mouth is 3/4 open.


Table 2 (continued). Submit-F Restrain-F Distress Behavior Oral Self D Temper Tantrum F Jerk F Squeal F Coo F Nipple Deny F Wean F 1, 2 2 1 1, 2 1,2 1,2 1,2 1,2 1,2 16 Infant smacks its lips together and withdraws from another animal. Infant is grabbed and held by another animal while the infant is locomoting and is therefore unable to get away. Infant puts any part of its body into its mouth for sucking, chewing, biting, or any other oral activity. Infant exhibits an uninterrupted sequence of high-pitched squealing and jerking. Infant's body jerks, usually accompanied by squeals. Unlike a Temper Tantrum, the jerk occurs in isolated episodes. Infant emits isolated squeals or squeaks. A long, pure sound emitted by the infant during periods of distress. The lips are generally pursed forward and rounded. Infant is denied the mother's nipple when it attempts to suck on it. The mother pushes the infant away before contacting the nipple. Infant is pushed off the nipple by its mother.


17 In addition to the observed behaviors, an activity count was also kept throughout each session. The group pen was divided into 20 approximately equal-sized imaginary cubes whose borders were obvious physical structures (Figure 1). When the infant was observed to cross from one section to another, a hand-held counter was incremented by 1 count. At the end of the behavior session, the observer recorded the total number of increments shown by the counter which represented that animal's activity count for the session. Thirty-three behaviors (including the activity count) were recorded and included 18 duration behaviors, having discreet start and stop times and finite durations, and 15 frequency behaviors which were unitary events with a single time of occurrence (Table 2). Numerical modifiers were used in conjunction with behaviors to indicate the recipient or initiator of particular behaviors where indicated. All behavior sessions containing start and stop times and sequences were stored on RKOS hard disks and analyzed on the computer. Stored sessions were pooled together and analyzed by animal for a specified period of time depending on the experimental design (for details, see methods under Phase One and Phase Two). All duration and frequency scores were standardized by computing the percent total time observed (%TTO) and frequency per 1000 seconds (F/lOOOs), respectively. For further discussion


WESH CEIUNG GRID Each of tr'le 8 MCtlon WATER PIPE l< I ---1 ,A: :.:-J--LADDER t : ) I I IMAGINARY FLOOR GRID Figure 1. Diagram of Monkey Pen. Grid sections for Activity Count are demonstrated by the number 1. When an infant travelled from one section to another, the Activity Count was incremented by 1. 00


of the behavior collection and analyzing system, see Reite and Short (1983). Physiology Data Collection and Analysis 19 Transmitter implantation surgery. Between the ages of 5 and 8 months, the infants were implanted with multi-channel biotelemetry systems for physiological data collection. One week prior to the surgical implantation, the infant and its mother were removed from the group to an isolation cage in the treatment room of the laboratory. Both animals were sedated with approximately 10 mg per kilogram of body weight (mg/kg) of Ketamine hydrochloride (Ketalar, Parke-Davis), and the infant weighed. The infant's torso was fitted with a small leather vest to mimic its appearance after surgery, thus minimizing maternal rejection. After recovery from the sedation, the mother and infant were returned to the group until the evening before surgery when they were again removed to the treatment room. The next morning prior to feeding, the mother and infant were again sedated. The infant was removed to the operating room and was given 15-20 mg/kg of intravenous sodium pentobarbital (Pentobarbital Sodium Injection, Elkins/Sinn Inc.) for general anesthesia. The transmitter and battery pack (see below) were implanted extraperitoneally beneath the anterior abdominal wall


20 musculature. An electrode cable was routed subcutaneously across the chest, around the back of the neck to the top of the head where electrodes were placed epidurally in the right hemisphere for electroencephalogram (EEG) recording. Further details of the surgical procedure have been documented elsewhere (Reite and Pauley 1976: Reite et al. 1973). Following recovery from anesthesia, the infant was returned to its mother in the treatment room where the two remained for ten days. During this time, the infant received 50 mg/kg per day of Prostaphlin, a syn-thetic penicillin (Oxacillin Sodium, Bristol Laboratories) and 0.1 mg/kg/day of Kefsol (Cephazolin Sodium, Lilly Co.). The animal was watched carefully for any signs of infection. The sutures and the leather vest were removed after 10 days. If no infection or other problem was apparent, the mother and the infant were returned to the group. The biotelemetry system, data acquisition, and transmission. The totally implantable biotelemetry systems were pulse-amplitude-modulated (PAM-FM) systems which transmitted seven channels of physiological data including heart rate, body temperature, eye movement (electro-oculogram, EOG), neck muscle activity (electromyogram, EMG), and 3 channels of electro-encephalogram (EEG): frontal, central-temporal, and parietal-occipital. Each biotelemetry system consisted of


Figure 2. Transmitter and Battery before Implantation. !=transmitter, 2=battery. two packages (Figure 2), one containing the hybrid 21 electronic transmitter and the other containing a GTE 900 mAh 3.6-Vol lithium cell which provided approximately 1000 hours of transmission. A magnetically activated bistable latching reed switch was connected in series with the battery and was controlled by passing a small magnet over the abdomen of the animal. The signal transmitted by the telemetry system was picked up by one of 6 dipole FM antennae which were located in the upper bars or in sealed plastic piping throughout the pen. An automatic antenna selector was used to select the antenna which transmitted the best signal depending on where the infant was in the pen


22 (Pauley and Reite 1977). The signal, sent through a Dynaco receiver and a PAM demodulator, was displayed on an oscilloscope and the transmitted physiology was recorded on paper using a Grass Model 78 polygraph with 7P511 amplifiers. Simultaneously, an time-code was recorded on the paper by a Datametrics Model SP425 code generator. (For details, see Pauley and Reite 1981: Reite and Short 1983.) Collection and analysis of sleep physiology. Af ter the mother and the infant had been back in the group for at least 3 days following the 10-day recuperation period, the transmitter was turned on with a magnet and physiological data were transmitted continuously 24 hours a day. To record sleep physiology, the paper polygraph record was turned on each night at 2000 hours (lights out time) and ran uninterrupted until 0700 hours the next morning (lights on time). The paper polygraph recorded 6 channels of data: time-code, EOG, EMG, and three channels of EEG. The paper speed was 15 millimeters per second resulting in one page representing 20 seconds of data. Thus one night of sleep was recorded on 1,980 pages of paper record. Criteria for sleep staging followed that described by Reite et al. (1965). Each page of the sleep record was classified as 1 of 5 categories of sleep and wakefulness: Awake, Drowsy, Stage Two, Stage Three and Four (combined), and Rapid Eye Movement Sleep (REM). In


23 addition, unreadable data were occasionally transmitted and were categorized as undefined. In this study, undefined sleep represented no more than 30 minutes in one night, or 4.5% of the sleep record. The Awake record was characterized by fast, low voltage activity on all three EEG leads, usually accompanied by high voltage activity on the EMG, as well as evidence of EOG activity (Figure 3). The Drowsy record, the first stage of NREM sleep, contained fast EEG activity as well as some slower but higher voltage activity present intermittently at approximately 12 cycles per second (cps). The EMG and EOG leads were less active than in the Awake record, but still demonstrated some bursts of activity (Figure 4). Stage 2, or light sleep, was characterized by low to medium voltage, slow 3 cps activity with occasional 12 cps sleep spindles and K complexes (Figure 5). The EOG and EMG did not show the bursts of activity typical of arousal. Stage Two sleep is the second stage included in NREM sleep. Stage 3-4 sleep was a composite category representing both stage 3 and stage 4, or deep sleep, and is also classified as NREM sleep. This stage was characterized by very slow 1-2 cps high voltage activity in the EEG with diminished muscle activity and no evidence of eye movements (Figure 6).


24 EOG FR EEG CT EEG P-0 EEC Figure 3. An Example of Awake EEG. Note the fast, low voltage of the 3 EEG leads and the high actvity on the EMG and EOG leads. TC=time code, EOG=electro-oculogram, EMG= electromyogram, FR=frontal EEG, CT=central-temporal EEG, PO=parietal-occipital EEG. L I, l I\$J blof>W

25 EOG .i' K CO HPL EX "-sLEEP SPiliiDL.E FB EEG CT EEG P-O EEC Figure 5. An Example of Stage Two Sleep. Note the low to medium voltage EEG, sleep spindles and K complexes on the EEG channels and the lack of activity on the EOG and EMG leads. EOG FR EEC CT EEG P-O EEC Figure 6. An Example of Stage 3-4 Sleep. Note the very slow high voltage activity on the EEG leads. The activity on the EOG and EMG leads is artifact created by the large slow waves of the EEG.


26 EOG FR EECi CT EEG P-0 EEG Figure 7. An Example of REM Sleep. Note the rhythmicity of the EEG leads and the inactive EMG lead. Particularly note the large eye movements on the EOG lead. The EEG during REM sleep was typically low voltage fast activity but was slower and more rhythmical than the Awake EEG. During REM sleep, the infants exhibited evidence of rapid eye movements (REMs) on the EOG channel. The EMG channel generally remained at very low voltage throughout this period except for occasional brief body movements, or phasic bursts (Figure 7). Each page of the sleep record was visually scored and the stage assigned to each page was entered into a computer file. The entries began at lights out and con tinued until lights on time or at the infant's final arousal, whichever came first.


27 When the entire paper record was scored, the resulting computer file was analyzed by a program developed in the laboratory called SLEEPSTAGE. This program read each sleep file individually and calculated and printed a set of sleep measures for further analysis. The measures included: Sleep Latency (min), REM Latency (min), Number of Arousals, total time (min) spent in Awake, Drowsy, Stage Two, Stage 3-4, and REM sleep, Number of REM Periods, Mean REM Length (min), Inter-REM-Interval (min), Sleep Efficiency (%), REM Efficiency (%), Number of Stage Changes, and Total Sleep Time (min). A list of these sleep measures, their abbreviations and their definitions can be seen in Table 3. Evaluation of Maternal Rank To quantify the dominance hierarchy in each group, a 24-hour water deprivation test (Boelkins 1967) was performed every 3 months. The hierarchy, once established, never varied during the recording period of a particular infant. The water was turned off at 1300 hours the day before the test and remained off for 24 hours. When the water was turned on, the group was carefully observed for dominance-related behaviors. The sequence of the animals as they arrived at the drinking pipe to drink, the duration of drinking time, and the direction of displace-ment and fighting behaviors were recorded. A linear hierarchy was constructed from these data with the most


28 Table 3. Sleep Measures and Definitions Measure Abbreviation Sleep Latency REM Latency Number of Arousals Number of REM Periods Inter-REM-Interval Mean REM length Awake Drowsy Stage Two Stage 3-4 REM Efficiency Sleep Efficiency Tbtal Sleep Time Number of Stage Changes SLPIAT REMI.AT IRI SLPEFF TST Definition Time (min) from lights out to sleep onset as defined by 1 minute of uninterrupted sleep (any stage) Time (min) from sleep onset to onset of the first REM perioo Number of arousals to awake during the night Number of REM perioos Mean time (min) between onset of sequential REM periods Mean length of REM periods (min) Tbtal time (min) spent awake between sleep onset and final arousal Total time (min) spent in Drowsy sleep. Total time (min) spent in stage two sleep. Tbtal time spent (min) in stages three and four sleep Total time (min) spent in REM sleep Percent of time spent with REM Em patterns of total time in all REM periods, including interruptions Percent of time spent asleep between sleep onset to final arousal Tbtal sleep time (min) not including time awake The number of stage changes after sleep onset


29 dominant animal (the adult male) at the top. The hier archy was divided into thirds with the highest, middle, and the lowest ranking animals designated as groups 1, 2, and 3 respectively. The adults were the only animals observed for this test. The infants tended to drink with their mothers or opportunistically and therefore were assigned the ranks of their mothers. Phase One Eleven infants, 6 females and 5 males, were studied in Phase One (Table 1). Each infant was observed for two 3-minute behavior sessions 5 days a week, begin ning with birth through the first month of life. Behavior sessions on all laboratory infants were performed once in the morning and once in the afternoon to get a general daily sample of each infant's behavior. The sessions were limited to 3 minutes due to the large number of animals that needed to be observed. All behavior collected during this time period for each animal was pooled together and used for further analysis in Phase One. Physiological data were collected when the animals were 5 to 8 months of age. Four nights of sleep were collected for each infant. Means and standard deviations of each sleep measure were computed across the 4 nights and were the values used for further analysis in Phase One.


30 Phase Two Five of the 11 animals studied in Phase One, 3 females and 2 males, were studied in Phase Two (Table 1). At 5 to 8 months of age, each infant was observed for six Sminute behavior sessions, three each morning and three each afternoon, for 10 consecutive days. Behavior was analyzed on a daily basis for each animal. Physiological data were obtained each night following each day of behavior. Each night for all animals was analyzed separately and the values obtained were used for further analysis.


CHAPTER III RESULTS Assessment of Sleep Variability: Phase One and Phase Two Each sleep measure was evaluated for inter-individual and intraindividual variability to determine the normal variations of sleep in infant monkeys as represented by the animals in this study. To be thorough, the analysis was done with the data from both Phase One and Phase Two, as each represents a slightly different sample. A one-way analysis of variance was computed for each sleep measure in both phases across all the animals with animal as the grouping factorl. Coefficients of variation of all sleep measures for each phase were also computed. Mean values and standard deviations for the sleep measures for each of the 11 animals in Phase One can be seen in Table 4a, and for each of the 5 animals of Phase Two in Table 4b. 1The purpose of this ANOVA is to document the differences between the means of each of the animals and examine the magnitudes of the variances within and between them. It has been argued that consecutive nights of sleep within an individual are not independent due to the possibility that one night's sleep may be affected by the quality of sleep of the preceding night. Since this paper is examining the affects of daily environmental and behavioral events on the quality of the sleep pattern, the nights of sleep are considered as independent events.


Table 4a. Phase One: Mean Values and Standard Deviations of Sleep Measures AROUS R!HP AN IlK! STO l-4 R!H IRI HEAMRL R!H!FF SLP!FF TST STOCHO SLPLAT 1\NIHI\L __! I 1 I I I I I i lei I I ---8 6 8 7 11. 6 11.8 20 ) 21.2 2). J 2). 4 29. I 29. 2 )I. l 5 4 17 17 27 16 90 l4 29 16 70 24 II 7 6 ) 45 18 18 12 22 20 68 17 88 44 )9 20 57 4 94 41 liZ 44 112 l1 124 ll 74 10 124 )0 78 19 26 10 21 1 26 16 )2 10 29 4 )0 8 18 6 20 6 14 6 20 15 4 o.o o o 9.) o.s 9.5'1.) 9.0 0.8 9.0 1.8 8.5 1.1 9.5 0.6 11.) 1.) 9.0 1 6 9.o o.8 10.5 t.) 98 11 69 24 50 19 98 )1 10) 14 82 25 59 20 76 17 4 5 17 l2 57 l9 16 l 22 2 19 6 ll 1 II 2 2S 2 15 l 15 2 tO 2 28 4 10 2 260 41 252 21 281 27 215 I) 211 11 241 29 285 26 261 27 244 29 275 12 299 u 179 14 169 9 141 II ll8 to 156 9 148 14 18) 8 181 16 222 ]2 159 14 190 21 76 100 107 74 71 8) 9) 99 9) 62 82 9 75.6 1.9 11 61.5 1.5 1 61.5 5.4 1 56.9 . 7 62.) 9.5 12 58.0 7 ) 10 60.1 2.0 10 so.o 5.1 10 6).2 8.8 58.1 . 6 56. 7 6.7 11.5 2.0 87 4 1).) 1.6 8) l ll. l 1.7 82 l 10.6 1.1 79 I ll.S 1.1 75 10 12.2 1 1 81 5 11.7 0.7 84 5 11.1 2.1 85 l 12.8 l.l 82 4 8.9 1.7 82 8 9.9 0.9 79 l 86 6 90 5 90 5 82 6 8) 2 115 5 90 4 88 l 94 I 81 5 91 4 518 18 556 II 551 17 447 24 5ll 1 500 51 577 24 579 Jl 569 )0 529 t9 581 16 ))6 47 ll8 29 280 62 )58 27 ]54 51 152 27 109 64 296 JO 289 14 296 21 284 49 CROUP ll 21 88 28 2) 6 9.1 0.9 71 22 19 1 26) 24 170 l5 86 14 60.4 6.2 11.6 82 l 88 4 540 41 ll7 10 Table 4b. Phase Two: Mean Values and Standard Deviations of Sleep Measures R!HLAT 1\ROUS R!HP AWAKE DROWSY TWO STO l-4 R!H IRI R!HEFF SLP!FF TST STOCHO SLPLAT 1\NIHI\L -.NO. I I e I e I e I elel e I e I e I e I I I I I -------8 7 :24 20. J 28 2).) 41 29.2 28 ll.l 22 -19 25 1 II 16 81 86 112 Ill 74 15 4l 58 )I 52 28 It 9.2 0 8 74 25 24 1 261 21 156 IS )) 11 9.1 l.t 81 29 17 l 275 12 161 12 19 4 9.8 0.6 58 ll 14 1 )07 29 161 2] 21 6 8.6 1.0 89 26 26 4 274 It 167 ll 14 6 10.0 1.) 46 25 10 l 109 26 185 17 101 79 101 61 87 9 1 II 12 8 61 2 62.5 59.) 62.8 60.0 ) 8 6.9 1.7 8.7 9.9 -------11 1 1.6 81 l 89 4 550 12.2 1.1 74 7 87 5 5)) 12.8 1.7 8) 5 90 ) 585 9.4 1.5 76 9 86 4 529 __ !!-! _ II 21 17 20 25 _ 6 94 19 21 8 9. 1 0. 6 70 17 18 7 286 21 166 II !!..._!2__!.!._:_!_1_._5 _ __ --!!_4 _ 89 __ .. -558 29 11111 In lnutee eacept AROUS, R!HP, end STGCHO which frequency of end REHEFF and SLPEFF which devlatlone for the entire group repreeent the variability of the aeane for each Measure. 346 lSI )04 101 287 )l 48 )9 28 )8 JIB 10 w tv


33 Interindividual variation. Results of the analyses of variance of Phase One show that variability between animals is significantly greater than variability within animals for all sleep measures except REMEFF. For Phase Two, results showed significantly greater variability between animals than within animals for all sleep measures except for SLPLAT and IRI. The discrepancy of results between the two phases of these three measures is likely to be due to sampling differences. Since the Fvalues of the complementary anovas of SLPLAT and IRI of Phase One and the complementary anova for REMEFF in Phase Two are all significant to it is likely that the non-significant results for SLPLAT and IRI in Phase Two and REMEFF in Phase One are due to sampling differences and small sample sizes. Therefore, there is no indication that these measures consistently show that they are equally as variable between animals as within animals, and it is likely that the variability of all sleep measures is greater between animals than the variability observed within animals. Means calculated across the 4 nights for each animal in Phase One and across the 10 nights for each animal in Phase Two were used to examine the relative interindividual variation of the sleep measures. The rationale behind this analysis was that a mean across several nights of sleep for an animal represents a more accurate assessment of that animal's sleep pattern than


34 Table s. Interindividual and Intraindividual Variability Profiles for Phase One and Phase Two. Inter individual (%) Intra individual (%) Sleep Measure Phase One Phase Two Phase One Phase Two SLPLAT 82 27 56 68 REMLAT 32 20 36 35 AROUS 26 33 33 33 AWAKE 30 25 30 35 DROWSY 37 39 15 19 MEANRL 12 13 12 15 STGCHG 10 9 12 12 REMP 10 7 11 10 REM 16 20 10 11 IRI 10 2 9 11 TWO 9 7 9 7 THREE 15 7 8 10 REMEFF 4 5 6 7 SLPEFF 5 3 5 4 TST 8 5 4 3 one night alone. Using these mean values, means, standard deviations and coefficients of variation for each sleep measure were computed across all animals in each phase. A "variability profile" of the 15 sleep measures for each phase was created using the coefficients of variation. The interindividual variability profiles for each phase are listed in Table 5. Homogeneity of the profiles between the phases was tested with Friedman's test as suggested by Sokal and Braumann (1980) and was significant to The sleep measures were ranked from high to low within each phase, then the resulting ranked sets of the two phases were tested for similarity. A significant p-value suggests that the ranks of the variation coefficients of the sleep measures differ


35 consistently over the two phases. Therefore, the relative interindividual variabilities of the sleep measures were similar between the two phases. Although the relative variabilities of the sleep measures within each phase cannot be tested statistically, it can be seen from Table 5 that SLPLAT, REMLAT, AWAKE, AROUS and DROWSY are the most variable in both phases, though the order within in each phase differs slightly. Further, TST, REMEFF, and SLPEFF demonstrate the least variability of all the sleep measures. IRI in Phase Two demonstrates an extremely low variation coefficient. It is thought that this score is spuriously low due to the low sample number of Phase Two. The relative variabilities of the 5 different stages of sleep and wakefulness are similar for both phases: Drowsy was the most variable, followed by Awake, REM, Stage 3-4 and Stage Two. Stage 3-4 and Stage Two had the same variation coefficient in Phase Two. Intraindividual variation. Coefficients of variation for each sleep measure for each animal were computed to examine the relative variability of the sleep measures within all animals in both phases. As illustrated graphically in Figure Sa for Phase One and Figure Bb for Phase Two, the trend of variability profiles of the sleep measures for all animals were consistent with few exceptions. Tests evaluating homogeneity of variability profiles across all animals in


36 each phase using Friedman's test were significant to To obtain one score that was representative of the variation within animals, means of the variation coefficients across all animals in each phase were computed for each sleep measure. The results of this analysis of the relative variability of the sleep measures within animals can be seen in Table 5. SLPLAT, REMLAT, AWAKE, AROUS, and DROWSY are the most variable sleep measures and SLPEFF, REMEFF, and TST are the least variable and were consistent across phases in the ranks of their variability coefficients The homogeneity of the variability profiles for both phases within and between animals shown in Figure 9 was also significantly similar In order to establish whether the relative variabilities between the sleep measures were significant, t-tests were performed between each of the mean variation coefficients for Phase One and Phase Two. SLPLAT was the most significantly variable sleep measure of all measures in both phases REMLAT, AROUS, and AWAKE were similar in variability to each other, but were significantly more variable than all the measures that followed, and DROWSY was less variable than SLPLAT, REMLAT, AROUS,and AWAKE, but was significantly more


110 100 10 11'1 10 .... z 70 w u i: 10 .. w 0 u 60 z 0 40 II: c JO > 10 10 0 Figure Sa. 110 100 10 ao 11'1 .... z 70 w u i;: 10 .. w 0 u 50 z 0 40 c iii c JO > 20 10 0 37 1 J 6 7 I I 10 11 12 13 14 15 SU:EP hiiEASURES Phase One: Range of Intraindividual Variability Profiles of Sleep Measures by Animal. l=SLPLAT, 2=REMLAT, 3=AROUS, 4=AWAKE, S=DROWSY 6=MEANRL, 7=STGCHG, S=REMP, 9=REM, lO=IRI ll=TWO, 12=3-4, 13=REMEFF, 14=SLPEFF, lS=TST. 1 J 4 6 7 I I 10 11 12 13 14 15 SLEEP h!IEASURES Figure Sb. Phase Two: Range of Intraindividual Variability Profiles of Sleep Measures by Animal l=SLPLAT, 2=REMLAT, 3=AROUS, 4=AWAKE, S=DROWSY 6=MEANRL, 7=STGCHG, S=REMP, 9=REM, lO=IRI ll=TWO, 12=3-4, 13=REMEFF, 14=SLPEFF, lS=TST.


38 110 100 10 ., 10 ... z 70 lol C VITHII Pfi.&3E OlE u i: 10 II. lol + 8E'N EEl Pfl.&31 OlE 0 VITHIJI PfiASI 'NO 0 u 50 6 BE'NE Pfl.&31 'NO z 0 j: 40 oC ii oC 30 > 20 10 a 2 3 4 5 I 7 I I 10 11 12 13 14 15 SLEEP I.IE.\SURES Figure 9. Variability Profiles for Both Phases: Within and Between Animals. l=SLPLAT, 2=REMLAT 3=AROUS, 4=AWAKE, 5=DROWSY, 6=MEANRL, 7=STGCHG 8=REMP, 9=REM, lO=IRI, ll=TWO, 12=3-4, 13=REMEFF, 14=SLPEFF, 15=TST. variable than the other sleep measures. REMEFF, SLPEFF, and TST were significantly lowest in variability.1 When the 5 stages of sleep and wakefulness are assessed for both phases, AWAKE was the most variable, followed by DROWSY, and REM. In Phase One, Stage TWO was next followed by Stage which had the least variability of the 5 stages in Phase One. In Phase Two, Stage 3-4 showed more variability than Stage TWO. 1It must be noted that this assessment of intraindividual variability is a result of multiple t-tests and the number of significant values may be spuriously high.


Phase One: Rearing Environment and Sleep Variability The value used to represent the intraindividual variability of the sleep measures in the remaining analysis of Phase One is the coefficient of variation 39 rather than the standard deviation. It was demonstrated by the analysis of variance that the mean values of the sleep measures were significantly different between the animals, therefore, the coefficient of variation was thought to be a more accurate assessment of variability across the animals since the magnitude of the standard deviation tends to be directly proportional to the size of the mean (Lande, 1977). Gender, group, maternal rank and age at run. To determine whether gender, group, or maternal rank had any effect on the variability of sleep patterns, the variances of the individual means and intraindividual variation coefficients of each subgroup were computed and tested against its counterpart subgroup with an F-test. None of the sleep measures revealed any significant differences with respect to gender, group, or maternal rank. To determine whether age at run, the first day that physiology data were collected, had any effect on the variability of the sleep measures, the individual means and the intraindividual variation coefficients for each sleep measure were regressed on the age of the


40 infants in days when physiological data were collected. None of these regressions was significant, thus sleep variability was not a function of age in this study. Behavioral consolidation. The behavioral variables represented in Table 2 were consolidated into 5 composite scores to form five behavioral categories: Mother-Infant (MOTHINF), Activity (ACTIVITY), OtherAnimal (OTHERAN), Aggression/Dominance (AGGDOM), and Distress (DISTRESS). Each category included behavior that best represented either comfort or stress to the infant such as the quality of the mother-infant bond or the aggression observed in the infant's environment. Activity was included because previous literature has shown it to have some effect on sleep patterns, and as a result may affect the variability of the sleep. The Mother-Infant category includes those behaviors that are a function of a close mother-infant bond and represents time spent in ventral-ventral contact with mother. These behaviors were Cradle, Enclose, and Carry Mother, and time spent on the Nipple. Contact Mother, Passive Support Mother, Receive Groom, and Receive Genital Exploration from mother were excluded here because some mothers have been observed to push their squealing infants off within the first month, pin them to the ground and groom them. Therefore, these behaviors were excluded because they might confound the Mother-Infant category in Phase One.


41 The Activity category represents those behaviors where the infant is actively moving about the pen and consists of walking, running, or climbing (represented by Locomotion}, Play, and Activity Count. Initiate Play and Auto-Play were excluded because of infrequent occurrence in Phase One. The Other-Animal category includes behaviors received from other animals that might be comforting to the infant. This consists of time spent in ventral-ventral contact with or leaning against other animals and includes Cradle, Enclose, Carry, and Passive Support. Also included were Contact Lady which represents time spent contacting adult females, and Receive Groom and Receive Genital Exploration from others. The Aggression/Dominance category represents those behaviors that involve hierarchical disputes that affect and may be stressful to the infant. The behaviors forming this category include frequency and duration of fights in the group, Fight(F} and Fight(D}, and any aggression the mother or the infant receives from another animal, including Receive Threat, Receive Aggression, Receive Punishment, and Mother Receives Aggression. Submit was also included in this category. Restrain was excluded from this category in Phase One because, at an early age, it often occurs when the mother is trying to protect her infant. If the infant is active and running around, however, the mother (or another animal} may


42 restrain the infant as a mild form of punishment. It was consequently excluded because it might confound the category. The final category, Distress, represents behaviors manifested by the infant that demonstrate dis-tress as well as behaviors that are thought to cause the infant distress. These behaviors include any crying behavior, Jerk, Squeal, Coo, and Temper Tantrum, incidences where the infant is refused the nipple, Wean and Nipple Deny, and also Oral Self which is included in Phase One because it is often observed to occur when the infant is denied the nipple. In a primary consolidation, those behaviors within each category that were mutually exclusive or were variations of the same general behavior were added together to form composite behaviors. The results from this primary consolidation can be seen in Figure 10. In order to add the remaining behaviors to form the final categories, the results of the primary con-solidation were ranked across animals and the resulting ranks of each behavior in the category were added together to form the final behavioral categories. The main purpose of this secondary consolidation was to add frequency and duration behaviors together so that both could be included in the behavioral categories. The results of the secondary consolidation can be seen in Figure 10.


43 Figure 10. Phase One: Behavioral Consolidation Primary Consolidation 1 Secondary Consolidation 2 Mother-Infant Behavior Cradle MOTHINF ----Nipple--------------------------Activity Locomotion +++++---------------Play ---------------=--ACTIVITY Activity Count-----------------+ Other Animal Enclose Carry Passive Support OTHERAN Contact Lady Receive GroomReceive Exploration-Distress Oral Se 1 f"---------------------+ Squeal Nipple Deny-1+++'s indicate that the designated variables have been added together to form an intermediate value 2The secondary consolidation is represented by the abbreviated variable name for the category.


44 Early behavior and sleep variability. To determine whether early behavior might account for the variation of sleep measures between the infants 4 to 7 months later, multiple stepwise regressions were performed on the individual mean values of the sleep measures where the sleep measure was the dependent variable and the behavioral categories were the independent variables. The computer program used to perform this analysis was BMDP P2R (Dixon 1983). To choose predictor variables, the program used the partial correlations of each independent variable with the dependent variable and the resulting F-tc-enter value. The F-tc-enter value computed for each independent variable was the appropriate F statistic used for testing the significance of the linear regression of the dependent variable on the independent variable. Of the 15 sets of sleep measure means, REMLAT and IRI showed significant regressions with one of the behavioral categories. The best predictor for the interindividual variability of REMLAT was AGGDOM which accounted for 54 percent of the variability. The regression equation, which is significant to 0.05, is REMLAT = 4.31X + 10.7 where X is AGGDOM. When ACTIVITY was added to the equation, the percent of interindividual variability accounted for (the multiple coefficient of determination)


45 rose to 68 percent and was significant to 0.05. When the regression equation included DISTRESS, the variation accounted for was 71 percent and was significant to o.os. Although the addition of these 2 behavioral categories increased the predictive value of the equation, the increases were not significant thus were not included in the equation. The best predictor variable for IRI was OTHERAN and resulted in a negative regression that was significant to o.os. The regression equation with OTHERAN as the predictor variable accounts for 41 percent of the interindividual variability of IRI. The equation is IRI = -0.67X + 68.39 where X is OTHERAN. Addition of other behavioral categories resulted in nonsignificant regressions and did not significantly improve the prediction of IRI. Since the multiple regression tests only whether the dependent variable is predicted by the independent variable, product-moment correlations were performed between the two sleep measures that showed significant regressions and the behavioral categories. Results from these product-moment correlations are shown in Table 6. To determine the behavioral categories that best predicted intraindividual variability of sleep patterns later in life, multiple stepwise regressions, using the BMDP program P2R, were performed on the intraindividual variation coefficients of each sleep measure with the


46 Table 6. Phase One: Product-moment Correlations between Behavioral Categories and Individual Sleep Measure Means. Only those sleep measures that resulted in significant regressions are shown. r=0.602, MOTHINF ACTIVITY OTHERAN AGGDOM DISTRESS SLPLAT -0.203 -0.202 0.392 0.519 REMLAT 0.344 0.732 0.344 REM ... 0.380 -0.182 -0.440 IRI 0.378 -0.310 -0.642 0.116 -0.275 sleep measure as the dependent variable and the behavioral categories as the independent variables. Of the 15 sleep measures, SLPLAT, REMLAT, REM, showed significant regressions with one or more of the behavioral categories. The best predictors for intraindividual variability of SLPLAT were DISTRESS and OTHERAN which together account for 57 percent of the variability. The regression equation, which is significant to 0.05, is SLPLAT = 117.88 2.23X 1.80Y where X is DISTRESS and Y is OTHERAN. Addition of other behavioral categories resulted in nonsignificant regres-sions and did not significantly improve the prediction of SLPLAT. The best predictor variable for the intra-individual variability of REMLAT was DISTRESS and resulted in a negative regression that was significant to 0.01. The regression equation with DISTRESS as the


predictor variable accounts for 61 percent of the variability of REMLAT. The equation is REMLAT = 65.4 -1.64X 47 where X is DISTRESS. When ACTIVITY was added to the equation, the percent of variability accounted for rose to 63 percent. When MOTHINF was included in the regression, 68 percent of the variability of REMLAT was accounted for. Although the addition of these two behavioral variables still resulted in a significant regression the increase of prediction of REMLAT was not significant. Therefore the best equation for prediction of REMLAT contains only DISTRESS. The intraindividual variability of the sleep measure REM was predicted better by the early behavioral categories than all the other sleep variables. The best predictor variable for REM was AGGDOM which accounted for 70 percent of the variability of REM. The equation, significant to 0.001 was REM = 2.58 + 0.43X where X is AGGDOM. When the regression equation included OTHERAN as a predictor variable, the multiple coefficient of determination was 72 percent and the regression was significant to 0.01. However, the increase of the prediction from 70 percent to 72 percent is not significant. Further, the addition of DISTRESS and MOTHINF also resulted in a significant regression and together with AGGDOM and OTHERAN accounted for 82 percent


48 Table 7. Phase One: Product-moment Correlations between Behavioral Categories and Variation Coefficients of Sleep Measures. Only those sleep measures that resulted in a significant multiple regressions are shown. r=0.602, SLPLAT REMLAT REM MOTHINF 0.412 0.444 0.514 ACTIVITY OTHERAN -0.405 -0.506 -0.473 -0.208 -0.036 -0.535 AGGDOM -0.178 0.837 DISTRESS -0.643 -0.782 0.262 of the variability of REM. Although the addition of these 3 behavioral categories increased the predictive value of the equation, the increases were not significant and thus were not included in the equation. To further elucidate these regressions, product-moment correlations were performed between the intra-individual variation coefficients of three sleep measures that showed significant regressions and the behavioral categories. Results from these product-moment cor-relations are shown in Table 7. Phase Two: Daily Events and Sleep Var1ability Behavioral consolidation. As in Phase One, the behavioral variables represented in Table 2 were consolidated to form the 5 behavioral categories: Mother-Infant, Activity, Other-Animal, Aggression/Dominance, and Distress. Again, each category included behavior that best represents either comfort or stress to the infant. Activity was also included. The categories in Phase Two are basically composed of the same behaviors as the


49 categories in Phase One, but since the animals were older in Phase Two, the behavioral repertoire of the infants changed accordingly. In addition to the behaviors used in the Mother-Infant category of Phase One, Contact Mother, Receive Groom, and Receive Genital Exploration were also included in the Mother-Infant category of Phase Two. When the infant is older, much of its time is spent sitting in contact with its mother rather than being enclosed by her. Further, Receive Groom and Receive Genital Exploration from mother were also included in Phase Two because they are positive interactive behaviors between the infant and its mother. At this age, it is highly unusual for the infant to be pushed away by its mother in such a manner as can be seen during the first month of life, so it is doubtful that these behaviors could confound the Mother-Infant category of Phase Two. In addition to the behaviors represented by the Activity category of Phase One, the Activity category of Phase Two also includes Initiate Play and Auto-Play. The Other Animal category in Phase Two is similar to that in Phase One. In addition to the behaviors utilized in the Aggression/Dominance category of Phase One, this category in Phase Two included Restrain. The mother is less likely to protect her infant with a mild restrain at this age,


and usually a restrain is observed as a mild form of punishment. 50 Behaviors in the Distress category were the same in both phases. The behavioral consolidation procedure was similar to that of Phase One. The results of the primary and secondary consolidation of behavior in Phase Two can be seen in Figure 11. Daily behavior and sleep physiology. To examine the association of daily behavior and nightly sleep physiology within animals, multiple regressions were performed on the 10 days of behavior and the 10 nights of sleep for each of the 5 infants where the behavioral categories were considered the independent variables and each sleep measure was the dependent variable. The resulting regression lines for all 5 animals for each sleep measure were then tested for equality with respect to their slopes and intercepts. The purpose of this analysis was to document any predictive function of the daily behavior on nightly sleep physiology across all 5 animals. To determine which behavioral categories would be used as predictor variables in the multiple regressions, product-moment correlations were performed between sleep and behavioral categories for each of the animals. Direction of correlation (+ or -) and significance across


51 Figure 11. Phase Two: Behavioral Consolidation Mother-Infant Behavior Primary 1 Consolidation Secondary 2 Consolidation Enclose +++++ Carry 7 Nipple ------------++-MOTHINF Contact Receive Receive +++++ Exploration-Activity Locomot1on Play Activity Count-----------------+: ACTIVITY Auto-Play-----------------------Initiate Play----------------Other Animal Enclose Passive Support OTHERAN Contact Lady Receive Groom+++++ Receive Genital Exploration/ Flght(D) + Fight(F)-------------+ Mother Receives Aggressio Receive Receive Threat Receive Punishment Submit---------------------------+ Restrain + Distress Squeal Coo +++++--------------DISTRESS Temper Tantrum Nipple Deny-1+++'s indicate that the designated variables have been added together to form an intermediate value. 2The secondary consolidation is represented by the abbreviated variable name for the category.


52 all five animals for each correlated pair of behavior and sleep was examined. Those behavior-sleep pairs that showed consistency across the animals for direction of correlation as well as significance to 0.05 in several of the animals were chosen for predictor variables in the multiple regressions and can be seen in Table a. This manner of selection was adapted from the method used to choose predictor variables in a stepwise regression. As demonstrated in Table 8, behaviors chosen often differed between the sleep measures. The appropriate behavioral categories were then entered in a stepwise manner for each of the sleep measures. The program used to perform this analysis was BMDP PlR (Dixon 1983). The program first analyzed each sleep measure for all the animals together, then each animal separately for each of the independent variables entered (behavioral categories). The program then tested the equality of regression lines between groups (animals). None of the sleep measures showed significant regressions across all animals with any of the behavioral categories, as measured by the analysis of variance of the regressions (Table 8) Significant regressions did occur in the data but appeared to be randomly scattered throughout. When the equality of regression lines were tested across animals for each of the regressions, most showed significant differences between the animals. Those sleep measures that did demonstrate equality of regression lines between


53 Table 8. Phase Two: Multiple Regressions of Behavior and Sleep Variables. The behavioral variables are independent variables and the sleep measure is the dependent variable. Sleep Significance by Animall Group Variable Behavior 8.7 20.3 23.3 29.2 31.1 Significance2 SLPLAT ACTIVITY NS NS NS NS NS NS ACTIVITY OTHERAN NS NS NS NS NS NS REMLAT DISTRESS AGGDOM NS NS NS NS NS 0.05 DISTRESS AGGDOM ACTIVITY NS NS NS NS NS NS AROUS OTHERAN NS 0.05 NS NS NS 0.0001 REMP OTHERAN NS 0.05 NS NS NS 0.05 OTHERAN ACTIVITY NS NS NS NS NS 0.05 AWAKE DISTRESS NS NS NS NS NS 0.01 DISTRESS OTHERAN NS NS NS NS NS 0.05 DROWSY DISTRESS NS o.os NS NS NS 0.00001 DISTRESS MOTHINF NS NS NS NS NS 0.00001 TWO OTHERAN o.os NS NS NS NS 0.00001 OTHERAN DISTRESS NS NS NS NS NS 0.001 THREE DISTRESS NS NS NS NS NS o.os DISTRESS OTHERAN NS NS NS NS NS o.os 1values given represent p-value resulting from the analysis of variance testing the significance of the regression.




55 the animals were SLPLAT and IRI. SLPLAT was regressed with ACTIVITY and OTHERAN, and IRI was regressed with ACTIVITY and AGGDOM. However, since none of the individual regressions were significant (except for one in IRI which may be an erroneous result due the large number of regressions performed in this part of the study), it is probable that these behaviors are not determining factors in the variability of SLPLAT and IRI. In order to test whether daily events may have a small but significant effect on the sleep measures across all 5 animals, t-tests were performed between the distribution of the 5 correlation coefficients of each sleep/behavior combination and the expected distribution which was normally distributed around the mean of zero. Four of the dyadic combinations resulted in t-tests significant to 0.05 or less. These were Stage 3-4 and DISTRESS (t=7.319, REM and ACTIVITY (t=3.028, p<0.05), MEANRL and ACTIVITY (t=6.476, and MEANRL and MOTHINF (t=-3.297, Cross-Species Comparisons Table 9 compares the means of sleep measures found in this study with mean sleep values observed in other primate studies including infant pigtailed macaques (Reite and Short 1980), infant bonnet macaques (Reite and Earle, unpublished manuscript), immature chimpanzees


Table 9. Cross-Species Comparisons: Mean Values. Sleep stages are represented as the percentage of Total Sleep Time (TST). 56 II. PLAT lliLIT IIDUI 1P UIIAIIE IDIDIIT UID U-4 UEI REMIL Ill ILPEfF IUEFF TIT liTICHI P11111 Dne ll II 21 l u.o 3 47.4 ll. 4 u.t u o II 12 117 P11111 lD 21 .. 21 3 11.1 l.l 51.2 30.0 15.4 12 .. 550 liB ,.,, ... lDfenta' 2 77 22 11.0 5.4 4t.l 27 11.5 12 59 541 lonnet lnt ent1 ll 101 l5 7 ll.' 3.3 50 29.0 ... 12 o 17 14 c" ,., .. ,,,. 121 7 .. 51 20.1 l l II .. Caalar 1 to s 14 uo 2 l.l 2.1 45.1 20 30.5 21 .. 5U 34 'Ill tl 1nt1 Sr. art IUIOI eat1 Eerle tal 8 A11pt1lll froe RUII It 11. IU711 0Ah,te froe IIIIIUI et el. IU741 II .. l lwAII .. Ie (McNew et al 1971), and human children years of age (Williams et al 1974)1 To compare the sleep stages across species, the time in minutes of each sleep stage was converted to percentage of total sleep time. The pigtail study done by Reite and Short (1980) used the same methodology as the present study but used a different set of animals. T-tests performed between Phase One and Phase Two and the 1980 study demonstrated that the mean values for all sleep measures were similar between the studies. When compared with the bonnet macaques recorded under the same conditions as the pigtails in this study, t-tests showed that all the sleep measures were similar across the species except the pigtails had significantly longer SLPLAT and less ARCUS than the bonnets. 1sleep data from children 3 to 5 years of age were chosen because it is thought that the rate of development of the monkey infant is 6 times that of humans(Caveness 1962). If that is the case, then the infants in this study have a comparative developmental age of 2.5 to 4 years.


57 Sleep measures of the chimpanzees and the human children showed several differences with the present study. The chimps showed lower REMP, less %3-4 sleep, greater %REM, greater %DROWSY, and greater MEANRL than the pigtails. The human children similarly exhibited lower REMP, less %3-4 sleep, greater %REM, and greater MEANRL than the pigtails and also had fewer ARCUS, spent much less time in %AWAKE, had shorter IRI, greater SLPEFF, fewer STGCHG, and a longer TST. In addition, they had a shorter SLPLAT than the pigtails but were similar to the bonnets for this measure. Table 10 and Figure 12 demonstrate the interindividual variation coefficients for the pigtails of the present study and the bonnets. The relative variation for the sleep measures of Phase One and Phase Two are slightly different than those presented in Table 5 due to the computation of each sleep stage as a percentage of the total sleep time. Generally, the relative variabilities of the sleep measures remain the same across the two macaque species. SLPLAT, %DROWSY, %AWAKE, REMLAT, and ARCUS are the most variable, and TST, SLPEFF, and REMEFF are the least variable measures. Further, a Friedman's test demonstrated that the rank of the variability coefficients differed consistently for the sleep stages across both species


58 Table 10. Variation Coefficients of Present Data and Bonnet Data. Sleep Phase Phase Bonnets1 Measure One Two SLPLAT 82 27 46 REMLAT 32 20 39 AROUS 26 33 31 %AWAKE 32 27 28 %DROWSY 44 39 33 MEANRL 12 13 11 STGCHG 10 9 NA %REH 13 15 20 IRI 10 2 10 %TWO 11 4 10 % 3-4 11 6 14 REMP 10 7 12 REMEFF 4 5 5 SLPEFF 5 3 3 TST 8 5 3 lR e1te and Earle (unpublished manuscript) NA Not Available 110 100 10 10 ., .... a PHASE OlE z w 70 u + PH.LSE niO i: 10 0 10 .. '3 w 0 u 50 z 0 j:: & 30 > 20 10 0 2 l 4 5 I 7 I I 10 11 12 13 14 15 SL.EEP hiEASURES Figure 12. Variability Profiles of Different Species Represented in Table 9.l=SLPLAT, 2=REMLAT 3=AROUS, 4=AWAKE, S=DROWSY, 6=MEANRL, 7=STGCHG B=REMP, 9=REM, lO=IRI, ll=TWO, 12=3-4 13=REMEFF, 14=SLPEFF, lS=TST.


CHAPTER IV DISCUSSION The presence of intraindividual and interindividual variability of the sleep measures observed among the pigtails in this study is consistent with previous findings of sleep variability (Moses et al. 1972; Reite et al. 1976; Clausen et al. 1974). Interspecies variability of the sleep measures across most of the species examined is also consistent with the literature (Balzamo et al. 1977; Allison and Cicchetti 1976; Bert et al. 1970; Bert and Pegram 1969). However, the sleep patterns of the pigtails and the bonnets were very similar. The differences in SLPLAT and AROUS may simply be the result of a differing social structure between these two closely related species (Reite and Earle unpublished manuscript), although other closely related species have not shown similar sleep patterns (Bert et al. 1972; Bert et al. 1970). The animals in these other studies were adolescent or adult and therefore the age of the subjects when data were collected might be a factor in these results as it has been shown that variability of sleep increases with age (Hauri 1972). Further, these animals were chair-


restrained and adaptation to a such a strange and restrictive testing environment may also contribute to interspecies differences. It has been observed that different species of mammals adapt differently to a laboratory testing environment (Allison and Van Twyver Bert et al. Bert 1972). 60 Differences between the macaque sleep and the chimpanzee and human sleep were quite evident. Greater REM time and longer REM periods observed in the higher species, with the chimpanzees demonstrating values between the macaques and the human children, is consistent with the theory that ontogenically, REM time is greater in those animals whose brain demonstrates relative immaturity at birth with respect to that of the adult animal (Balzamo et al. Webb and Cartwright 1978). The longer IRI in the humans and the fewer REMP found in both the higher primates may be related to the hypothesis that those species with increased body weight and decreased basal metabolic rate tend to have an increased IRI and fewer number of REM periods (Allison and Cicchetti Berger Weiss and Rolden 1964). IRI has also been observed to increase in primates with higher phylogenetic position (Balzamo et al. 1977). The human children appeared to sleep better than the pigtails and bonnets. SLPLAT, AROUS, and %AWAKE were significantly less in the humans than in the macaques. Kripke et al. (1968) also found more frequent arousals in rhesus


61 macaques than are observed in young humans. Further, the humans were also observed to sleep longer and more effi-ciently than the macaques, and had fewer stage changes, an indication of subjective sleep disturbance in humans (Karakan et al. 1971). These findings may be a result of increased disturbances in the monkeys' group environ-ments. Bert and Collomb (1966) have hypothesized that smaller primates may need a higher state of arousal for protection purposes than less threatened species like chimpanzees and humans. The children and the chimpanzees spent less time in Stage 3-4 as compared to the macaques, however, which was not expected. The chimpanzee has been observed to show larger amounts of deep sleep due to its safe sleeping environment and lack of predators (Pegram van Lawick-Goodall 1968). Humans also have safe sleeping environments and lack predators and were observed to have fewer interruptions during sleep so it is unclear why the amount of time in deep sleep was decreased in these two species as compared to the macaque species. Although differences are observed in the mean values, the relative variability of the sleep measures, assessed in this study by the variability profile, is remarkably consistent within and between animals of the same species as well as across the two macaque species examined. The sleep measures tended to fall into 3 groups of variability: SLPLAT, REMLAT, AROUS, AWAKE and DROWSY


62 being the most variable; REMP, IRI, MEANRL, STGCHG, and time spent in REM, Stages TWO, and 3-4 being of middle variability; and TST, REMEFF, and SLPEFF being the least variable. Further, preliminary examination of McNew's chimpanzees and Williams' human children suggests that this consistency in relative variation of the sleep measures may be demonstrated by these two higher species of primates as welll. This recurring trend suggests that although the sleep measures may differ slightly in the mean values of their sleep patterns, the inherent variability of each measure is relatively constant across species and may in fact be subject to the same factors either intrinsically or environmentally. Although age and gender did not appear to be factors in the pigtails of this study or in Clausen's study on humans (1974), it is unclear whether this consistency of the variability profiles of sleep patterns would be observed across lifetimes for the two sexes. Since variability of sleep measures has been observed to increase with age in humans and frequently depends on the sex of the subjects (Williams et al. 1974), the relative variability profile may be affected over the longterm. For example, deep 1The standard deviations and therefore the coefficients of variations for these two species could only be estimated from the data presented in the papers of McNew and Williams et al. (1971 and 1974, respectively). Since the coefficient of variation is very sensitive to changes in the standard deviation, particularly when a small mean value is involved, the analysis is purely suggestive with these two higher species.


63 sleep and REM decrease as the organism ages and perhaps the factors relating to these changes in the sleep pattern differ between subjects, thus increasing the variability. Sleep measures with the most variability, SLPLAT, REMLAT, AROUS, AWAKE, and DROWSY tend to be those that would be most affected by random disturbances in the surrounding environment of the sleeping individual. This trend is apparent both in the nonhuman primates where the subject is sleeping in close proximity to others and in humans where the subject is sleeping alone in the laboratory (Williams et al. 1974: Clausen et al. 1974). Such factors as the presence of other animals, the laboratory itself, and the recording apparatus might contribute to these findings. The measures with moderate variability, MEANRL, STGCHG, REMP, IRI, REM, and Stages TWO and 3-4 may be affected less by the sleeping environment and thus may be less subject to random changes that might occur from night to night. They may, however, be subject to other unknown intrinsic or environmental factors which may help contribute to their variability. The least variable measures, TST, REMEFF, and SLPEFF are measures that are directly tied to the day/night cycle as defined by the experimenter, the sun, and/or intrinsic circadian rhythms. They are thus less likely to change very dramatically from night to night, individual to individual, or


64 from species to species as long as the species are prone to the same sleep/wake cycle. Consistent with previous findings on adult humans (Clausen et al. 1974), the relative variability profile remains intact when the percentages of the sleep/wake stages are considered, with %Awake and %Drowsy as the most variable, and the sleep stages, %3-4, %REM, and %Two, less variable. In humans adults, however, %3-4 is more variable than %REM and %Two. The pigtails and bonnets, however, showed %REM more variable than %Two and %3-4. This difference between lower primates and humans with respect to the variability of Stage 3-4 might be a function of species and/or age differences. It would be interesting for further research to examine more closely the consistency of relative variability in sleep patterns over phylogenetically diverse species with a special consideration to age as well as gender. That the variability was greater across animals than within animals is consistent with Clausen's finding in adult humans (1974). Further, the homogeneity of the variability profiles were relatively similar intra-individually for both phases, but some discrepancies across the phases were observed interindividually. These results suggest that intraindividual variation of sleep in infant monkeys can accurately be assessed by a rela-tively small sample size (four nights appeared to be enough in this study) whereas sample sizes need to be


65 somewhat larger to accurately assess interindividual variation. (It appears from this study that more than five individuals are needed, though the 11 animals of Phase One appeared to be enough at least to show a similar interindividual variability profile to those of other species.) The results of Phase One suggest that both intraindividual as well as interindividual variability of sleep patterns may partly be a function of early rearing environment and behavior. Intraindividually, variability of 3 of the sleep measures, SLPLAT, REMLAT, and REM, and interindividually, variability of two of the sleep measures, REMLAT and IRI were shown to be predicted by behavior observed during the first month of life. DISTRESS and OTHERAN accounted for 57% of the intraindividual variability of SLPLAT. The negative correlation between these variables (Table 7) suggests that the animals which show greater distress and spend more time with other animals than their mothers tend to be less flexible with respect to their sleep latency than those who are not as distressed and stay closer to mother. Although not statistically significant, a correlation between the actual length of SLPLAT and the behavioral variables suggest that these less flexible animals also tended to have longer sleep latencies (Table 6). Perhaps these animals are less variable in their sleep latency because they rarely go to sleep


quickly and take some time after lights out to settle down, a pattern that may have been developed early in life. 66 AGGDOM accounted for 70% of the intraindividual variability of time spent in REM. The positive correlation between AGGDOM and the variation coefficient of REM (Table 7) suggests that those animals that are submissive and/or are the subject of frequent hostility from other animals have greater variability in their REM time than those who do not receive aggression as frequently. A negative correlation between the actual time spent in REM and AGGDOM (Table 6), though not statistically significant suggests that time spent in REM tends to be less in those animals that are the subject of aggression and dominance behaviors when young. REM does tend to be less in those animals that are more likely to be subject to predation (Allison and Cicchetti 1976) and perhaps this may be a factor within species as well. Further, REM in these animals may fluctuate more widely due to their safety factor that they have developed during a particular night, thus explaining the increased variability. OTHERAN accounted for 41% of the intervariability of IRI. The negative correlation between the length of IRI and OTHERAN suggests that those animals that spend more time with animals other than mother tend to have shorter Inter-REM-Intervals than those that do not.


67 Variability of REMLAT was predicted by early behavior and environment both intraindividually and interindividually. DISTRESS accounted for 61% of the intraindividual variability observed for REMLAT. The negative correlation between these variables (Table 7) suggests that more distressed animals in the first month of life show less variability in their REM latencies than those who are not. There was no relationship between the actual length of REMLAT and DISTRESS (Table 6). AGGDOM accounted for 54% of the interindividual variability of REMLAT. The positive correlation between the actual length of REMLAT and AGGDOM (Table 6) suggests that those animals that experience more aggression early in life tend to have longer REM Latencies. REM Latency is of particular interest as it has been discussed as a biological marker for depressive disease in humans (Kupfer 1976; Kupfer and Foster 1972). This measure has also shown large variability in depressives (Ansseau et al. 1985) as has been observed in normal subjects. Since the variability of REM Latency may in part be a function of an individual's behavior and early rearing environment, perhaps controlling for these factors in future studies may aid in reducing the large variability of REM Latency and aid in its standardization as a biological marker in depression. Sleep patterns have been observed to differ across individuals with different temperaments


68 (Weissbluth 1981: DeKoninck et al. 1983: Hartmann 1973a: Monroe 1967). It is likely that temperament is intimately involved in the predictive value of early behavior and environment to the variability of the sleep measures. Temperament may be an influential factor in determining early behavior and the assimilation of and adaptation to the environment. Therefore, sleep variability may be directly related to an individual's temperament. Further research is needed to assess temperament as an independent variable and its relationship to behavior as well as the variability of sleep measures. Day to day variations in behavior and environment in this study were not observed to show a consistent relationship with nightly sleep physiology across the five animals of Phase Two. Similarly, Reite et al. (1976) in a study examining socially-deprived surrogate-reared pigtail infants, observed that night to night variability persisted in a relatively unchanging environment. They concluded that variability in sleep may be free of environmental determinants. Animals in totally monotonous environments however, such as the third stage of chair-restraint adaptation in baboons described by Bert (1972), show that sleep under these conditions varies little from night to night. Although extremes in daily behavior and environment have been observed to affect sleep patterns, normal daily behavior and environment may not be so important.


69 A few behavior-sleep trends were observed when the distributions of the correlation coefficients were examined in Phase Two. Positive associations were observed between Stage 3-4 and DISTRESS, and ACTIVITY and REM and MEANRL, and a negative association was observed between MEANRL and MOTHINF. Although increased NREM sleep in neonates has been observed immediately following circumcision (Emde et al. 1971), slow-wave-sleep has generally been associated with increased activity in previous studies (Baekeland and Lasky 1966: Zloty et al. 1973: Shapiro et al. 1975). However, this is not always conclusive (Adamson et al. 1974: Webb and Agnew et al. 1974). Further, REM has generally been associated with emotional stress rather than physical stress (Hauri 1982). Although the correlations between MEANRL and ACTIVITY, and Stage 3-4 and DISTRESS were significant to 0.01, the possibility of making a Type I error as a result of the multiple correlations is greatly increased. Further, discrepencies exist between the current findings and previous findings. Therefore, caution must be exercised when accepting the validity of these results. More controlled studies need to be done examining both physical and emotional stress within the same individuals to clarify these findings. There are many other factors that may be affecting the night to night variation observed in sleep patterns that were not considered here. Individual


70 animals did show some relationships between behavior and sleep (Table 8). It is possible that some animals are more easily influenced by certain daily events than others, depending on personality factors, genetics, or previous experience such as that discussed in Phase One. These factors are difficult to quantify and study, but in light of the complexity of these animals, it is very possible that they may have some influence. There are other more tangible factors that were not considered here that may have influence on the variability of sleep patterns in these monkeys. The animals were not monitored by laboratory personnel for the three hours prior to lights out time and this was frequently a time for severe group fighting (personal observation). It has been shown that pre-sleep stress, both emotional and physical, have an effect on the following night's sleep patterns (Haynes et al. 1981: Browman and Tepas 1976). Also, the events occurring throughout the night within the group or within the primate laboratory were also not taken into consideration. It is likely that nightly group activities within the study pen or proximate pens might disturb the study animal and may account for some of the arousals, and consequently the depth of sleep achieved by the individual. These factors should be considered in future studies investigating sleep variability.


71 Clearly, variability of sleep is a complex phenomenon. Sleep of the higher primates is basically similar in electroencephalographic patterns and stage cycling (Balzamo et al. 1977). The differences in the absolute values of the sleep measures are likely to be due to phylogenetic adaptations to the individual niches. Despite these absolute differences, relative variability of the sleep measures is remarkably consistent across species. Even on the individual level, the relative variability of the sleep measures is consistent. Whether this is due to genetic homeogeneity of these closely related species and/or similar environmental pressures is unknown. That variability may be partly acquired during the early environment of the individual, may be a result of the individual's developmental adaptation to its specified niche, suggests that it may contribute to the idiosyncracies of its sleep pattern. This may help explain why variability appears to increase with age. That is, idiosyncracies of sleep may increase with age and experience thus increasing interindividual sleep variability. Although nightly sleep variation does not appear to be solely a function of daily events, these results may be deceiving. These factors may in fact be very subtle, may be additive in their effects, and further,


72 may be individually assimilated, thus may be a a product of genetics, learning, and environment. Alternatively, the normal night variation of sleep may be primarily a baseline, a manifestation of the inherent capability of sleep to adapt, for example when the animal is subject to extreme conditions. An animal that can adapt quickly and effectively in order to maintain maximum vigilance yet satisfy the need for sleep under extreme conditions is an animal that is more likely to survive.


REFERENCES 1. Adamson, L., Hunter, W.M., Ogunremi, 0.0., Oswald, I., Percy-Robb, I.W. (1974). Growth hormone increase during sleep after daytime exercise. J. Endocrinol. 62:473. 2. Adams-Tucker, c. (1982). Proximate effects of sexual abuse in chidhood: A report on 28 children. Am. J. Psychiat. 139(10):1252-1256. 3. Agnew, H.W., Webb, W.B., Williams, R.L. (1967). Comparison of stage four and 1-REM sleep deprivation. Percept. Mot. Skills 24:851-858. 4. Allison, T., Cicchetti, o.v. (1976). Sleep in mammals: Ecological and constitutional correlates. Science 194:732-734. 5. Allison, T., Goff, W.R. (1968). Sleep in a primitive animal, the spiny anteater. Psychophysiology 5:200. 6. Allison, T., van Twyver, H. (1970). The evolution of sleep. Nat. Hist. 79:56-65. 7. Anders, T.F., Weinstein, P. (1957). Sleep and its disorders in infants and children: A review. Pediatrics 50(2):312-324. 8. Angus, R.G., Pearce, D.G., Buguet, A.G.G., Olsen, L. (1979). Vigilance performance of men sleeping under arctic conditions. Aviat. Space Environ. Med. 50(7):692-696. 9. Ansseau, M., Kupfer, D.J., Reynolds, C.F. III. (1985). Internight variability of REM latency in major depression: Implications for the use of REM latency as a biological correlate. Biol. Psychiat. 20:489-505. 10. Aserinsky, E., Kleitman, N. (1955). Two types of ocular motility during sleep. J. Applied Physiol. 8(1):1-10. 11. Baekeland, F., Lasky, R. (1966). Exercise and sleep patterns in college atheletes. Percept. Motor Skills 23:1203-1207.


12. Balzamo, E., Bradley, R.J., Bradley, D.M., Pegram, G.V., Rhodes, J.M. (1972a). Sleep ontogeny in the chimpanzee: From birth to two months. Electroenceph. Clin. Neurophys. 33:41-46. 13. Balzamo, E., Bradley, R.J., Rhodes, J.M. (1972b). Sleep ontogeny in the chimpanzee: From two months to forty-one months. Electroenceph. Clin. Neurophys. 33:47-60. 14. Balzamo, E., Santucci, v., Seri, B., VuillonCacciuttolo, G., Bert, J. (1977). Nonhuman primates: Laboratory animals of choice for neurophysiologic studies of sleep. Lab. Anim. Science 27(5):879-886. 74 15. Berger, R.J. (1969). The sleep and dream cycle. In Sleep: Physiology and Pathology, ed. A. Kales. J.B. L1pp1ncott Co., Phlladelphia, pp. 17-32. 16. Berger, R.J. (1975). Bioenergetic functions of sleep and activity rhythms and their possible relevance to aging. Fed. Proc. 34:97-102. 17. Bert, J. (1972). Adaptation of sleep to recording conditions in a baboon. In The Sleeping Brain, ed. M. H. Chase. Perspectives in the Bra1n Sc1ences Vol. I, UCLA, Ca., pp. 28-33. 18. Bert, J., Balzamo, E., Chase, M., Pegram, v. (1975). The sleep of the baboon, Papio papio, under natural conditions and in the laboratory. Electroenceph. Clin. Neurophys. 19. Bert, J., Collomb, H. (1966). L'electroencephalogramme du sommeil nocturne chez le babouin. Etude par telemetrie. J. Physiol. Paris 58:285301. 20. Bert, J., Collomb, H., Martino, A. (1967). L'electroencephalogramme de sommeil d'un prosimien. Sa place dans !'organization du sommeil chez les primates. Electroenceph. Clin. Neurophys. 23:342-350. 21. Bert, J., Pegram, v. (1969). L'electroencephalogramme du sommeil chez les Cercopithecinae: Erythrocebus patas et Cercopithecus aethoips sabaeus. Folia Primat. 11:151-159. 22. Bert, J., Pegram, v. (1972). Sleep and environment. In Medical Primatology 1972. Proc. 3rd Con. Exp. Med. and Surg. Primates, Part II, ed. J. Moor-Janowski. Karger, Basel, pp. 244-249.


23. Bert, J., Pegram, v., Balzamo, E. (1972). Comparison du sommeil de deux macaques (Macaca radiata et Macaca mulatta). Folia Primat. 17:202-208. 24. Bert J., Pegram, v., Rhodes, J.M., Balzamo, E. Naquet, R. (1970). A comparative sleep study of two cercopithecinae. Electroenceph. Clin. Neurophys. 28:32-40. 25. Boelkins, R.C. (1967). Determination of dominance hierarchies in monkeys. Psychon. Science 7:317-318. 26. Browman, C.P., Cartwright, R.D. (1980).The first night effect on sleep and dreams. Biol. Psychiat. 15(5):809-812. 27. Browman, C.P., Tepas, D.I. (1976). The effects of presleep activity on all night sleep. Psychophysiology 13(6):536-540. 28. Bunnell, D.E., Bevier, w., Horvath, S.M. (1983). 75 Effects of exhaustive exercise on the sleep of men and women. Psychophysiology 20(1):50-58. 29. Caveness, W.F. (1962). Atlas of Electroencephalography in the Develop1ng Monkey: Macaca mulatta. Addison-Wesley Publishing Co., Boston. 30. Clausen, J., Sersen, E.A., Lidsky, A. (1974). Variability of sleep measures in normal subjects. Psychophysiology 11:509-516. 31. DeKoninck,J., Gagnon, P., Lallier, s. (1983). Sleep positions in the young adult and their relationship with the subjective quality of sleep. Sleep 6(1):52-59. 32. Dement, w.c., Cohen, H., Ferguson, J., Zarcone, v. (1970). A sleep researcher's odyssey: The function and clinical significance of REM sleep. In The Psychodynamis Implications of the PsychologiCal Studies on Dreams, eds. L. Madow and L.H. Snow. Thomas, Springfield. pp. 71-123. 33. Dement, w.c., Kleitman, N. (1957a). The relation of eye movements during sleep to dream activity: An objective method for the study of dreaming. J. Exp. Psycho!. 53:339-346.


34. Dement, w.c., Kleitman, N. (1957b). Cyclic variations in EEG during sleep and their relation to eye movements. Electroenceph. Clin. Neurophys. 9:673-690. 35. de Monchy, c., Kappers, M.J. (1980). A study of the causes of sleep disorders in young children. Ned. Tijdschr Geneeskd. 124(8):264-269. 36. Dixon, W.J.,ed.(l983). BMDP Statistical Software, 1983 Revised Printing. University of California Press, Berkeley, Cal1f. 37. Emde, R.N., Harmon, R.J., Metcalf, D., Koenig, K.L., Wagonfeld, s. (1971). Stress and neonatal sleep. Psychosom. Med. 33(6):491-497. 76 38. Globus, G., Freidmann, J., Cohen H. (1974). The effects of aircraft noise on sleep electrophysiology as recorded in the home. In Proceedings of the International Congress on Noise as a Publ1c Health Problem, ed. W.D. Ward. Environmental Protect1on Agency, Wash DC., pp. 587-591. 39. Hartmann, E. (1969). Mania, depression and sleep. In Sleep: Physiology and Pathology, ed. A. Kales. J.B. L1pp1ncott Co., Ph1ladelph1a, pp. 183-191. 40. Hartmann, E. (1973a). Sleep requirement: Long sleepers, short sleepers, variable sleepers and insomniacs. Psychosomatics 14:95-103. 41. Hartmann, E. (1973b). The Functions of Sleep. Yale Univ. Press, New Haven, Ct. 42. Hauri, P. (1982). The Sleep Disorders. Current Concepts. The Upjohn Co., Kalamazoo, Mi. 43. Haynes, S.N., Adams, A., Franzen, M. (1981). The effects of pre-sleep stress on sleep onset insomnia. J. Abn. Psychol. 90(6):601-606. 44. Healey, E.S., Kales, A., Monroe, L.J., Bixler, E.o., Chamberlin, K., Soldatos, C.R. (1981). Onset of insomnia, role of life-stress events. Psychosom. Med. 43(5):439-451. 45. Hirschberg, J. (1957). Parental anxieties accompanying sleep disturbances in young children. Bull. Meninger Clin. 21:129.


46. Horne, J.A. (1981). The effects of exercise upon sleep: A critical review. Biol. Psycho!. 12(4):241-290. 47. Kaufman, I.C., Rosenblum, L. (1966). A behavioral taxonomy for Macaca nemestrina and Macaca radiata: 77 Based on longitudinal observation of family groups in the laboratory. Primates 7:206. 48. Karakan, I., Williams, R.L., Salis, P.J., Hursch, C.J. (1971). New approaches to the evaluation and treatment of insomnia (preliminary results). Psychosomatics 12:81-88. 49. Kripke, D.F., Reite, M.L.,Pegram, G.V., Stephens, L.M., Lewis, O.F. (1968). Nocturnal sleep in rhesus monkeys. Electroenceph. Clin. Neurophys. 24:582-586. SO. Kupfer, D.J. (1976). REM latency, a psychobiological marker for primary depressive disease. Biol. Psychiat. 11:159-174. 51. Kupfer, D.J., Foster, F.G. (1972). Interval between onset of sleep and rapid eye movement sleep as an indicator of depression. Lancet ii:684-686. 52. Lande, R. (1977). On comparing coefficients of variation. Syst. Zool. 26:214-217. 53. Levere, T.E., Bartus, R.T., Hart, F.D. (1972). Electroencephalographic and behavioral effects of nocturnally occurring jet aircraft sounds. Aerospace Med. 43:384-389. 54. Lubin, A., Moses, J.M., Johnson, L.C., Naitoh, P. (1974). The recuperative effects of REM sleep and stage 4 sleep on human performance after complete sleep loss: Experiment 1. Psychophysiology 11:133-146. 55. McNew, J.J., Howe, R.C., Adey, W.R. (1971). The sleep cycle and subcortical-cortical EEG relations in the unrestrained chimpanzee. Electroenceph. Clin. Neurophys. 30:489-503. 56. Moldofsky, H., Scarisbrick, P. (1976). Induction of neurasthenic musculoskeletal pain syndrome by selective sleep deprivation. Psychosom. Med. 38:35-44.


57. Monroe, L.J. (1967). Psychological and physiological differences between good and poor sleepers. J. Abnorm. Soc. Psycho!. 72:255-264. 58. Monroe, L.J. (1969). Transient changes in EEG patterns of married good sleepers: The effects of altered sleeping arrangement. Psychophysiology 6(3):255-264. 59. Montgomery I., Trinder, J., Paxton, S.J. (1982). Exercise expenditure and TST: Effect of exercise. Sleep 5(2):159-168. 60. Moses, J.A., Lubin, P., Naitoh, P., Johnson, L.C. (1972). Reliability of sleep measures. Psychophysiology 9:78-82. 61. Moruzzi, G. (1972). The sleep-waking cycle. In Reviews of Physiology 64, ed. R.H. Adrian. Springer-Verlag, Berlin, pp. 2-165. 62. Oswold, I. (1970). Sleep, the great restorer. New Sci. 46:170-172. 63. Otto, E. (1973). Physiological analysis of sleep disturbances induced by noise and increased room temperature. In Sleep, eds. W.P. Koella and P. Levin. First European Congress on Sleep Research. Basel, s. Karger. pp. 414-418. 64. Parmeggiani, P.L., Rabini, c. (1970). Sleep and environmental temperature. Arch. Ital. Biol. 108:369-387. 65. Pauley, J.D., Reite, M. (1977). Automatic antenna selector for use in biotelemetry applications. Physiol. Behav. 18:169wl70. 66. Pauley, J.D., Reite, M. (1981). A microminiature hybrid multichannel implantable biotelemetry system. Biotelemetry Patient Monitg. 8:163-172. 78 67. Pearlman, c., Becker, M. (1974). REM sleep deprivation impairs bar-press acquisition in rats. Physiol. Behav. 13:813-817. 68. Reite, M. (1977). Maternal separation in monkey infants: A model of depression. In Animal Models in Psychiatry and Neurology, eds. I. Hanin and E. Usdin. Pergamon Press, New York, pp. 127-140. 69. Reite, M., Earle, H. (unpublished manuscript). Nocturnal sleep patterns in social group living bonnet (Macaca radiata) monkey infants.


70. Reite, M., Pauley, J.D., Kaufman, I.e., Stynes, A.J., Marker, v. (1974). Normal physiological patterns and physiological-behavioral correlations in unrestrained monkey infants. Physiol. Behav. 12:1021-1033. 71. Reite, M., Pauley, J.D. (1976). Multichannel implantable biotelemetry: Problems, pitfalls, and rewards. In Biotelemetry III, eds. T.B. Fryer, H.A. Miller and H. Sandler.-Academic Press, New York, pp. 59-62. 72. Reite, M., Short, R. (1977). Nocturnal sleep in isolation-reared monkeys: Evidence for environmental independence. Devel. Psychobiol. 10(6):555-561. 73. Reite, M., Short, R. (1978). Nocturnal sleep in infant monkeys. Arch. Gen. Psychiat. 35:1247-1253. 74. Reite, M., Short, R. (1980). A biobehavioral developmental profile (BDP) for the pigtailed monkey. Devel. Psychobiol. 13(3):243. 75. Reite, M., Stynes, A.J., Vaughn, L., Pauley, J.D., Short, R. (1976). Sleep in infant monkeys: Normal values and behavioral correlates. Physiol. Behav. 16:245-251. 76. Reite, M., Walker, s., Pauley, J.D. (1973). Implantation surgery in infant monkeys. Lab. Primate News 4:1-6. 77. Reynolds, C.F., Coble, P.A., Kupfer, D.J., Shaw, D.H. (1982). Depressive patients and the sleep laboratory. In Sleeping and Waking Disorders: Indications and Techn1ques, ed. c. Guilleminault. Menlo Park, Ca. pp. 245-264. 78. Rosa, R., Bonnet, M., Kramer, M. (1980). The relationship of sleep and anxiety in anxious subjects. In Sleep Research Vol. 9, eds. M.H. Chase, D.F. Kripke, and P.L. Walter. Brain Information Service/Brain Research Institute, Los Angeles, pp. 177. 79 79. Roth, T., Kramer, M., Trinder, J. (1972). The effect of noise during sleep on the sleep patterns of different age groups. Can. Psychiat. Assoc. J. 17:197-201. 80. Schmidt, H.S., Kaelbling, R. (1971). The differential laboratory adaptation of sleep parameters. Biol. Psychiat. 3:33-45.


81. Selye, H. (1956). The Stress of Life. McGrawHill, New York. 82. Shapiro, C.M., Griese!, R.D., Bartel, P.R., Jooste, P.L. (1975). Sleep patterns after graded exercise. J. Appl. Physiol. 39:187-190. 83. Simpson, G.G. (1944). Tempo and Mode in Columbia University Press, New York. 84. Sokal,R.R., Braumann, C.A. (1980). Significance tests for coefficients of variation and variability profiles. Syst. Zool. 29:50-66. 85. Sokal, R.R., Rohlf, F.J. (1981). Biometry. W.H. Freeman and Co., San Francisco. 80 86. Tilley, A.J., Empson, J.A.C. (1978). REM sleep and memory consolidation. Biol. Psycho!. 6:293-300. 87. Torsvall, L.(l981). Sleep after exercise: A literature review. J. Sports Med. 21:218-225. 88. Trinder, J., Stevenson, J., Paxton, S.J., Montgomery, I. (1982). Physical fitness, exercise, and REM sleep cycle length. Psychophysiology 19(1):89-93. 89. van Lawick-Goodall, J. (1968). The behavior of free-living chimpanzees in the Gombe Stream Reserve. Anim. Beh. Monogr. 1:165-311. 90. Webb, W.B., Agnew, H.W., Jr. (1974). Sleep and waking in a time-free environment. Aerospace Med. 45:617-622. 91. Webb, W.B., Cartwright, R.D. (1978). Sleep and dreams. Ann. Rev. Psycho!. 29:223-252. 92. Webb, w.B., Campbell, s.s. (1979). The first night effect revisited with age as a variable. Waking and Sleeping 3:319-324. 93. Weiss, T., Rolden, E. (1964). Comparative study of sleep cycles in rodents. Experientia 20:280-281. 94. Weissbluth, M. (1981). Sleep duration and infant temperament. Pediatrics 99(5):817-819. 95. Williams, H.L., Hammack, J.R., Daly, R.L., Dement, w., Lubin, A. (1964). Responses to auditory stimulation, sleep loss and the EEG stages of sleep. Electroenceph. Clin. Neurophys. 16:269-279.


96. Williams, H.L., Holloway, F.A., Griffiths. W.J. (1973). Physiological Psychology: Sleep. Am. Rev. Psychol. 24:279-316. 97. Williams, R.L., Karakan, I., Thornby, J.I., Salis, P.J. (1972). The electroencephalogram sleep patterns of middle-aged males. J. Nerv. Ment. Dis. 154:22-30. 98. Williams, R.L., Karakan, I., Hursch, C.J. (1974). Electroencephalography (EEG) of Human Sleep: Clinical Appl1cat1ons. W1ley, New York. 99. Wilson, A.C., Sarich, V.M. (1969). A molecular time scale for human evolution. Proc. Nat. Acad. Sci. 63:1088-1093. 100. Yablokov, A.V. (1974). Variability of Mammals. Amerind Publishing Co., New Delhi. 101. Zepelin, H., Rechtschaffen, A. (1974). Mammalian sleep, longevity, and energy metabolism. Brain Behav. Evol. 10:425-470. 102. Zloty, R,B., Burdick, J.A., Adamson, J.D. (1973). Sleep of distance runners. Activitas Nervosa Superior Praha 15:217-221. 81