Mammalian Sleep

W0797-008 1/6/05 5:40 PM Page 91
8
Mammalian Sleep
Harold Zepelin
Jerome M. Siegel
Irene Tobler
ABSTRACT
The belief that some mammals do not sleep (e.g., prey species,
because of a need for constant vigilance; shrews, because of
Knowledge about sleep comes primarily from research on
the need for incessant foraging) has been superseded by sys-
mammalian species, whose daily sleep quotas range from 4 to
tematic observations. Some species, in some circumstances,
19 hours, with rapid eye movement (REM) sleep occupying
may be able to postpone sleep for long periods, or sleep may
10% to 50% of this time. Findings of REM sleep or elements of
simply be difficult to recognize, as in the ever-swimming,
it in monotremes have filled a gap in its evolutionary history.
blind Indus dolphin, whose sleep occurs in periods measured
To some, this suggests that REM was inherited from reptiles,
in seconds as it contends with strong river currents.3
although the absence of REM in living reptiles casts doubt on
this view.
The function of sleep remains controversial.1 On one hand,
SLEEP CRITERIA
restorative theories hold that brain processes during sleep
Sleep can usually be identified by sustained quiescence in a
sustain waking behavior (e.g., visual function, learning). On
species-specific posture accompanied by reduced responsive-
the other hand, the negative correlation of sleep quotas
ness to external stimuli, but a definition of mammalian sleep
with body size across species suggests that sleep is a state of
requires several additional criteria, such as quick reversibility
enforced rest most urgent in species with low energy reserves.
to the wakeful condition and characteristic changes in the
Because most of the variance in sleep quotas remains unac-
electroencephalogram (EEG). Quick reversibility distinguishes
counted for statistically, supplementary theories are in order.
sleep from coma and hypothermic states (e.g., hibernation).
There are strikingly strong correlations of REM sleep quotas
With only minor exceptions, EEG changes reliably confirm
with degree of maturity at birth—that is, altricial species, born
sleep-related change in behavior and brain activity. Another
with a low percentage of adult brain weight after a short gesta-
fundamental property of sleep, derived from comparative
tion period, have higher REM sleep quotas, whereas precocial
studies in many species, is its homeostatic regulation. See
species have lower quotas. Given other fetal characteristics of
Chapter 7.
altricial species (e.g., lapse of thermoregulation), REM sleep
These definitional criteria exhibit notable interspecies
may be a carryover from fetal life.
variation. Quiescence does not necessarily mean immobility;
for example, some cetaceans reportedly swim while sleeping.4
In terrestrial mammals, lateral and sternoabdominal recum-
Most studies on sleep have been performed in mammals.
bency with eyes closed are the postures most commonly asso-
Human beings, cats, rats, and, more recently, many mouse
ciated with sleep, but there are striking variations (Figs. 8–1
strains have been the most frequent subjects of sleep research,
and 8–2). The horse, elephant, and giraffe, for example, sleep
but about 100 other species have also been studied. There
some while standing. Some species (e.g., cattle) sleep while
are at least two published reports about the daily sleep of
they are ruminating, and many mammals can sleep with eyes
mammalian species for each one pertaining to other classes.2
semiopen. Choice of sleeping site is another element of
This not only makes for mammal-centeredness in thinking
species-specific sleep behavior and varies with mode of life
about sleep but also affords the opportunity for extensive
and social organization. Burrows, caves, and trees are com-
interspecies comparisons that can shed light on the purpose
mon sites because of the safety they afford, but some species
of sleep, which is still without adequate explanation. This
(e.g., the zebra) sleep in the open and seem to rely on the
chapter considers relevant theories in the light of available
presence of vigilant conspecifics for protection.5-8 Ritualistic
findings.
presleep activity is characteristic of some species, ranging
Despite their relative abundance, the mammalian data
from the circling of a chosen spot (seen in dogs and foxes) to
represent less than 3% of roughly 4260 extant species.
the construction of a nest each evening by chimpanzees.
Figure 8–1. Sea otter sleeping “moored”
to a float of algae. (From Bourliere F: The
Natural History of Mammals, 3rd ed. New
York, Alfred A Knopf, 1967, p 68.)
91

---

W0797-008 1/6/05 5:40 PM Page 92
92
Normal Sleep and Its Variations
11 Hz in the opossum, 10 to 16 Hz in rats, 10 to 13 Hz in mice,
and 12 to 16 Hz in primates). Slow wave activity (SWA; 0.5 to
4 Hz) differs in its peak frequency depending on species—that
is, it is more concentrated at lower frequencies in some
species (human beings, rats) than in others (most mice). The
considerable differences in amplitude between species are
difficult to interpret because technical aspects of the record-
ings confound them. Although the changes in the EEG spec-
trum within NREM sleep, especially in SWA, are known to be
continuous, NREM sleep is sometimes subdivided into “light”
and “deep” sleep, on the basis of the amount of delta wave
activity. In most primates, distinct features of the EEG have
led to the definition of two stages in NREM sleep.
Another fundamental characteristic of mammalian sleep is
its homeostatic self-regulation, which keeps the amount and
Figure 8–2. Giraffe in a zoo, presumably in paradoxical sleep.
(From Immelman K, Gebbing H: Schlaf bei Giraffiden. Z Tierpsychol
depth of sleep in equilibrium with prior wakefulness. Sleep
1962;19:84-92.)
loss creates a debt that is repaid, in part, by some lengthening
of subsequent sleep and, in addition, by the intensification of
SWA, as indicated by its increased amplitude and density14
(see Chapters 7 and 33).
Mammalian sleep, as is well known, also includes para-
These preparatory behaviors justify the description of sleep as
doxical sleep (PS), or REM sleep, which is distinguished by
appetitive, instinctive behavior.
desynchronized, low-amplitude EEG activity in association
The timing of daily sleep varies with the species and in each
with eye movements, twitching of the extremities, and pos-
case is complementary to the activity pattern, which may be
tural atonia. Eye movements vary in prominence and may
diurnal, nocturnal, crepuscular, or arrhythmic. Sleep tends to
even be absent, as in the mole. Detectability of atonia also
be concentrated in a single period each day in adult humans
varies. Rhythmic theta activity (4 to 8 Hz) originating in the
and the great apes, although the latter and people in many
hippocampus is a reliable indicator of PS and can be recorded
cultures take a midday nap. In most mammals, however, sleep
in some animals with implanted epidural electrodes over the
is polyphasic, with sleep episodes interrupted by periods of
parietal or occipital cortex. Theta activity is less evident in
wakefulness. Species also vary in the degree of responsiveness
frontal recordings. Many studies in mammals benefit from
to external stimuli during sleep, some awakening more readily
both recording sites by combining frontoparietal electrodes.
than others.9,10
Other striking features of PS are ponto-geniculo-occipital
Sleep onset in mammals is associated with a slowing of
(PGO) spikes, cardiorespiratory irregularity, and largely inhib-
EEG activity, a rising of EEG amplitude, and a decrease in
ited thermoregulatory responsiveness.15 In all species, PS,
muscle activity, followed in most species by the appearance of
defined by several of its behavioral components, is at its height
spindling activity, and in all cases culminating in sustained
in early life, either in the fetus or in the neonate. It is initially
slow activity at relatively high amplitude. Spindling and slow
the predominant state—for example, occupying 90% of the
waves are the hallmarks of mammalian quiet or non–rapid
kitten’s first 10 days of life and perhaps even more time in the
eye movement (NREM) sleep (the term NREM sleep is used
infant rat.16 By virtue of this pattern, PS has been considered
synonymously with slow wave sleep [SWS] in nonhuman
ontogenetically primitive sleep, especially because PS time is
mammals).11,12 The characterization and quantitative analysis
reduced as electrophysiologic signs of quiet sleep and wake-
of sleep spindles in nonhuman mammals and their signifi-
fulness emerge with maturation. However, recent studies
cance has received little attention. It is known that spindles
investigating the development of EEG patterns in newborn rat
are not merely a transition element in the EEG but that they
pups have questioned this notion. The alternative view is that
occur also throughout NREM sleep. In some species, however,
the vigilance state typical for newborns is a manifestation of
the distinction between wakefulness and sleep is not always
the immature nervous system, which becomes progressively
clear-cut. Especially in carnivores, ungulates, and insectivores,
more organized and evolves simultaneously both into the
there are frequent or protracted periods when spindles or slow
typical EEG and behavioral manifestations of PS and into
EEG components appear sporadically against background
NREM sleep.17
behavioral activity that may not be clearly different from that of
The alternation of quiet sleep with periods of REM sleep
wakefulness. This characterizes the state commonly described
constitutes the sleep cycle, also known as the NREM-REM
as drowsiness, also referred to as “light sleep” or “spindle
(or REM-NREM) cycle, which can be considered the basic
sleep.” The transition from wakefulness to sleep depends on the
organizational unit of mammalian sleep (Fig. 8–3). Duration
derivation—in an occipital derivation, it is abrupt, whereas in
of the cycle varies widely from species to species, as do daily
the frontal derivation, it is gradual and can be subdivided into
sleep quotas and the percentage of sleep time occupied by
“light” and “deep” sleep.13 It should be noted that spindles in
REM sleep. The cyclic organization of sleep is a characteristic
mammals are most apparent in frontal regions, but this does
shared by mammals and birds, which also share the behav-
not mean that sleep there is more superficial.
ioral and to some extent the EEG criteria of sleep. Notably,
Spindling activity varies from species to species; it occurs
birds have a smaller difference in the EEG spectrum between
over a wide range of amplitudes and at a variety of frequencies
wakefulness and NREM sleep than mammals. It is still an
(e.g., 11 waves per second in dogs, 6 to 7 Hz in the sloth, 8 to
open question whether the bird brain that lacks the structures

---

W0797-008 1/6/05 5:40 PM Page 93
Mammalian Sleep
93
MOUSE
RAT
SWA
W
N
R
0
2
4
6
8
10
12
14
16
18
20
22
24
Hours
Figure 8–3. Twenty-four-hour distribution of non–rapid eye movement (NREM) sleep (N), REM sleep (R), wakefulness (W), and slow wave activ-
ity (SWA—i.e., “delta” power) in a C57BL/6 mouse and a Sprague-Dawley rat recorded in a 12-hour light, 12-hour dark cycle. Note the polypha-
sic sleep–wakefulness pattern (typical for all nonhuman primates) and the short NREM-REM sleep cycles of 12 to 15 minutes that are typical for
small rodents.
responsible for generating slow waves has the capacity to
However, studies of mammals have localized the neurons
respond to sleep pressure by increasing NREM sleep intensity.
driving and responding to REM sleep processes to the brain-
Because relatively little is known about avian sleep physiology,
stem, and these neuronal groups appear to be present in
other unsuspected differences from mammalian sleep may
reptiles (see Chapter 10). This can be taken advantage of by
exist. On the basis of current knowledge, however, what
recording from these neurons during active and quiescent
chiefly distinguishes avian from mammalian sleep is the much
lower percentage of REM sleep in birds (about 5% of sleep
time, on the average, as opposed to 15% to 30% in mammals),
the occurrence of REM sleep in clusters, the much briefer
REM periods (often less than 10 seconds), and the corre-
Birds
spondingly short sleep cycles. It should be noted, however,
Turtles
that there is an immense diversity of bird species, of which
Marsupials
few have been recorded.
Crocodilians
Monotremes
65
EVOLUTIONARY HISTORY
Snakes and lizards
Eutherian mammals
The essential similarity of sleep in birds and mammals may
Dinosaurs
well be a clue to the history and function of sleep.18 The sleep
cycle first appears in evolutionary history in association with
endothermy, which is the maintenance of a high, constant body
temperature by metabolic means, as found only in birds and
mammals, enabling them to occupy nocturnal niches and survive
in cold climates. The alternation between REM and NREM states
has yet to be explained, but it is possible that the cycle evolved
Millions of years ago
independently (in parallel) in birds and in mammals, or in
their immediate forebears (mammal-like reptiles) (Fig. 8–4).
Research on living reptiles has not produced convincing
Mammal-like
evidence of REM sleep or of sleep organization that might
248
reptiles
suggest a cycle. Although some studies of reptiles have claimed
to see evidence for eye movements and EEG changes during
quiescent periods, it remains unclear whether these phenomena
were simply transient awakenings or REM sleep.19 Brain struc-
ture and sensorimotor organization in reptiles differ greatly
Stem reptiles
from those in birds and mammals, so it is difficult to draw
320
conclusions about reptilian sleep based on simply observing the
Figure 8–4. Temporal relationships and lines of descent for birds,
electrical activity from screw electrodes positioned on the skull,
mammals, and reptiles. Solid lines indicate availability of fossil record;
as can be done in mammals.
dotted lines indicate still-uncertain relationships.

---

W0797-008 1/6/05 5:40 PM Page 94
94
Normal Sleep and Its Variations
states in reptiles. In the first study using this approach, the
of overt sexual behavior in the bottle-nosed dolphin.32 However,
neuronal activity of the midbrain and pontine regions respon-
as myoclonus is seen in terrestrial mammals in NREM sleep32
sible for REM sleep generation was studied in the turtle.20
and nocturnal erections are not always linked to REM sleep in
This study found no evidence of cyclicity in neuronal activity
terrestrial mammals,33 it is unclear whether these observations
during extended quiescent periods and hence no evidence of
indicate REM sleep in cetaceans. Unequivocal REM sleep, as
REM sleep. It would be extremely valuable to confirm these
identified by EEG and electromyographic (EMG) recording and
results in other reptilian species. Pending such evidence, these
evidence of elevated arousal thresholds, has not been demon-
data suggest that REM sleep may not have existed in reptilian
strated in any cetacean. Even if the twitches that have been
species but may have evolved rapidly with endothermy.
observed are eventually demonstrated to be signs of REM
A striking finding of the study in turtles was that most
sleep, a mystery remains. The maximal number of twitches
brainstem neuronal activity in the observed portions of the
seen in cetaceans during relatively quiescent periods is of the
midbrain and pontine reticular formation decreased to minimal
order of 10 to 20 per day, clustered in two or three short time
levels, often completely ceasing, within seconds after move-
periods.31 In terrestrial mammals such as the rat, visual scor-
ment. In contrast, most brainstem cells in the same brain
ing of twitches as has been done in cetaceans registers thou-
regions of mammals show tonic activity even during quiet
sands of twitches each day, most in REM sleep (J. M. Siegel
waking. This tonic waking activity presumably allows for more
and O. I. Lyamin, unpublished observations). Therefore, even
rapid response to sensory inputs. One can speculate that
if the twitches in cetaceans mark a REM sleep–like state, the
tonic waking activity is accompanied by the need for the
amount of REM sleep in cetaceans would be the lowest among
inactivity–activity cycle that accompanies mammalian sleep.
the mammals. Further investigations of cetacean sleep aimed
Belief that the emergence of PS was relatively recent was
at addressing the question of the amount and nature of REM
encouraged by its reported absence (although quiet sleep was
sleep–like phenomena might provide fundamental insights
present) in the Australian short-nosed echidna,21 one of the
into the evolution and function of REM sleep.
three surviving monotremes (egg-laying mammals) that
diverged early from the main paths of mammalian evolution
COMPARATIVE THEORIES
(see Fig. 8–4). It may seem that some characteristic of the
echidna (mode of reproduction, fossorial adaptation, or low
The emerging evolutionary perspective has undermined the
body temperature) obviates the need for PS and explains its
previously dominant influence of commonsensical restorative
absence, but this is not the case. The presence of PS in birds
theory, which holds that sleep is for relief of bodily or cerebral
shows that it is not related to viviparity. PS is clearly present
deficits caused by waking activity. Restorative theory cannot
in other fossorial mammals (e.g., the mole and blind mole rat)
readily explain the dramatic interspecies variation in daily
and in species with very low body temperature (e.g., the sloth).
mammalian sleep quotas (Table 8–1). (For a comprehensive
Need for qualification of the belief that PS is absent in the
compilation, see Principles and Practice of Sleep Medicine, first
echidna is indicated, however, by a finding that, at some times
edition, pp. 39-41. See also Zepelin and Rechtschaffen,34
when the echidna’s EEG indicates quiet sleep, there are bursts
Meddis,35 and Elgar et al.36,37).
of neuronal activity in its brainstem similar to activity charac-
Inspired by this variation, comparative theories have been
teristic of REM sleep in therian mammals.22 This is said to
advanced as alternatives. Guided by an assumption that sleep
indicate that REM and NREM sleep did not evolve sequen-
varies with complexity of the brain, some of these assert that
tially but as a differentiation of a primitive state that held the
sleep has cerebral functions. PS has attracted interest in this
seeds of both sleep states. Furthermore, there is now unequiv-
respect because of its cerebral activation. For example, taking
ocal evidence of REM sleep in another of the three surviving
note of instinctive behaviors (e.g., rage) released during PS in
monotremes, the platypus, occupying 6 to 8 hours per day
cats whose postural atonia is surgically abolished, Jouvet38
(more than in any other mammal) and accompanied by eye
suggested that PS evolved for daily reprogramming of innate
movements, atonia, twitching, and an elevated response
behaviors to preserve them in species that rely chiefly on
threshold, as generally found in mammalian PS, although with
learning. This view is attractive as an explanation for the
EEG voltage that may be at a level characteristic of quiet
absence of PS in reptiles and its meagerness in birds, which
sleep in eutherian mammals.23,24
rely heavily on instinctive behaviors. On the other hand, if the
Complicating speculation about the history of PS are reports
theory is correct, PS quotas should be greatest in mammals
of its absence in the bottle-nosed dolphin and the common
with the most learning ability (e.g., primates), but as can be
porpoise, which cannot be considered primitive mammals25,26
judged by the data in Table 8–1, this is not the case.
(see Chapter 10). Most NREM sleep in these species is uni-
In what can be called the eraser theory of REM sleep, Crick
hemispheric, consisting of synchronized, slow activity in one
and Mitchison have treated its reported absence in the echidna
cerebral hemisphere and desynchronized activity characteristic
as evidence that it amounts to a mechanism for reverse learning,
of wakefulness in the other. There is no bilateral high-amplitude
in which stimulation of the forebrain weakens the synaptic
(delta) sleep. This sleep organization seems necessary to
strength of undesirable “parasitic modes” of neuronal activity,
guarantee respiratory function.27 It has been suggested that PS
thus fine-tuning the brain’s operation.39 The echidna, it is
is also absent on this account, although it is present in the
said, gets by without REM sleep because its surprisingly large
northern fur seal28 and the manatee,29 both of which also have
neocortex makes reverse learning unnecessary. If true, an
unihemispheric sleep. Other reports make it advisable to
inverse relationship between size of neocortex and REM sleep
reserve judgment regarding the absence of PS in any cetaceans.
quotas is to be expected in other species, but supportive data
Its possible presence is suggested by a report of quiescent
are lacking.40 Human total sleep time, REM sleep, and percent
hanging behavior accompanied by twitching in captive beluga
of sleep time spent in REM sleep are not unusual. Other
whales.30,31 Penile erections have been reported in the absence
mammals have much higher or lower amounts of each state.

---

W0797-008 1/6/05 5:40 PM Page 95
Mammalian Sleep
95
could effect a metabolic saving of about 15%,47 and the savings
Table 8–1.
Daily Sleep Quotas in a Sample of
would largely depend on the animal’s nutrition48 and its capacity
Mammalian Species
to remain in quiet wakefulness. There are some species with
Total Daily
Daily REM
high sleep quotas and low metabolic rates, a condition trace-
Species
Sleep Time* (hr)
Time (hr)
able to energy-deficient diets.48 An outstanding example is the
endangered koala (see Table 8–1), whose diet consists of rare
Echidna
8.5
?
Platypus
14.0
7.0
types of eucalyptus leaves with low nutritional value.49 Other
Opossum
18.0
5.0
examples are edentates (e.g., the sloth) and armadillos. Such
Koala
14.5
?
species cannot afford high activity levels. Their extended sleep
Mole
8.5
2.0
seems necessary to ease metabolic pressure and is consistent
Bat
19.0
3.0
with a role for sleep in energy conservation. In humans, the
Baboon
9.5
1.0
overnight saving is more likely to be only 5% to 11%, taking
Humans
8.0
2.0
into account the effects of body movement and arousals.50-52
Armadillo
17.0
3.0
An argument against the view that sleep is for energy conser-
Rabbit
8.0
1.0
vation enforced by rest is the reported continuous movement
Rat
13.0
2.5
of dolphins while asleep.3 This is considered evidence of some
Hamster
14.0
3.0
Dolphin
10.0
?
sleep function other than enforced rest.
Seal
6.0
1.5
Guinea Pig
9.5
1.0
Correlational Findings
Cat
12.5
3.0
Ferret
14.5
6.0
The second version of energy conservation theory considers
Horse
3.0
0.5
the reduction of metabolic rate during sleep to be of minor
Elephant
4.0
?
importance. The principal contribution of sleep (with no qual-
Giraffe
4.5
0.5
ification regarding PS) is held to be the enforcement of rest so
as to set a limit on activity and energy expenditure. This view
Values are rounded to the half hour and exclude drowsiness.
Some values are averages for two or more members of the
emerged from Zepelin and Rechtschaffen’s34 comparative
same genus.
study of sleep parameters, potential life span, and other
*Total daily sleep time includes daily REM time.
constitutional variables in 53 mammalian species. The study
REM, rapid eye movement; ?, reported absence of REM sleep or
assessed a long-standing belief that species with high daily
uncertainty.
sleep quotas have relatively long potential life spans because
they benefit from lowered metabolism during sleep. The con-
trary proved to be the case: long-sleeping species are typically
short-lived. They also tend to be small in size and high in
This poses a problem for sleep-learning or “information
basal or resting metabolic rate per unit of body weight. There
processing” theories of sleep function.41
was an impressive correlation (0.63) between sleep quotas
and metabolic rate.53 Together with knowledge that the meta-
ENERGY CONSERVATION
bolic cost of physical activity varies inversely with body size54
and that daily food requirements relative to body size are dis-
A major alternative to restorative theory is the view that mam-
proportionately high in small mammals,55 the finding on
malian sleep is for energy conservation, as suggested by its
metabolic rate led to the conclusion that sleep sets a limit on
association with endothermy. Two versions of this view are
energy expenditure to the extent necessary to balance a
frequently mistaken to be the same. One of these, advocated
species’ energy budget.
by Berger42 and coworkers, holds that sleep is for reduction of
Table 8–2 summarizes relevant correlations found in
energy expenditure below the level attainable by rest alone.
updated analyses of EEG and behavioral sleep data for 85 taxa
Interdependence of quiet sleep and endothermy is inferred
(species or genera), with missing data for some taxa in
from their concurrent maturation in mammalian infancy and
each analysis. Despite recent revisions in the sleep data (Asian
the uninterrupted operation of thermoregulatory processes
elephant,56 giraffe57), added data for newly studied species
during quiet sleep at a reduced temperature level.42,43
(koala,58 ferret59), and some adjustments and expansion of the
The reduced capacity for thermoregulation during PS (see
data for constitutional variables, the results are quite similar
Chapter 24) is considered evidence that PS is a vestige of a
to previous findings. Previous and present correlations of
reptilian state of ectothermic inactivity. The notion that quiet
body weight, brain weight, and encephalization quotient with
sleep, torpor, and hibernation are related dormant states with
total daily sleep time and cycle length are virtually identical.
a common purpose, mainly energy conservation, is not
Consistency was somewhat less for correlations of metabolic
upheld by recent data (see Chapter 7). The upsurge and inten-
rate and the correlations of PS measures. Correlations of quiet
sification of SWA directly following bouts of torpor or hiber-
sleep time differed most from previous findings. This was
nation suggest compensation for loss of sleep during the
because the present analyses made no adjustment of sleep
hypothermic bouts.44-46 The metabolic expense of heating the
quotas for drowsiness.
body toward euthermia in the periodic interruptions of hiber-
Analyses of partial correlations found, as in previous
nation is close to the energy saving that results from the
research, that brain weight and cycle length were positively
immobilization and lower temperature during hibernation.
correlated independently of body weight (r = .64; P < .001),
Above all is the question of whether reduction of metabolic
but body weight and cycle length did not correlate independ-
rate is sufficient as the raison d’être for sleep. At most, sleep
ently, thus leaving brain weight alone as a likely determinant

---

W0797-008 1/6/05 5:40 PM Page 96
96
Normal Sleep and Its Variations
Table 8–2.
Correlations between Sleep Parameters and Constitutional Variables
Constitutional
Total Daily
Quiet Sleep
Paradoxical
Paradoxical
Variables
Sleep Time (hr)
Time (hr)
Sleep Time (hr)
Sleep (%)
Cycle Length
Body weight
–.53*
–.53*
–.45*
–.12
.83*
(85)
(65)
(65)
(65)
(33)
Brain weight
–.55*
–.48*
–.52*
–.25
.89*
(71)
(56)
(54)
(56)
(32)
Metabolic rate
33†
.30‡
.13
–.09
–.82*
(65)
(51)
(50)
(50)
(29)
Encephalization
–.17
–.10
–.20†
–.30‡
.52†
quotient
(69)
(55)
(53)
(55)
(32)
Common logarithmic transformations were used for the constitutional variables. Log (1 + X) transformations were used for paradoxical
sleep values. Number of cases per coefficient is in parentheses.
*P < .001
†P < .01
‡P < .05
of cycle length. Brain weight and body weight both failed to
sustained running, climbing, and other vigorous activities. A
correlate with sleep time independently of each other, as was
mammal’s basal energy requirement is at least five times that
found in a previous set of analyses.60 It therefore seems best
of an ectotherm similar in size and body temperature.61 This the-
to consider the correlations of sleep time with brain weight
ory is also in accord with findings on prolonged sleep deprivation
and body weight, without distinction between them, as con-
in rats. With ultimately fatal consequences, the rats suffered
sequences of body size. Multiple regression analysis showed
from increased metabolism, as shown by weight loss in spite of
that brain weight and body weight together (i.e., body size)
increased food intake, along with reduced body temperature
could account for 31% of the variance in sleep time.
in spite of increased metabolic rate. Prolonged deprivation of
Findings similar to those discussed earlier were obtained
REM sleep alone has similar effects.62-64 These findings indicate
by Elgar et al.36,37 in a study based solely on electrographic
sleep data for 69 species, with family as the unit of analysis.
Agreement between the two sets of findings is important
because the choice of families as the taxonomic level for analysis
guards against possible inflation of sample size owing to similar-
1000
ities between species with shared phylogenetic backgrounds.
The correlation findings underwrite the view that sleep
occurs to enforce rest and keep energy expenditure at an
100
Thermoneutral t = 9.30 M 0.44
s
b
affordable level. There may be less pressure for sleep in large
species because of their greater energy reserves. The propor-
tion of fat to body mass increases with size, ranging from less
10.0
than 5% in the smallest mammals up to 25% to 30% in
species weighing 1000 kg.55 As body size increases, the ratio of
surface to body mass decreases and thickness of fur increases.
1.0
Large mammals consequently have lower thermal conductance
25
20
(flow of heat to the environment) and wider thermoneutral
10
0.1
zones. This lessens requirements for active heat production to
Survival time (days)
0
–20
maintain body temperature. Lindstedt and Boyce,60 as illus-
°C
trated by Figure 8–5, have shown that fasting endurance
0.01
(survival time) at thermoneutrality is a function of body size,
and the size advantage is accentuated in the cold. Even at
thermoneutrality, as the figure shows, survival time is short for
0.001
small species. Many are frequently only hours away from
0.001
0.01
0.1
1.0
10.0
100
1000
death by starvation. They may require more sleep to avoid
Body mass (kg)
exhaustion of energy reserves. Also consistent with an energy
conservation role are the relatively high sleep quotas that all
Figure 8–5. Fasting endurance (or survival time) as a function of
mammals have early in the maturational period, when energy
body size for mammals. The top line represents survival time at ther-
must be channeled into growth.
moneutral ambient temperature, where t =
0.44
s
9.30 M
. Steeper lines
b
This version of energy conservation theory befits the require-
represent survival times at elevated metabolism induced by cold.
(From Lindstedt SL, Boyce MS: Seasonality, fasting endurance, and
ments of endothermy, which is not only for thermoregulation
body size in mammals. Am Nat 1985;125:873-878. The University of
alone but also for the increased aerobic capacity required for
Chicago Press. All rights reserved.)

---

W0797-008 1/6/05 5:40 PM Page 97
Mammalian Sleep
97
an indispensable role for sleep in energy regulation, as suggested
ecologic basis. Its prominence in ungulates (e.g., the horse)
by the relationships of sleep quotas and constitutional variables.
may be a compromise between sleep and alertness to predatory
The convergence here of experimental and correlational findings
threat. On the other hand, there is no simple explanation for
is theoretically promising.
the prominent drowsiness of carnivores (e.g., the cat), which
seems like purposeless fraying of sleep.
Immobilization Theory
Predation is the ecologic variable that has attracted most
interest. Because of the scarcity of data on the extent to which
Because correlational findings in support of energy conserva-
individual species are subject to predation or have suffered
tion theory fall short of explaining even half of the variance in
from it in their evolutionary past, judgments of its influence
sleep quotas, the way is open for other viewpoints. Restorative
are open to question. Findings by Allison and Cicchetti,65
theories having failed to explain species differences, behav-
however, raised the possibility that PS quotas are reduced by
ioral theories have come to the fore, expanding the concept of
predatory threat because the elevation of sensory thresholds
sleep as a state of enforced rest. It is suggested, for example,
during PS puts prey at a disadvantage.
that large mammals sleep relatively little because they require
Relatively neglected in ecologic theorizing about sleep is
extra time for foraging.65 A related view is that the function of
the role of species differences in reproductive strategies and
sleep amounts to adaptive nonresponding,35,65 meaning that it
life histories. Adaptation to the environment occurs not only
prevents activity (e.g., foraging) when it would be dangerous
through fine-tuning of physical and behavioral characteristics
or inefficient, and it blocks harmful reactions that might occur
but also through changes in the number of offspring and the
in an animal merely resting but aware of ongoing events.
timing of their maturation. Interspecies variation in sleep may
Meddis35 (p. 54) elaborated, “The benefits of sleep depend
be secondary to such adaptations. Commanding attention in
upon the species. For some the conservation of energy is most
this respect is the variation of REM sleep with maturational
valuable, for others, the protection from predation… for others,
variables, as illustrated in Figures 8–7 and 8–8.
the timing element.” In effect, however, such theory is untestable,
In precocial species, that is, those born fairly mature (e.g.,
for there will always be some need that sleep can be said to
the guinea pig, sheep), the REM sleep percentage at birth
meet. Filling spare time can be considered a function of sleep.
is low and near the adult level. In altricial species, that is, those
But how much spare time does a species have? With circular
born immature (e.g., the rat, the cat), the REM sleep percent-
reasoning, it is argued that time in sleep is a measure of spare
age is initially high and remains relatively high even after
time (Fig. 8–6).
maturation.16 Also indicative of influence by maturational
timing is the inverse correlation of REM quotas with gestation
ECOLOGIC INFLUENCES: THE
time.34,36
ALTRICIAL–PRECOCIAL DIMENSION
Previous findings of correlation between daily REM sleep
time and degree of maturity at birth were confirmed with
It is often assumed that species differences in sleep reflect
expanded data for up to 65 viviparous species or genera
environmental influences, but this is not readily apparent.
(echidna and platypus excluded). For eutherian mammals,
Probably the clearest case of an ecologically determined char-
acteristic is the unihemispheric sleep found in some marine
mammals. Drowsiness in some species seems to have an
100
Rat
90
80
Cat
70
60
50
40
30
20
Guinea pig
Active-REM sleep as % of total sleep
10
0
1
5
10
15
20
30
40
Adult
Age (days)
level
Figure 8–7. Maturational changes in rapid eye movement (REM)
sleep as a percentage of total sleep time in two altricial species (rat
Figure 8–6. Nine African lions share a ribbon of shade and sleep in
and cat) and a precocial species (guinea pig). (Reprinted from Jouvet-
the heat of day. (From Schaller GB: Life with the King of Beasts. Natl
Mounier D, Astic L, Lacote D: Ontogenesis of the states of sleep in rat,
Geog 1969;135:494-519. By permission of Dr. George B. Schaller.
cat, and guinea pig during the first postnatal month. Dev Psychobiol
Copyright 1969, National Geographic Society. All rights reserved.)
1970;2:216-239, by permission. Copyright 1970, John Wiley & Sons.)

---

W0797-008 1/6/05 5:40 PM Page 98
98
Normal Sleep and Its Variations
4
European
hedgehog
Cat
Hamster
3
Armadillo
Dog
Red
fox
Tree shrew
Pig
Laboratory
rat
2
Man
Northern
Squirrel
House
fur seal
monkey
Chinchilla
mouse
Chimpanzee
Daily adult REM time (hr)
Macaque
1
Cotton
Baboon
Rabbit
rat
Guinea pig
Cow
Horse
Sheep
Rock
Goat
hyrax
0
10
20
30
40
50
60
70
Neonatal percent of adult brain weight
Figure 8–8. Daily rapid eye movement (REM) sleep time in adult mammals as a function of neonatal maturity, represented by the percentage
of adult brain weight in neonates of the same species.
the four-point altricial–precocial (A-P) ratings were assigned
and they have the highest amount of REM sleep time of any
previously by an expert mammalogist without knowledge of
animal yet studied. The absence or minimal presence of PS
the sleep data (Table 8–3). For this analysis, the author gave
in cetaceans is understandable on the basis of their extreme
marsupials a rating of 1 based on the criteria of the scale. The
precociality.
results, shown in Table 8–4, were independent of body weight
These analyses also found a correlation of .49 (P < .001)
and brain weight and differed from previous findings chiefly
between daily quotas of PS and quiet sleep. In the 49 species
in terms of a significant correlation in the present research
with complete data, this correlation with body weight and
between the PS percentage and the gestation period. It is note-
worthy that in the four marsupial species, the two with the
shortest gestation periods had PS quotas and percentages
radically higher than the others—the highest in the entire
sample. The recent findings in the platypus strengthen these
Table 8–4.
Correlations of Paradoxical Sleep
correlations, as these animals are highly immature at birth
Parameters with Measures of Neonatal
Maturity and Reproductive Variables
Measures of Neonatal
Maturity and
Paradoxical
Reproductive
Sleep
Paradoxical
Table 8–3.
Altricial–Precocial Scale for Neonates
Variables
Time (hr)
Sleep (%)
of Viviparous Species
Altricial–precocial rating
–.66*
–.45†
Scale Value
Description
(65)
(64)
Neonatal brain weight (%)
–.61*
–.55†
1
Eyes closed; naked; rolls; sometimes can
(27)
(27)
cling (1.5: fur shows)
Gestation period
.63*
–.39†
2
Eyes barely closed or just open; furred;
(60)
(59)
crawls well (2.5: eyes open; can cling)
Litter size
.51*
.41†
3
Eyes open; furred; can stand
(63)
(62)
4
Eyes open; furred; can walk and
follow or swim
To minimize skewness, common logarithmic transformations
were used for gestation period and litter size. Log (1 + X)
Intermediate scale values (e.g., 2.8) were assigned to some
transformations were used for paradoxical sleep values.
species.
Number of cases per coefficient is in parentheses.
Adapted from Eisenberg JF. The Mammalian Radiations.
*P < .001
Chicago, Ill: University of Chicago Press; 1981. Copyright
†
1981 by The University of Chicago. All rights reserved.
P < .01

---

W0797-008 1/6/05 5:40 PM Page 99
Mammalian Sleep
99
brain weight partialed out was .28 (P < .05). These results
Clinical Pearl
are consistent with the view that PS provides endogenous
stimulation to the brain to promote recovery from sleep.66
Although many sleep parameters differ between species,
The striking relationships between REM quotas and neona-
human sleep does not appear to be qualitatively unique.
tal characteristics give no clear indication of the maturational
This factor makes animal models suitable for the investiga-
events that are responsible. If one adopts Parmeggiani’s67 view
tion of many aspects of sleep pharmacology and pathology.
that REM physiology is under rhombencephalic regulation
(as opposed to the hypothalamic regulation in the rest of
Acknowledgment
sleep), then maturation of the hypothalamus may be the crit-
ical factor. Given the expression of fetal characteristics (lapse
Thanks to Cathleen S. Zepelin for artwork and assistance with
of thermoregulation, respiratory irregularity, and twitching) in
data analysis.
REM sleep during maturity, REM sleep could be considered a
REFERENCES
carryover from fetal life. The difference between mammals and
birds in the representation of REM may well be explained by
1. Rechtschaffen A: Current perspectives on the function of sleep.
unidentified differences between them in maturational timing.
Perspect Biol Med 1998;41:359-390.
The association of REM with altriciality takes on added
2. Campbell SS, Tobler I: Animal sleep: A review of sleep duration
significance with the realization that reptiles are strictly pre-
across phylogeny. Neurosci Biobehav Rev 1984;8:269-300.
cocial and that altriciality evolved in birds and mammals in
3. Pilleri G: The blind Indus dolphin, Platanista indi. Endeavour
1979;3:48-56.
conjunction with endothermy. Evolutionary theorists agree
4. Lyamin OI, Mukhametov LM, Siegel JM: Relationship between
that the metabolic cost of endothermy favored altriciality,
sleep and eye state in Cetaceans and Pinnipeds. Arch Ital Biol
which meant the birth of exothermic rather than endothermic
2004;142:557-568.
young. One view is that altriciality reduced the energy require-
5. Bourliere F: The Natural History of Mammals, 3rd ed. New York,
ments for maturation by placing reliance on parental body
Alfred A Knopf, 1967.
heat for temperature regulation in the young, allowing their
6. Hediger H: Comparative observations on sleep. Proc R Soc Med
food to be channeled primarily into growth.68 Another view is
1969;62:1-4.
that altriciality mainly allowed greater flexibility in meeting
7. Hediger H: The biology of natural sleep in animals. Experientia
energy requirements by shortening gestation and incubation
1980;36:13-16.
periods, putting off much of the expense of propagation until
8. Moss C: Portraits in the Wild. Boston, Houghton Mifflin, 1975.
9. Van Twyver H: Sleep patterns in five rodent species. Physiol
after birth, thus splitting the expense between parents, both of
Behav 1969;4:901-905.
whom could then forage for the young.69 Precociality in some
10. Zepelin H: Sleep in the jaguar and the tapir: A prey-predator
mammals evolved as a later adaptation.
contrast. Psychophysiology 1970;7:306.
Rather than outright inheritance from reptiles, as the finding
11. Rechtschaffen A, Kales A: A Manual for Standardized
of REM sleep in the platypus has suggested to some writers,
Terminology, Techniques, and Scoring System for Sleep Stages of
REM sleep seems to be part of the evolution of endothermy,
Human Subjects. Washington, DC, Public Health Service, US
which distinguishes mammals from reptiles. It is simultane-
Government Printing Office, 1968.
ously part of the transition from the strict precociality of
12. Ursin R: The two stages of slow wave sleep in the cat and their
reptiles. The limited indications of REM sleep features in the
relation to REM sleep. Brain Res 1968;11:347-356.
echidna are evidence of this transitional origin as opposed to
13. Huber R, Deboer T, Tobler I: Sleep deprivation in prion protein
deficient mice and control mice: Genotype dependent regional
straightforward inheritance from reptilian forebears.
rebound. Neuroreport 2002;13:1-4.
14. Borbély AA, Tobler I: Homeostatic and circadian principles in
SUMMARY
sleep regulation in the rat. In McGinty DJ (ed): Brain Mechanisms
of Sleep. New York, Raven Press, 1985, pp 35-44.
Evolutionary history suggests that mammalian sleep developed
15. McGinty DJ, Siegel JM: Sleep states. In Satinoff E, Teitelbaum P
in association with endothermy, paralleling a similar develop-
(eds): Motivation. (Handbook of Behavioral Neurobiology, vol 6.)
ment of sleep in birds. Consistent with this impression are sev-
New York, Plenum Press, 1983, pp 105-181.
eral clusters of correlated variables that must also be taken into
16. Jouvet-Mounier D, Astic L, Lacote D: Ontogenesis of the states
of sleep in rat, cat, and guinea pig during the first postnatal
account in theorizing about the function of sleep. One is the cor-
month. Dev Psychobiol 1968;2:216-239.
relation of daily mammalian sleep quotas with body weight and
17. Frank MG, Heller C: The ontogeny of mammalian sleep: A
weight-specific metabolic rate, indicating greater requirements of
reappraisal of alternative hypotheses. J Sleep Res 2003;12:25-34.
sleep in species with low energy reserves. Another is the correla-
18. Allison T, Van Twyver H: The evolution of sleep. Nat Hist 1970;
tion of REM sleep parameters with A-P status, gestation time,
79:56-65.
and litter size, indicating the influence of maturational time-
19. Siegel JM: The evolution of REM sleep. In Lydic R, Baghdoyan
tables on the prominence of REM sleep and suggesting that it is
HA (eds): Handbook of Behavioral State Control. Boca Raton,
a byproduct of altriciality and endothermy. Less clearly related to
Fla, CRC Press, 1999, pp 87-100.
endothermy is the positive correlation of the NREM-REM cycle
20. Eiland MM, Lyamin OI, Siegel JM: State-related discharge of
length with brain weight, a relationship that is also apparent in
neurons in the brainstem of freely moving box turtles, Terrapene
carolina major. Arch Ital Biol 2001;139:23-36.
the increase of cycle length with maturation in mammals. On
21. Allison T, Van Twyver H, Goff WR: Electrophysiological studies
the basis of an old comparative study that showed a high corre-
of the echidna, Tachyglossus aculeatus: I. Waking and sleep. Arch
lation of mammalian brain weight with cortical acetylcholine
Ital Biol 1972;110:145-184.
content and cholinesterase activity,70 it could be speculated that
22. Siegel JM, Manger PR, Nienhuis R, et al: The echidna
cholinergic mechanisms known to influence cycle length are
Tachyglossus aculeatus combines REM and non-REM aspects in a
probably its operative link with brain weight.71
single sleep state. J Neurosci 1996;16:3500-3506.

---

W0797-008 1/6/05 5:40 PM Page 100
100
Normal Sleep and Its Variations
23. Siegel JM: Monotremes and the evolution of REM sleep. Sleep
46. Trachsel L, Edgar DM, Heller HC: Are ground squirrels
Res Soc Bull 1997;4:31-32.
sleep-deprived during hibernation? Am J Physiol 1991;260:
24. Siegel JM, Manger PR, Nienhuis R, et al: Sleep in the platypus.
R1123-R1129.
Neuroscience 1999;91:391-400.
47. Shapiro CM, Goll CC, Cohen GR, et al: Heat production during
25. Mukhametov LM, Supin AY: EEG study of different behavioral
sleep. J Appl Physiol 1984;56:671-677.
states in free-moving dolphin, Tursiops truncatus [in Russian]. Zh
48. McNab BK: The influence of food habits on the energetics of
Vyssh Nerv Deiat 1975;25:396-401.
eutherian mammals. Ecol Monogr 1986;56:1-19.
26. Mukhametov LM, Poliakova IG: EEG investigation of sleep in
49. Lee A, Martin R: Life in the slow lane. Nat Hist 1990;8:34-42.
porpoises, Phocoena phocoena [in Russian]. Zh Vyssh Nerv Deiat
50. Ravussin E, Lillioja S, Anderson TE, et al: Determinants of
1981;3:333-339.
24-hour energy expenditure in man: Methods and results using
27. Mukhametov LM, Supin AY, Poliakova IG: Interhemispheric
a respiratory chamber. J Clin Invest 1986;78:1568-1578.
asymmetry of the electroencephalographic sleep patterns in
51. Garby L, Kurzer MS, Lammert O, et al: Energy expenditure
dolphins. Brain Res 1977;124:581-584.
during sleep in men and women: Evaporative and sensible heat
28. Mukhametov LM, Liamin OI, Poliakova IG: Sleep and wakeful-
losses. Hum Nutr Clin Nutr 1987;41:225-233.
ness in Callorhinus ursinus [in Russian]. Zh Vyssh Nerv Deiat
52. Goldberg GR, Prentice AM, Davies HL, et al: Overnight and
1984;34:465-471.
basal metabolic rates in men and women. Eur J Clin Nutr
29. Mukhametov LM, Lyamin OL, Chetyrbok IS: Sleep and wake-
1988;42:137-144.
fulness in an Amazonian manatee. In Horne JA (ed): Sleep 1990:
53. Zepelin H: Mammalian sleep, metabolic rate, and body size.
Proceedings of the Tenth European Congress on Sleep Research,
Sleep Res 1986;15:62.
Strasbourg, France, 1990, May 20-25. Bochum, Germany,
54. Schmidt-Nielsen K: Locomotion: Energy cost of swimming, flying,
Pontenagel Press, 1990, pp 119-122.
and running. Science 1972;177:222-228.
30. Lyamin OI, Mukhametov LM, Siegel JM, et al: Unihemispheric
55. Calder WA III: Size, Function, and Life History. Cambridge,
slow wave sleep and the state of the eyes in a white whale. Behav
Mass, Harvard University Press, 1984.
Brain Res 2002;129:125-129.
56. Tobler I: Behavioral sleep in the Asian elephant in captivity.
31. Lyamin OI, Shpak OV, Nazarenko EA, Mukhametov LM: Muscle
Sleep 1992;15:1-12.
jerks during behavioral sleep in a beluga whale (Delphinapterus
57. Tobler I, Schwierin B: Behavioural sleep in the giraffe (Giraffa
leucas L). Physiol Behav 2002;76:265-270.
camelopardalis) in a zoological garden. J Sleep Res 1996;5:21-32.
32. Coleman RM, Pollak CP, Weitzman ED: Periodic movements in
58. Nagy KA, Martin RW: Field metabolic rate, water flux, food con-
sleep nocturnal myoclonus: Relation to sleep disorders. Ann
sumption, and time budget of koalas, Phascolarctos cinereus, in
Neurol 1980;8:416-421.
Victoria. Aust J Zool 1985;33:655-665.
33. Affanni JM, Cervino CO, Marcos HJ: Absence of penile erections
59. Marks GA, Shaffery JP: A preliminary study of sleep in the ferret,
during paradoxical sleep: Peculiar penile events during wakeful-
Mustela putorius furo: A carnivore with an extremely high
ness and slow wave sleep in the armadillo. J Sleep Res 2001;10:
proportion of REM sleep. Sleep 1996;19:83-93.
219-228.
60. Lindstedt SL, Boyce MS: Seasonality, fasting endurance, and
34. Zepelin H, Rechtschaffen A: Mammalian sleep, longevity, and
body size in mammals. Am Nat 1985;125:873-878.
energy metabolism. Brain Behav Evol 1974;10:425-470.
61. Bennett AF, Ruben JA: Endothermy and activity in vertebrates.
35. Meddis R: The Sleep Instinct. London, Routledge and Kegan
Science 1979;206:649-654.
Paul, 1977, p 54.
62. Rechtschaffen A, Gilliland MA, Bergmann BM, et al: Physiological
36. Elgar MA, Pagel MD, Harvey PH, et al: Sleep in mammals. Anim
correlates of prolonged sleep deprivation in rats. Science 1983;
Behav 1988;36:1407-1419.
221:182-184.
37. Elgar MA, Pagel MD, Harvey PH, et al: Sources of variation in
63. Rechtschaffen A, Bergmann BM, Everson CA, et al: Sleep depri-
mammalian sleep. Anim Behav 1990;40:991-994.
vation in the rat: X. Integration and discussion of the findings.
38. Jouvet M: The function of dreaming: A neurophysiologist’s point
Sleep 1989;12:68-87.
of view. In Gazzaniga MS, Blakemore C (eds): Handbook of
64. Rechtschaffen A, Bergmann BM: Sleep deprivation in the rat by
Psychobiology. New York, Academic Press, 1975, pp 499-527.
the disk-over-eater method. Behav Brain Res 1995;69:55-63.
39. Crick F, Mitchison G: The function of dream sleep. Nature
65. Allison T, Cicchetti DV: Sleep in mammals: Ecological and
1983;304:111-114.
constitutional correlates. Science 1976;194:732-734.
40. Zepelin H: Encephalization and species differences in REM sleep:
66. Vertes RP: A life-sustaining function for REM sleep: A theory.
A test of the Crick-Mitchison theory. Sleep Res 1985;14:85.
Neurosci Biobehav Rev 1986;10:371-376.
41. Siegel JM: The REM sleep-memory consolidation hypothesis.
67. Parmeggiani PL: Regulation of physiological functions during
Science 2001;294:1058-1063.
sleep in mammals. Experientia 1982;38:1405-1408.
42. Berger RJ: Bioenergetic functions of sleep and activity rhythms
68. Hopson JA: Endothermy, small size, and the origin of mam-
and their possible relevance to aging. Fed Proc 1975;34:97-102.
malian reproduction. Am Nat 1973;107:446-452.
43. Glotzbach SF, Heller HC: Central nervous regulation of body
69. Case TJ: Endothermy and parental care in the terrestrial verte-
temperature during sleep. Science 1976;194:537-539.
brates. Am Nat 1978;112:861-874.
44. Walker JM, Berger RJ: Sleep as an adaptation for energy conser-
70. Tower DB, Elliott KAC: Activity of acetylcholine system in cerebral
vation functionally related to hibernation and shallow torpor.
cortex of various unanesthetized mammals. Am J Physiol 1952;
Prog Brain Res 1980;53:255-278.
168:747-759.
45. Deboer T, Tobler I: Sleep EEG after daily torpor in the
71. Sitaram N, Moore AM, Gillin JC: Experimental acceleration and
Djungarian hamster: Similarity to the effects of sleep depriva-
slowing of REM sleep ultradian rhythm by cholinergic agonist
tion. Neurosci Lett 1994;166:35-38.
and antagonist. Nature 1978;274:490-492.

---