Cold Exposure and Sleep in the Rat: Effects on Sleep Architecture and the
Electroencephalogram
Matteo Cerri, MD, PhD1; Adrian Ocampo-Garces, MD, PhD2; Roberto Amici, MD1; Francesca Baracchi1; Paolo Capitani3; Christine Ann Jones, PhD1; Marco Luppi,
PhD1; Emanuele Perez, MD1; Pier Luigi Parmeggiani, MD1; Giovanni Zamboni, MD1
Department of Human and General Physiology, Alma Mater Studiorum-University of Bologna, Italy; 2Programa de Fisiología y Biofísica, Instituto de
Ciencias Biomédicas, Facultad de Medicina, Universidad de Chile, Santiago, Chile; 3Department of Electronics, Computer Science and Systems, Alma
Mater Studiorum-University of Bologna, Italy
1
power during non-rapid eye movement sleep was decreased in animals
exposed to the lowest ambient temperatures and increased during the first
day of the recovery. In contrast, rapid eye movement sleep was greatly
depressed by cold exposure and showed an increase during the recovery.
Both of these effects were dependent on the ambient temperature of the
exposure. Moreover, theta power was increased during rapid eye movement sleep in both the exposure and the first day of the recovery.
Conclusion: These findings show that sleep-stage duration and electroencephalogram power are simultaneously affected by cold exposure. The
effects on rapid eye movement sleep appear mainly as changes in the
duration, whereas those on non-rapid eye movement sleep are shown by
changes in delta power. These effects are temperature dependent, and
the decrease of both parameters during the exposure is reciprocated by
an increase in the subsequent recovery.
Key words: Low ambient temperature, sleep deprivation, NREM sleep,
REM sleep, single REM sleep, sequential REM sleep, delta power density, theta power density
Citation: Cerri M; Ocampo-Garces A; Amici R et al. Cold exposure and
sleep in the rat: effects on sleep architecture and the electroencephalogram. SLEEP 2005;28(6):694-705.
Study Objectives: Acute exposure to low ambient temperature modifies
the wake-sleep cycle due to stage-dependent changes in the capacity to
regulate body temperature. This study was carried out to make a systematic analysis of sleep parameters during the exposure to different low
ambient temperatures and during the following recoveries at ambient temperature 24°C.
Design: Electroencephalographic activity, hypothalamic temperature, and
motor activity were studied during a 24-hour exposure to ambient temperatures ranging from 10°C to -10°C and for 4 days during the recovery.
Setting: Laboratory of Physiological Regulation during the Wake-Sleep
Cycle, Department of Human and General Physiology, Alma Mater Studiorum-University of Bologna.
Subjects: Twenty-four male albino rats.
Interventions: Animals were implanted with electrodes for electroencephalographic recording and a thermistor for measuring hypothalamic
temperature.
Measurements and Results: Wake-sleep stage duration and the electroencephalographic spectral analysis performed by fast Fourier transform
were compared among baseline, exposure, and recovery conditions. The
amount of non-rapid eye movement sleep was slightly depressed by cold
exposure, but no rebound was observed during the recovery period. Delta
different low ambient temperatures was shown to primarily affect REM-sleep occurrence and that the observed decrease was
proportional to the ambient temperature of exposure, while morecomplex effects were observed on non-REM (NREM) sleep.1,2
With respect to the latter, “spindle sleep” was progressively depressed at ambient temperatures below 0°C, while slow-wave
sleep was maximally decreased at an ambient temperature of 5°C
but increased toward control levels as the temperature was lowered.1,2,7 This depression in sleep occurrence by the exposure to
low ambient temperature has now been confirmed by many studies on different species.8-15
Exposure to low ambient temperature also clearly influences
sleep occurrence when animals are allowed to recover at normal
ambient temperature in the laboratory. In particular, a rebound of
REM sleep, which was proportional to the degree of the previous
REM sleep loss, has been observed in the cat.16 These results have
been confirmed in recent studies on the albino rat, in which it was
observed that both REM-sleep loss and the subsequent REM-sleep
rebound were quantitatively related to the thermal load (duration
of the exposure × decrease in ambient temperature with respect
to normal laboratory conditions).17-19 It has also been shown that
the amount of NREM sleep is less affected during recovery, since
no substantial increase in its amount has been found in either the
cat1,2 or the rat.14 However, an increase in the electroencephalogram (EEG) power density in the delta band (0.75-4 Hz) during
NREM sleep has been observed in the rat.14
INTRODUCTION
EXPOSURE TO AN AMBIENT TEMPERATURE OUTSIDE
THE APPROPRIATE THERMONEUTRAL RANGE CHANGES THE AMOUNT AND DISTRIBUTION OF THE DIFFERENT stages of the wake-sleep cycle in several different species.1-3
These changes may be a consequence of the differences in the
capacity to regulate body temperature across the different stages
of the wake-sleep cycle.3,4 In particular, it has been shown that
thermoregulatory responses like shivering, panting, and peripheral vasomotion are suppressed during rapid eye movement (REM)
sleep5 due to a change in the activity of the central thermostat at
the preoptic-hypothalamic level.6
In early studies carried out in the cat, a short-term exposure to
Disclosure Statement
This was not an industry supported study. Drs. Zamboni, Cerri, OcampoGarces, Amici, Baracchi, Capitani, Jones, Luppi, Perez, and Parmeggiani
have indicated no financial conflicts of interest.
Submitted for publication November 2004
Accepted for publication February 2005
Address correspondence to: Giovanni Zamboni, MD, Department of Human
and General Physiology, Alma Mater Studiorum-University of Bologna,Piazza
P.ta S. Donato, 2, I-40126 Bologna, Italy; Tel: 39 051 2091742; Fax: 39 051
251731; E-mail: gzamboni@biocfarm.unibo.it
SLEEP, Vol. 28, No. 6, 2005
694
Cold Exposure and Sleep in the Rat—Cerri et al
These studies suggest that exposure to low ambient temperature can be considered as a powerful physiologic tool for use in
sleep-deprivation studies, since the degree of deprivation can be
modulated by changing either its duration (time of exposure) or its
intensity (ambient temperature levels). This contrasts with other
more traditional methods of sleep deprivation in which the degree
of deprivation can only be affected by changing its duration.
As yet, studies lack either a complete analysis of the wakesleep cycle during the period of exposure to low ambient temperature or an assessment of the recovery for a period longer than
1 day. Therefore, we carried out an experiment aiming to cover
these aspects of wake-sleep regulation.
In the present paper, the effects of cold exposure on the wakesleep cycle and the subsequent recovery period are addressed in
terms of the amount, distribution, and EEG power (in specific,
frequency bands) of the different wake-sleep stages. Also, the
analysis of the recovery is extended over 4 consecutive days.
A quantitative analysis of the relationship between the sleep
loss during cold exposure and its gain during the recovery with
respect to the controversial homeostatic/nonhomeostatic aspects
of sleep regulation20-27 will be addressed in a further paper. Preliminary results have been presented in abstract form.28,29
exchange of air. The bottom was fitted with an electric heater and
contained 2 Plexiglas cages (24 × 24 × 60 cm; cage floor, 20 cm
above ground) in which animals were individually kept. Box temperature was measured by using a temperature sensor (National
Semiconductor, Santa Clara, CA, USA) placed in the middle of
the box at cage floor level. The sensor was connected to an external proportional controller. The box was kept in a sound-proof
room and connected to the recording and control instruments on
the outside. EEG, hypothalamic temperature, and the motor activity of each animal were continuously recorded during each experimental session except between 9:00 and 9:15 AM when cage
bedding, water, and food were changed.
Following 2 days of recording for the baseline condition (BL1
and BL2), animals were exposed for 24 hours to different low
ambient temperatures (E1). The exposure was started at the onset (9:00 AM) of the light period of the light-dark cycle. Four experimental groups of animals (n=6 for each group) were studied.
Each group was exposed to a different ambient temperature, ie,
10°C±1°C, 5°C±1°C, 0°C±1°C, and -10°C±1°C (E(10), E(5),
E(0), and E(-10), respectively). Following cold exposure, animals
were returned to an ambient temperature of 24°C and allowed
to recover for 4 consecutive days (R1-R4). Animals remained in
the home cage during the temperature changes, and the selected
ambient temperature in the recording box was obtained over a
range of 15 to 45 minutes, depending upon the previous ambient
temperature.
METHODS
Twenty-four male Sprague-Dawley rats (Charles River), which
had free access to food and water and were kept at 24°C±0.5°C
ambient temperature (ambient temperature), under a 12-hour:12hour light-dark cycle (light: 9:00 AM-9:00 PM; 100 lux at cage
level), were used. The experiments were carried out according to
the European Union Directive (86/609/EEC) and were under the
supervision of the Central Veterinary Service of the University of
Bologna and the National Health Authority.
Signal Analysis
User software was developed (QuickBASIC, Microsoft, CA,
USA) to handle the data. In each experiment, the EEG and hypothalamic temperature signals were amplified (amplification factor
for the EEG approximately 7000), filtered (for the EEG: highpass filter: -40 dB at 0.35 Hz; low-pass filter: -6 dB at 35 Hz)
and, after analog-digital conversion, (sampling rate: EEG, 128
Hz; hypothalamic temperature, 8 Hz) were stored on a personal
computer (486/100 DX-4). The EEG signal was subjected to online fast Fourier transform, and EEG power values were obtained
for 4-second epochs in the delta (DPW: 0.75-4 Hz), theta (TPW:
5.5-9 Hz), and sigma (SPW: 11-16 Hz) bands.
Motor activity was monitored by means of a passive infrared
detector (Siemens, PID11, Munich, FRG) placed at the top of
each cage. The signal was amplified and integrated before analog-digital conversion (sampling rate: 8 Hz) in order to make the
output proportional to the amplitude and the duration of movement (motor-activity intensity). During the period of adaptation,
the amplitudes of the signals for each animal were routinely controlled using an oscilloscope to ensure that they were kept within
an appropriate scale of values. This system detected most of the
movements related to the normal behavior of the rat, such as exploring, grooming, feeding, and small movement during twitching or brief awakenings in either REM sleep or NREM sleep.
However, it did not always detect changes while the animal was
drinking.
Surgery
Animals were placed under deep general anaesthesia (diazepam, Valium Roche, 5 mg/kg intramuscular; ketamine-HCl, Ketalar, Parke-Davis, 100 mg/kg intraperitoneal) and were implanted
epidurally on the right side with 2 stainless-steel electrodes for
frontal-parietal EEG recording (3.0 mm anterior - 3.0 mm lateral
to Bregma; 4.0 mm posterior - 1.5 mm lateral to Bregma).
A thermistor mounted inside the tip of a stainless-steel needle
(21 gauge), which had been coated with 3 layers of electrode varnish, was positioned above the left anterior hypothalamus to measure hypothalamic temperature. The plastic plugs to connect EEG
electrodes and the thermistor to the recording apparatus were embedded in acrylic dental resin anchored to the skull by small stainless-steel screws, which had also been implanted epidurally on the
outer limit of the surgical field.
Experimental Design
Animals were allowed to recover from surgery and to adapt to
the recording apparatus in a thermoregulated and sound-attenuated box for at least 1 week. The box (120 × 50 × 118 cm ) consisted
of a commercial chest freezer adapted by replacing the top with
stainless-steel panels fixed to the frame of the freezer door. The
top was also adapted for carrying a system of fiber optics, which
were used for lights, infrared sensitive cameras that were used to
monitor animal behavior, and electric fans to allow a continuous
SLEEP, Vol. 28, No. 6, 2005
Data Analysis and Wake-Sleep Staging
REM-Sleep Parameters
EEG, hypothalamic temperature, and motor-activity signals
were visually scored in order to assess REM-sleep parameters,
695
Cold Exposure and Sleep in the Rat—Cerri et al
ie number and duration of REM-sleep episodes and total amount
of REM sleep. The main criteria used for this assessment were
based on the analysis of EEG, motor activity, and hypothalamic
temperature.17,18 Particular consideration was given to the changes
in hypothalamic temperature, since they can be considered as an
index of state-dependent changes in autonomic activity.30 For example, a REM-sleep episode was considered as such only if the
EEG changes associated with this sleep stage were concomitant
to an increase in hypothalamic temperature. Moreover, the end of
a REM-sleep episode was considered as such only if the EEG and
postural changes were associated with a decrease in hypothalamic
temperature. The time for the minimal duration of a REM-sleep
episode was fixed at 8 seconds.
A more-detailed analysis of REM sleep was carried out according to the partition of REM-sleep episodes into single and
sequential episodes.17,18 The necessity of visually scoring REMsleep parameters arises from the precision needed to identify the
interval between 2 consecutive REM-sleep episodes (REM-sleep
interval), on which such a partition is based. The time for the
minimal duration of a REM-sleep interval was fixed at 8 seconds.
Single REM-sleep episodes are defined as those that are both preceded and followed by long REM-sleep intervals (> 3 minutes),
whereas sequential REM-sleep episodes are those that are separated by at least 1 short REM-sleep interval (≤ 3 minutes) and are
found in clusters. The duration of the REM-sleep cluster has been
calculated as the total duration of the constituent sequential REMsleep episodes, without considering the intervening intervals. For
brevity, REM sleep that occurs in the form of either single or sequential episodes will be addressed as “single REM sleep” and
“sequential REM sleep,” respectively.
scored for Wake and NREM sleep by 2 experienced researchers.
The visual analysis showed that the minimal duration of an episode of Wake was 4 seconds, while, in accordance with observations of REM sleep, a duration of at least 8 seconds (two 4-second
epochs) was considered necessary for the definition of an episode
of NREM sleep. When each visually scored 4-second epoch was
matched with its IWN index, the majority of epochs visually identified as Wake had an IWN ranging from 0 to +1, while the majority
of those identified as NREM sleep had an IWN ranging from 0 to
-1. According to these results, the 4-second epochs with an IWN >
0 were then automatically attributed to Wake, while those with an
IWN < 0 were attributed to NREM sleep. A comparison between
the 2 methods for classifying the epochs as either Wake or NREM
sleep showed an agreement of 96.3%±0.1%, and a posthoc revision of the mismatches found that, unexpectedly, in some cases,
automatic scoring was more reliable than visual scoring. When
this is considered, the agreement between the 2 methods is increased to 97.4%±0.1%.
Wake and NREM-Sleep Parameters
RESULTS
Following blind visual scoring for REM sleep and the removal
of the 4-second epochs that showed EEG artifacts (which were
less than 1% and were visually scored at the end of the procedure),
REM-sleep intervals were automatically separated into Wake and
NREM sleep by using a data transformation.
A relative index for both Wake (IW) and NREM sleep (IN)
was initially calculated for each 4-second epoch: IW = [(motor
activity)e/(motor activity)bl]2; IN = [(DPW × SPW)e/(DPW ×
SPW)bl]2, where e is the actual value of the 4-second epoch and bl
represents the average 24-hour value of the 2 baseline recordings.
The 2 relative indexes were used to make a Wake-NREM sleep index: IWN = (IW - IN)/(IW + IN). The values of this index ranged from
-1.0 to +1.0 and allowed each 4-second epoch to be attributed to
either Wake or NREM sleep, since it was positive when IW was
higher than IN (typical of Wake) and negative when IN was higher
than IW (typical of NREM sleep).
The rationale for this data transformation was based on maximizing differences between the 2 states and minimizing disturbances due to the experimental conditions. DPW, SPW, and motor activity were used because they are inversely related during
REM-sleep intervals. A relative index allows different data to be
expressed in the same units and their differences to be amplified
by simple mathematical manipulations.
Automatic scoring using the IWN index for separating Wake and
NREM sleep was validated on the baseline recordings of 12 randomly selected rats that had been visually scored for REM sleep
and REM-sleep intervals. REM-sleep intervals were then blindly
For each of the 4 experimental groups, the time spent in the
different wake-sleep stages and the average values of both hypothalamic temperature and motor activity during the 2 days of
the baseline recording (BL1, BL2), the day of exposure to low
ambient temperature (E1), and the 4 days of recovery at normal
laboratory ambient temperature (R1-R4) are shown in Figure 1.
Data refer to each 24-hour period and are presented as the percentage difference from the respective average baseline levels for
wake-sleep stages and motor activity but as absolute differences
(°C) for hypothalamic temperature.
For all the parameters studied, the largest variations with respect to the baseline condition were observed during either E1 or
R1. During R1, a significant increase in the average amount of
Wake and motor activity (P<.01) and a significant decrease in the
average amount of NREM sleep, REM sleep, and hypothalamic
temperature (P<.01) can be seen. However, the percentage decrease in the amount of REM sleep was larger than that observed
for NREM sleep. Also, the decrease in the amount of REM sleep
was directly proportional to ambient temperature levels, as was
that in the form of single and sequential episodes (P<.01).
During R1, a statistically significant (P<.01) decrease in the
amount of time spent in Wake was present following all the exposure conditions, while no relevant changes in the amount of
NREM sleep were observed. Also, there was a large and significant increase in the occurrence of REM sleep (P<.01), which
SLEEP, Vol. 28, No. 6, 2005
Statistical Analysis
Statistical analysis was carried out by means of repeated-measure analysis of variance for both the time spent in the different
wake-sleep stages and the number and duration of episodes, hypothalamic temperature levels, and power densities in the different stages. A number of preplanned nonorthogonal comparisons,
which allowed the data relative to each experimental day to be
compared to the respective average baseline values, were made
by means of the modified t tests and the Bonferroni correction of
the significance level.31
Analysis of the Effects of Low Ambient Temperature on the Time
Spent in Each Wake-Sleep Stage
696
Cold Exposure and Sleep in the Rat—Cerri et al
50
%
Wake
25
**
**
0
was directly proportional to ambient temperature levels. This was
considered to be due to the comparable increase in the occurrence
of sequential REM-sleep episodes (P<.01); however, it should be
noted that a small but significant increase (P<.01) in the amount
of single REM sleep was also present.
The changes occurring in R1 appear to be quite different from
those observed in R2, R3, and R4. These are characterized by an
overall increase in the amount of REM sleep, which was statistically significant (P<.01) in R3 and was consistent with a corresponding increase (P<.01) in the amount of sequential REM
sleep. These changes were concomitant with a small, but statistically significant (P<.01), decrease in hypothalamic temperature
levels that occurred over these 3 days of recovery.
Since the largest variations in the time spent in the different
wake-sleep stages were observed during cold exposure and in the
first day of recovery, a more-detailed analysis of the circadian
aspects of these changes was made. In Table 1, the percentage
of time spent in the different wake-sleep stages during either the
light or dark periods of the light-dark cycle in baseline (average
of BL1 and BL2), E1, and R1 is shown for each of the 4 experimental groups. Data relative to REM sleep are also given as the
amount of either single or sequential REM sleep.
The major effects of cold exposure are manifest during the light
period of the light-dark cycle, during which the overall pattern
is a significant change (P<.01 vs the respective baseline values)
which is in direct or inverse proportion to ambient temperature for
NREM sleep, REM sleep, or Wake, respectively. For both Wake
and NREM sleep, no significant changes with respect to baseline
were observed during the dark period of the exposure, but at E
-10), animals appeared to spend less time in Wake and more time
in NREM sleep. In contrast, REM sleep appears to decrease in
direct proportion to ambient temperature during either the light
or dark period of the light-dark-cycle, but, during the dark period, the decrease was statistically significant (P<.01) only for
E(-10). Moreover, cold exposure strongly reduced the amplitude
of the normal circadian REM-sleep occurrence observed in baseline. The pattern of change in the amount of single and sequential
REM sleep appears to closely follow that observed for total REM
sleep.
The light and dark periods of the first day of the recovery at
normal laboratory ambient temperature were characterized by
a large REM-sleep rebound, which showed an inverse relationship to the ambient temperature of the previous exposure and was
mainly due to an increase in the occurrence of sequential REM
sleep. However, whereas the increases in both total and sequential
REM sleep during the light period were always statistically significant with respect to baseline levels (P<.01), in the dark period,
there was a significant increase (P<.01) in total REM sleep during the recovery for E(0) or E(-10) and in sequential REM sleep
(P<.05) for E(-10). The dark period of R1 for animals exposed to
E(-10) also showed a significant increase (P<.01) of single REM
sleep.
The increase in REM sleep was concomitant with a decrease in
the occurrence of either Wake during the dark period or NREM
sleep during the light period. However, the changes were only
statistically significant with respect to the baseline levels in animals that had been previously exposed to either 0°C or -10°C
(Wake: E(0), P<.01; E(-10) P<.05; NREM sleep: E(0), P<.05).
E(10)
E(5)
E(0)
E(-10)
*
*
-25
-50
50
%
NREMS
25
*
0
**
*#
-25
-50
100
*
*
**
REMS
75
%
50
25
*
0
-50
*
*
**
-75
-100
100
%
Single REMS
#*
50
*
0
*
-50
**
**
-100
200
150
Sequential REMS
%
100
50
*
0
-100
0,8
**
*
*
#
#
*
-50
°C
#
*
-25
Thy
**
**
0,4
*
0,0
*
#
*
R2
R3
R4
-0,4
-0,8
100
%
MA
50
0
*
-50
-100
BL1 BL2
E1
R1
Figure 1—Percentage of variation with respect to average baseline
levels (mean ± SEM; average baseline levels shown as 0) of (1) the
time spent in the different wake-sleep stages (Wake; non-rapid eye
movement sleep [NREMS]; and rapid eye movement sleep [REMS]);
(2) the time spent in REM sleep in the form of either single (Single
REMS) or sequential (Sequential REMS) episodes; and (3) the motor
activity (MA). The variation of the hypothalamic temperature (Thy)
is shown as the absolute value (°C). Results are from 4 groups of
animals during 2 days of baseline recording at ambient temperature
24°C (BL1, BL2), 1 day of exposure to low ambient temperature (E1),
and 4 days of recovery at ambient temperature 24°C (R1-R4). Each
group is shown under its respective ambient temperature of exposure
(E): E(10), E(5), E(0), and E(-10). The horizontal bars indicate a comparison in which the 4 groups have been taken as a whole. #P<.05 vs
baseline; *P<.01 vs baseline.
SLEEP, Vol. 28, No. 6, 2005
697
Cold Exposure and Sleep in the Rat—Cerri et al
Table 1—Total Amount of the Wake-Sleep Stages During the Exposure to Low Ambient Temperature
WAKE
Light
27.6 ± 3.3
28.7 ± 2.2
27.2 ± 0.9
27.5 ± 1.9
E(10)
E(5)
E(0)
E(-10)
NREM sleep
E(10)
61.7 ± 3.3
E(5)
60.4 ± 2.1
E(0)
61.4 ± 1.6
E(-10)
61.8 ± 1.0
REM sleep
E(10)
10.7 ± 0.9
E(5)
11.0 ± 0.5
E(0)
11.5 ± 1.1
E(-10)
10.7 ± 0.9
SINGLE REM sleep
E(10)
5.3 ± 0.5
E(5)
6.4 ± 0.9
E(0)
6.3 ± 0.6
E(-10)
6.2 ± 0.7
SEQUENTIAL REM sleep
E(10)
5.4 ± 0.8
E(5)
4.6 ± 0.9
E(0)
5.2 ± 1.4
E(-10)
4.4 ± 0.9
Baseline
Dark
72.6 ± 1.3§
72.3 ± 0.9§
73.0 ± 1.4§
74.9 ± 2.9§
Exposure
Light
Dark
43.5 ± 3.4†
74.5 ± 1.6§
53.9 ± 3.9†
72.7 ± 1.6§
57.0 ± 2.0†
77.0 ± 2.3§
63.1 ± 5.4†
70.3 ± 5.2
Recovery (day 1)
Light
Dark
26.2 ± 3.2
71.5 ± 2.5§
28.0 ± 2.2
70.0 ± 1.3§
27.5 ± 1.5
65.1 ± 1.7†§
29.3 ± 1.3
64.8 ± 1.7*§
23.2 ± 1.2§
23.1 ± 0.8§
22.8 ± 1.3§
22.5 ± 0.8§
51.8 ± 3.4†
43.3 ± 3.5†
42.0 ± 2.0†
36.7 ± 5.4†
22.1 ± 1.4§
24.7 ± 1.5§
21.3 ± 1.9§
28.8 ± 5.2
60.1 ± 3.2
57.1 ± 2.1
55.6 ± 1.8*
53.3 ± 1.5
23.0 ± 2.0§
23.7 ± 0.9§
26.3 ± 1.5§
25.3 ± 0.8§
4.2 ± 0.7§
4.6 ± 0.8§
4.2 ± 0.3§
4.4 ± 0.5§
4.8 ± 0.6†
2.8 ± 0.6†
1.0 ± 0.2†
0.3 ± 0.1†
3.4 ± 0.4
2.6 ± 0.4
1.7 ± 0.6
0.8 ± 0.3†
13.6 ± 1.3*
14.9 ± 0.7†
16.8 ± 0.6†
17.4 ± 1.3†
5.5 ± 0.8§
6.3 ± 0.7§
8.6 ± 0.3†§
9.9 ± 1.3†§
2.5 ± 0.4§
2.6 ± 0.4§
2.6 ± 0.4§
2.3 ± 0.2§
2.3 ± 0.2†
1.8 ± 0.4†
0.6 ± 0.2†
0.2 ± 0.1†
1.9 ± 0.6
1.7 ± 0.3
1.2 ± 0.5
0.5 ± 0.2 *
5.5 ± 0.8
6.5 ± 1.0
6.8 ± 0.8
7.1 ± 0.9
3.9 ± 0.7
3.8 ± 0.4‡
4.0 ± 0.6§
4.8 ± 0.6† ‡
1.6 ± 0.3§
2.1 ± 0.4§
1.6 ± 0.3§
2.2 ± 0.6‡
2.5 ± 0.6†
1.0 ± 0.3†
0.3 ± 0.1†
0.1 ± 0.1†
1.5 ± 0.3
0.9 ± 0.2
0.5 ± 0.2
0.3 ± 0.2
8.2 ± 1.3†
8.4 ± 1.3†
10.0 ± 2.5†
10.3 ± 1.4†
1.6 ± 0.2§
2.5 ± 0.5§
4.5 ± 0.5§
5.1 ± 1.5*§
Data are presented as mean ± SEM and represent percentage of 12-hour period. E(10), E(5), E(0), and E(-10) refer to ambient temperature of exposer, ie, 10°C±1°C, 5°C±1°C, 0°C±1°C, and -10°C±1°C, respectively; NREM, non-rapid eye movement sleep; REM, rapid eye movement sleep.
*P<.05 and †P<.01 vs respective baseline condition.
‡P<.05 and §P<.01 vs respective light period.
that were frequently observed during the light period of E1 and
R1 were mainly due to changes in the number of episodes. As can
be seen in the light period of the exposure condition, the number of single and sequential REM-sleep episodes and of REMsleep clusters progressively declined towards zero as the exposure temperature was lowered (P<.01 with respect to baseline for
all comparisons, except for sequential REM-sleep episodes and
REM-sleep clusters in E(10)). The average duration of the episodes was reduced to a lesser extent but did reach significance for
single REM sleep episodes in E(5) (P<.05) and E(-10) (P<.01),
for sequential episodes in E(5) (P<.01) and E(0) (P<.01), and for
REM-sleep clusters in E(5) (P<.01). During the dark period of
exposure, the number and the duration of single and sequential
REM-sleep episodes and of REM-sleep clusters were less affected by the ambient temperature than during light, with the exceptions of the duration of single episodes in E(-10) (P<.01) and
sequential episodes and clusters in E(0) and E(-10) (P<.01).
During the recovery period, the REM-sleep rebound was
mainly due to an increase in the frequency of the episodes. In the
light period, a large and significant increase, with respect to the
baseline levels, in the number of sequential REM sleep episodes
and REM sleep clusters was observed following all the exposures
except to that at ambient temperature 10°C (sequential episodes:
E(5), E(0), E(-10), P<.01; REM-sleep clusters: E(5), E(-10),
P<.01; E(0), P<.05). However, whereas the number of single and
sequential REM-sleep episodes and of REM-sleep clusters was
higher in the dark period than in baseline conditions, the difference was only statistically significant for single REM-sleep episodes in E(0) (P<.01) and REM-sleep clusters in E(-10) (P<.05).
No statistically significant changes were observed in the duration
Analysis of the Effects of Low Ambient Temperature on the Dynamics of Wake-Sleep Stages
Changes in the time spent in the different wake-sleep stages
during either cold exposure on the first day of the recovery period
have been further analyzed in terms of the number and duration
of episodes. The results of the analysis with respect to baseline
(average of BL1 and BL2), E1, and R1 are shown in Table 2 for
Wake and NREM sleep and in Table 3 for REM sleep. As shown
in Table 2, the number of Wake and NREM sleep episodes were
similar for the light and dark periods of each experimental condition and were significantly increased (P<.01), with respect to the
respective baseline levels, during the exposure to the lowest ambient temperatures (E(0) and E(-10)).
The fragmentation of wake-sleep stages was accompanied by
an increase, although not significant, in the average duration of
Wake episodes in the light period, whereas in the dark period, the
average duration significantly decreased in E(5) (P<.05) and in
both E(0) and E(-10) (P<.01). Also, the duration of NREM-sleep
episodes was characterized by a large decrease that was proportional to the lowering of ambient temperature in both light and
dark periods. This decrease was significant (P<.01) in the light
period for all the experimental conditions and only for E(0) and
E(-10) in the dark period and was consistent with the disappearance of the circadian variation in the duration of the NREM sleep
episodes. No statistically significant changes in either number or
duration of Wake and NREM-sleep episodes were observed during the first day of the recovery, and the normal circadian variation was reattained except for the number of episodes in E(-10).
Table 3 shows that the large changes in REM-sleep occurrence
SLEEP, Vol. 28, No. 6, 2005
698
Cold Exposure and Sleep in the Rat—Cerri et al
Table 2—Number and Duration of Wake and NREM Sleep Episodes During the Exposure to Low Ambient Temperature
Baseline
Light
WAKE
Episode, no.
E(10)
E(5)
E(0)
E(-10)
Episode duration, sec
E(10)
E(5)
E(0)
E(-10)
NREM sleep
Episode, no.
E(10)
E(5)
E(0)
E(-10)
Episode duration, sec
E(10)
E(5)
E(0)
E(-10)
Exposure
Recovery (day 1)
Light
Dark
Dark
Light
Dark
361 ± 29
355 ± 24
354 ± 19
362 ± 15
245 ± 18§
231 ± 22§
244 ± 13§
215 ± 29‡
405 ± 25
387 ± 27
497 ± 32†
669 ±100†
233 ± 18§
320 ± 33*
355 ± 15†§
657 ± 86†
333 ± 34
319 ± 8
305 ± 11
287 ± 17
227 ± 11§
223 ± 21‡
229 ± 10‡
208 ± 8
33 ± 2
38 ± 5
34 ± 1
32 ± 2
131 ± 11§
141 ± 17§
130 ± 9§
170 ± 28§
47 ± 4
62 ± 7
50 ± 4
44 ± 3
141 ± 14§
102 ± 11*
94 ± 6†§
47 ± 3†
35 ± 3
41 ± 4
38 ± 2
44 ± 7
136 ± 11§
138 ± 15§
124 ± 8§
134 ± 8§
360 ± 29
354 ± 24
353 ± 18
362 ± 14
245 ± 18§
230 ± 22§
244 ± 13§
214 ± 29a
404 ± 25
387 ± 27
498 ± 32†
669 ±100†
232 ± 18§
320 ± 32*
355 ± 15†§
657 ± 86†
331 ± 34
318 ± 7
304 ± 11
285 ± 17
227 ± 11§
223 ± 21‡
229 ± 10‡
208 ± 8
76 ± 8
74 ± 4
75 ± 5
74 ± 5
42 ± 3§
45 ± 4§
41 ± 2§
44 ± 5§
56 ± 6†
49 ± 5†
37 ± 3†
26 ± 4†
42 ± 3
34 ± 3
26 ± 2†
22 ± 5†
81 ± 9
76 ± 3
78 ± 2
80 ± 5
44 ± 4§
48 ± 5§
50 ± 2§
53 ± 3§
Data are presented as mean ± SEM. E(10), E(5), E(0), and E(-10) refer to ambient temperature of exposure, ie, 10°C±1°C, 5°C±1°C, 0°C±1°C, and
-10°C±1°C, respectively; NREM, non-rapid eye movement sleep; REM, rapid eye movement sleep.
*P<.05 and †P<.01 vs respective baseline condition.
‡P<.05 and (§) P<.01 vs respective light period.
Table 3—Number and Duration of REM Sleep Episodes During the Exposure to Low Ambient Temperature
SINGLE
Episode, no.
E(10)
E(5)
E(0)
E(-10)
Episode duration, sec
E(10)
E(5)
E(0)
E(-10)
SEQUENTIAL
Episode, no.
E(10)
E(5)
E(0)
E(-10)
Episode duration, sec
E(10)
E(5)
E(0)
E(-10)
CLUSTER
Episode, no.
E(10)
E(5)
E(0)
E(-10)
Episode duration, sec
E(10)
E(5)
E(0)
E(-10)
Light
Baseline
25.7 ± 2.3
27.4 ± 2.7
27.2 ± 3.1
30.3 ± 2.5
88 ±
96 ±
101 ±
87 ±
37.3 ±
28.4 ±
29.7 ±
31.8 ±
63 ±
71 ±
74 ±
63 ±
7
9
6
5
5.0
6.7
7.5
6.2
4
5
3
6
Dark
Light
12.9 ± 2.0§
12.9 ± 1.7§
11.8 ± 2.3§
13.0 ± 0.9§
87 ±
85 ±
98 ±
76 ±
9.5 ±
12.3 ±
9.8 ±
13.3 ±
80 ±
76 ±
75 ±
71 ±
8
5
7
5
2.4§
2.9§
2.6§
4.2§
9
5
6
5
Exposure
Recovery (day 1)
Dark
Dark
Light
16.3 ± 1.9†
12.2 ± 3.0†
4.8 ± 2.1†
2.8 ± 1.0†
9.8 ± 1.3
9.3 ± 1.3
6.5 ± 2.6
5.5 ± 1.3†
25.0 ± 2.3
25.0 ± 3.1
25.2 ± 3.9
26.3 ± 2.8
16.5 ± 2.6‡
18.7 ± 2.0
19.5 ± 2.0
21.7 ± 2.0†
63 ± 8
68 ± 10*
75 ± 15
34 ± 17(4)†
77 ± 18
77 ± 7
78 ± 11(5)
42 ± 16†
96 ± 11
107 ± 6
122 ± 13
113 ± 7
103 ± 12
88 ± 6
89 ± 8
94 ± 6
27.7 ± 9.2
9.8 ± 2.6†
3.3 ± 1.2†
1.0 ± 0.6†
14.0 ±
5.5 ±
4.8 ±
7.5 ±
44 ± 4
43 ± 6(5)†
37 ± 9(4)†
44 ± 36(3)
62 ± 11
77 ± 11‡
52 ± 8(5)†
35 ± 22(5)†
4.5
1.3
1.6
4.5
51.0 ± 6.7
48.3 ± 6.5†
51.3 ± 10.3†
54.8 ± 6.9†
68 ±
76 ±
78 ±
80 ±
5
5
5
4
10.7 ±
14.5 ±
25.8 ±
29.7 ±
2.1§
2.4§
3.8§
8.2§
78 ± 18
71 ± 7
79 ± 6
71 ± 6
15.1 ± 1.9
11.8 ± 2.4
12.6 ± 2.8
13.4 ± 2.3
4.3 ± 1.0§
4.8 ± 1.1§
4.0 ± 0.7§
5.5 ± 1.6§
10.3 ± 3.0
4.2 ± 1.1†
1.7 ± 0.6†
0.7 ± 0.3†
5.0 ± 1.5
2.7 ± 0.7
1.7 ± 0.6
2.7 ± 1.7
18.8 ± 2.4
19.7 ± 2.2†
19.7 ± 3.3*
21.2 ± 2.3†
4.8 ± 0.9§
5.8.± 0.9§
10.7 ± 1.4§*
11.0 ± 2.3§
153 ± 8
174 ± 7
172 ± 10
148 ± 10
173 ± 18
174 ± 10
173 ± 14
168 ± 19
111 ± 4
101 ± 13(5)†
113 ± 25(4)
93 ± 73(3)
138 ± 16
159 ± 23
92 ± 17(5)†
32 ± 5(3)†
182 ± 13
182 ± 7
202 ± 23
210 ± 12
165 ± 34
177 ± 17
185 ± 15
172 ± 18
Data are presented as mean ± SEM. E(10), E(5), E(0), and E(-10) refer to ambient temperatureof exposure, ie, 10°C±1°C, 5°C±1°C, 0°C±1°C, and
-10°C±1°C, respectively; NREM, non-rapid eye movement sleep; REM, rapid eye movement sleep.
*P<.05, †P<.01 vs respective baseline condition.
‡P<.05, §P<.01 vs respective light period..
SLEEP, Vol. 28, No. 6, 2005
699
Cold Exposure and Sleep in the Rat—Cerri et al
200
180
*
140
120
*
*
#
100
80
* *
#
#
R1 vs. BL
E1 vs. BL
60
40
i NREMS (%)
Effects of Low Ambient Temperature on EEG Power Density in the
Delta and Theta Band During Different Wake-Sleep Stages
130
120
R1
#
BL
E1
R1
R1 E1
#
#
*
110
#
*
100
90
#
R1 vs. BL
80
#
E1 vs. BL
70
130
S (%)
Figure 3 shows the results of an analysis of the effects of low
ambient temperature on the EEG power density in the frequency
bands that characterize the different wake-sleep stages in the rat,
ie, delta in NREM sleep and theta during either Wake or REM
sleep. To emphasize the general pattern of change, data relative
to the 4 different experimental groups have been pooled, and the
time course of the power density in baseline (average of BL1 and
BL2), E1, and R1 is expressed as the percentage of the respective
average 24-hour level of baseline.
On the whole, relative delta power density in NREM sleep (Figure 3, top) appears to be depressed in the E1 and to be increased
BL
E1
R1
#
160
D lt
of episodes during either the light or the dark period of the recovery.
Figure 2 shows the accumulation rate of single and sequential
REM sleep during baseline (average of BL1 and BL2), E1, and
R1, in the 4 experimental groups. Data are expressed as the percentage of the average 24-hour amount of baseline values. In E1,
the overall pattern of accumulation was similar for the 2 types of
episodes; however, in R1, there was a striking difference. As can
be clearly seen, the rate of sequential REM-sleep accumulation
was closely proportional to the ambient temperature of the exposure, and, in E(0) and E(-10), it was maintained at a high level
even in the first hours of the dark period. In contrast, the rate of
single REM-sleep accumulation during the light period was similar to that found for baseline. However, during the dark period, in
accordance with the lowering of the previous exposure temperature, the rate progressively approached those observed for either
baseline or R1 in the light period.
BL
E1
R1
120
110
100
Single REMS
250
250
250
200
200
200
200
150
150
150
150
100
100
100
100
50
50
50
50
0
0
0
0
300
300
Th
250
90
BL
E1
R1
E(-10)
E(0)
E(5)
E(10)
300
300
300
300
i
300
300
250
250
250
250
200
200
200
200
150
150
150
150
100
100
100
100
50
50
50
50
L
D
0
0
0
0
L
D
L
D
L
D
Figure 2—Relative cumulative amount of time spent in rapid eye
movement sleep (REMS, mean ± SEM, 2-hour intervals; percentage
of average 24-hour cumulative baseline levels) in the form of either
single (Single REMS) or sequential (Sequential REMS) episodes, in 4
groups of animals, during a 24-hour exposure to different low ambient
temperatures (E1, filled circles) and during the first day of the following recovery period at ambient temperature 24°C (R1, filled triangles).
Baseline levels (ambient temperature 24°C) are shown (BL, empty
circles). Each group is shown under its respective ambient temperature
of exposure (E): E(10), E(5), E(0), and E (-10).
SLEEP, Vol. 28, No. 6, 2005
80
L
D
Figure 3—Relative power density (mean ± SEM, 2-hour intervals;
percentage of average 24-hour baseline levels) in the delta band (0.754 Hz) in non-rapid eye movement sleep (NREMS, top) and the theta
band (5.5–9 Hz) in either rapid eye movement sleep (REMS, middle)
or Wake (bottom), during a 24-hour exposure to different low ambient
temperatures (E1, filled circles) and in the first day of the following
recovery period at ambient temperature 24°C (R1, black triangles).
Baseline levels (ambient temperature 24°C) are shown (BL empty
circles). For clarity, only the positive SEM for values of theta power during both REM sleep and Wake is shown. #P<.05 vs baseline;
*P<.01 vs baseline. The horizontal lines underneath the graphs show
the comparisons that were significant for the 12 hours of light or dark,
respectively, and for the whole 24-hour period (#P<.05).
BL
E1
R1
E(-10)
E(0)
E(5)
#
70
Sequential REMS
E(10)
#
in R1. The consistent depression due to cold exposure was statistically significant in the dark period of E1 (P<.05 with respect
to the baseline). In contrast, a large and statistically significant
increase (P<.05 with respect to baseline) was observed during the
light period of the R1, whereas no changes were observed in the
following dark period.
Relative theta power density decreased progressively from the
early light period to the late dark period in both REM sleep and
Wake (Figure 3, middle and bottom diagrams, respectively), al700
Cold Exposure and Sleep in the Rat—Cerri et al
though, for both stages, this oscillation was less evident than that
observed for delta power in NREM sleep. However, a consistent
difference in the response to the environmental challenge was observed between REM sleep and Wake. As can be observed, the
relative theta power density in REM sleep was increased above
baseline levels in E1 and R1. This increase was very large and
statistically significant for the whole dark period with respect to
baseline (P<.05), whereas R1 was characterized by a statistically
significant (P<.05) increase that remained above baseline values
for the whole 24-hour period. In contrast, the relative theta power
in Wake was depressed, with respect to baseline values, during the
early period of the experiment (P<.05, for 2 consecutive 2-hour
periods) and increased, although not significantly, in the following dark period, but no significant change was observed in R1.
Since delta power is important for the quantitative characterization of NREM sleep, the individual effects of each low ambient temperature across the light-dark cycle were analyzed and
are shown in Figure 4. Delta power is expressed as the percentage of the respective average 24-hour level of baseline for each
temperature of exposure. The results show that both the decrease
in E1 and the increase in R1, which are shown in Figure 3, are
dependent on the level of the thermal load, and, in particular, a
substantial increase characterizes the light period of R1 for both
E(0) and E(-10).
treme environmental challenge that greatly perturbs wake-sleep
parameters. In particular, during the exposure, the circadian differences in terms of number and duration of episodes that differentiate the wake-sleep stages disappear during the exposure,
and both the light and dark period are characterized by rapid sequences of Wake and NREM-sleep episodes that were only interrupted by a REM-sleep episode approximately once in every 100
sequences. In Figure 5, an example from 2 selected recordings
during either baseline (ambient temperature, 24°C; top) or E(-10)
(ambient temperature -10°C, bottom) is shown (rat #216). The
histograms represent the time course, within a 30-minute period
(11:55 PM-12:25 AM), of the power density (delta, theta, and sigma bands), and motor activity, determined in 4-second epochs. A
comparison between the 2 experimental conditions emphasizes
the extreme fragmentation of the wake-sleep cycle at ambient
temperature -10°C, which is characterized by the occurrence of
low delta, low theta, and high sigma power density NREM-sleep
episodes interrupted by brief awakenings characterized by a rise
in motor activity.
In Figure 6, the time course of the relative delta power density
in NREM sleep and the relative theta power density in REM sleep
across the whole experimental period for E(-10) is shown. When
compared with baseline, the results clearly show that relative delta
power was greatly decreased in E1, increased to maximum value
in the light period of R1, and decreased during the light period of
R2. From the dark period of R2 onward, it kept close to baseline
levels. On the contrary, the theta power density was maintained
above baseline during both the dark period of E1 and R1.
Further Analysis of the Effects of Ambient Temperature -10°C on
Wake-Sleep Parameters
The previous analyses of the results have shown that the 24hour exposure to ambient temperature -10°C represents an ex220
220
BL
E1(10°C)
R1
200
180
BL
E1(5°C)
R1
200
180
160
160
140
140
120
120
DPW
400
%
200
TPW
600
%
300
SPW
600
%
300
MA
800
%
400
Ta, 24°C
0
0
D l
100
100
80
80
60
60
i NREMS (%)
40
40
220
220
BL
E1(0°C)
R1
200
180
160
*
*
0
#
0
23.55h
200
*
180
BL
E1(-10°C)
R1
400
%
200
TPW
600
%
300
SPW
600
%
300
MA
800
%
400
160
140
140
120
120
100
100
80
80
60
60
40
40
Ta, -10°C
0
0
#
0
L
D
L
D
Figure 4—Relative power density (mean ± SEM, 2-hour intervals;
percentage of average 24-hour baseline levels) in the delta band (0.754 Hz) in non-rapid eye movement sleep (NREMS) for 4 groups of
animals during a 24-hour period of exposure to different low ambient
temperatures (E1, filled circles) and during the first day of the following recovery period at ambient temperature 24°C (R1, black triangles).
Baseline levels (ambient temperature 24°C) are also shown (BL, empty
circles). Each group is shown under its respective ambient temperature
of exposure: E(10), E(5), E(0), and E(-10).
SLEEP, Vol. 28, No. 6, 2005
DPW
00.25h
0
23.55h
00.25h
Figure 5—Time course of relative (percentage of average 24-hour
baseline levels) (DPW, 0.75-4 Hz), theta (TPW, 5.5–9 Hz) and sigma
(SPW, 11–16 Hz) power densities and motor activity (MA), in 1 animal (rat #216) during baseline recording (ambient temperature [Ta]
24°C, upper diagram) and during the exposure to ambient temperature
-10°C (lower diagram). Figures shows 30 minutes of recording. Black
bars under the histogram relative to TPW indicate rapid eye movement
sleep episodes.
701
Cold Exposure and Sleep in the Rat—Cerri et al
250
the light period, than in the following dark period during which,
for example, the amount of Wake never significantly exceeds that
of normal circadian variation.
The overall increase in Wake during the exposure to low ambient temperature may be considered, from both an autonomic
and a behavioral point of view, as a means to improve thermoregulation. The increase in sympathetic activity that occurs during an acute exposure to low ambient temperature counteracts the
decrease in tonic sympathetic activity and the increase in tonic
parasympathetic activity that normally characterizes NREM sleep
and, thus, depresses its occurrence.32 Moreover, the thermoregulatory impairment that characterizes REM sleep5,33 can be considered as the basis for the strong and progressive inhibition of its
occurrence, which is in direct proportion to the thermal load. In
a laboratory condition in which burrowing, huddling, or migrating is impossible,34 the inhibition of REM-sleep occurrence may
be considered as an alternative form of behavioral thermoregulation.
Although the reduction in NREM-sleep occurrence caused
by the lowering of ambient temperature was mainly present during the light period, it should be noted that the wake-sleep cycle
was fragmented during both the light and dark period and that
this was more pronounced as the ambient temperature was lowered. During the dark period, this fragmentation was concomitant
with a decrease in the average duration of both Wake and NREM
sleep episodes that was greatest during the exposure to ambient
temperature -10°C. A further analysis of the pattern of the wakesleep cycle at such an extreme ambient temperature has shown
that consecutive brief NREM-sleep episodes were characterized
by a rise in sigma power density, which, in the rat, typically signals the transition period from NREM sleep to REM sleep.35-37
In rats kept at normal laboratory conditions, only 25% to 30% of
the spontaneous transitions from NREM sleep to REM sleep are
followed by a consolidated REM sleep episode, while this percentage rises to nearly 100% during the early recovery period that
follows a 24-hour period of total sleep deprivation.27 Moreover,
in a study in which rats were selectively deprived of REM sleep
by being forced awake during each transition period from NREM
sleep to REM sleep, the spontaneous rate of transition periods
from NREM sleep to REM sleep was shown to be an index of
the increase in pressure for REM sleep.38 Thus, the succession
of apparent transition periods from NREM sleep to REM sleep
observed in our experiment that are not followed by a REM-sleep
episode would suggest not only that the pressure for REM sleep
was strongly increased during the exposure to ambient temperature -10°C, but also that the wake-sleep cycle was apparently
changed into a sequence of repetitive attempts to start a REMsleep episode.
With respect to the recovery at normal laboratory ambient
temperature following each thermal load, only REM-sleep occurrence is significantly affected, and this is in agreement with previous studies on short-term cold exposure.1,2,14,17,18 The increase in
the amount of REM sleep during both the light and dark period of
the first day of recovery was proportional to the previous thermal
load and occurred at the expense of both NREM sleep and Wake.
However, although REM sleep occurrence was highest during
the first day of recovery it was still tonically increased above the
baseline levels during the subsequent recovery period (days 2-4)
and, thus, independent from the previously applied thermal load.
A similar effect on REM-sleep occurrence 2 to 4 days into the
Ta, -10°C
200
150
100
50
180
Ta, -10°C
140
100
60
BL
E1
R1
R2
R3
R4
Figure 6—Time course of the relative power density (mean ± SEM,
2-hour intervals; percentage of average 24-hour baseline levels) in the
delta band (0.75-4 Hz) in non-rapid eye movement sleep and in the
theta band (5.5–9 Hz) in rapid eye movement sleep during 2 days of
baseline at ambient temperature (Ta) 24°C (BL), 1 day of exposure to
ambient temperature -10°C (E1), and 4 days of recovery at ambient
temperature 24°C (R1-R4).
DISCUSSION
The results of our study confirm that the exposure to low ambient temperature greatly influences the wake-sleep cycle.3,7 However, this exposure induces relevant quantitative changes not only
in the time spent in the different wake-sleep stages, but also in the
EEG power density in specific frequency bands, albeit that both
of these are weighted differently in NREM sleep and REM sleep.
The Paradigm of Duration: Influences on the Time Spent in Each
Wake-Sleep Stage and on Their Occurrence
Our results are in agreement with others on the rat and other
species that show that Wake is enhanced during the exposure to
low ambient temperature, while the amount of REM sleep is depressed to a larger extent than NREM sleep.1,2,8-13,15 Moreover, as
initially observed in short-term exposure studies carried out in the
cat,1,2,7 the effect of cold on the amount of REM sleep during a
24-hour exposure is proportional to the thermal load, but this is
not the case for NREM sleep. With respect to this, a slight, but not
significant, increase in the amount of NREM sleep was observed
in the dark period of the lowest ambient temperature and is in
keeping with a similar finding in the cat that is able to reattain
control levels of slow-wave sleep during the exposure to extreme
low ambient temperatures.1,2,7 Furthermore, the results clearly indicate that the effects of the thermal load on all wake-sleep stages
are more pronounced during the first 12 hours of exposure, ie, in
SLEEP, Vol. 28, No. 6, 2005
702
Cold Exposure and Sleep in the Rat—Cerri et al
recovery following total sleep deprivation has been previously
described in the rat.39
The dynamics of REM sleep during both the exposure to low
ambient temperature and the following recovery period confirm
that, under such an environmental challenge, REM-sleep occurrence in the rat is mainly modulated by the frequency of episodes
rather than by their duration.17-19,40 Significantly, a large increase
in the number of sequential REM-sleep episodes, occurring in
clusters, was observed during both the light and dark period of
the recovery. Thus, under the pressure of a high and unrestrained
drive for REM sleep, the occurrence of episodes in a rapid sequence would be the main behavioral strategy and allows the need
for REM sleep to be satisfied without increasing the duration of
the episode.17-19 As observed in previous studies,17-19 the amount of
REM sleep in the form of single episodes was not increased, with
respect to baseline, during the light period of the first day of the
recovery, but an increase was found during the subsequent dark
period. Moreover, its relative accumulation rate did not overcome
that observed during the light period of baseline, even following
the lowest temperature of exposure E(-10) in which the REMsleep rebound was at its maximum (R1). These data suggest the
presence of a preset limit for the rate of single REM sleep accumulation.
band has been noted in NREM sleep,45 in the rat, hippocampal
theta rhythm is known as an EEG marker for both REM sleep and
behaviorally active waking.46
With respect to REM sleep, theta power has been shown to increase after sleep deprivation, and, although it has been proposed
as an intensity index for REM sleep,47,48 as yet no quantitative
model concerning REM-sleep intensity has been developed. If
the increase in theta power is considered as a sign of REM-sleep
intensity, the increase observed in the few episodes of REM sleep
occurring during the exposure to low ambient temperature would
suggest that REM-sleep intensity is high in the cold, while the
amount of REM sleep is low. Thus, as observed for the amount
of time spent in each sleep stage, REM sleep appears to be regulated differently with respect to NREM sleep. While REM sleep
is regulated mainly by its duration, the intensity with which it
occurs, even in a condition in which it is being actively inhibited
such as the cold, shows that a REM-sleep debt is accumulating
and that the animal tries to compensate for this at any given opportunity. This would imply that once REM sleep is started, it is
a compulsory behavior, and ambient conditions can only impose
a tight control on when an episode can begin. The increase in
theta power could not be due to the effects of cooling nervous
structures, since it has been clearly shown that when the brain is
cooled, as for example, in torpor,49 the peak in theta power shifts
to the lower frequency defining its band and, thus, would result in
a reduced theta power density.
When Wake is considered, it is likely that a decrease in theta
power reflects a situation in which the exploratory behavioral
activity of the animal is also decreased. Thus, during cold exposure, the reduced level of theta power during the light period is
in accordance to the exploratory behavior of the animal that was
observed to be notably absent during the initial exposure.
The Paradigm of EEG Power: Influences on the Power Density in
the Different Frequency Bands During the Wake-Sleep Stages
Many sleep studies in which the EEG has been quantitatively
analyzed in terms of power-spectra have shown that the power
density in the delta band (0.75-4.0 Hz) progressively declines during NREM sleep in normal conditions, whereas it is increased after
total sleep deprivation to an extent that is quantitatively related to
the duration of the previous deprivation.25 Thus, although the biologic meaning of the synchronization of the EEG during NREM
sleep is still unknown, delta power is considered to be an indicator of NREM-sleep intensity.25 With respect to this, delta power
has been shown to increase in conditions in which the pressure
for NREM sleep would be expected to be intense, such as when
rats have been previously subjected to stress, thereby enhancing
wakefulness,41 or following either an intracerebroventricular or
intraperitoneal administration of an adenosine A1 agonist.42,43
The results of our study show that, during the exposure to low
ambient temperature, delta power is depressed to an extent that
ranges from values close to those observed in baseline (E10) to
values very much lower than these (E-10). Such a decrease would
indicate that, in the rat, cold exposure affects NREM-sleep intensity much more than its duration. It would appear that the breakdown of NREM sleep into fragmented episodes, which is highest
at the lowest ambient temperatures, prevents the progressive increase of delta power that normally occurs during NREM sleep.
Although it has been shown that in different species a decrease
in brain temperature in the physiologic range induces both a frequency shift and a change in power density,44 the small decrease
in average values of hypothalamic temperature (about 0.5°C)
observed in the animals kept at the lowest ambient temperatures
would seem unlikely to explain the concomitant decrease in delta
power.
During both REM sleep and, to a lesser extent, Wake, significant changes in power density were found in the theta band. Although, a significant presence of EEG power density in the theta
SLEEP, Vol. 28, No. 6, 2005
CONCLUSIONS
Our study reinforces the use of cold exposure as a tool for sleep
deprivation,1,2,7 since this procedure not only allows the intensity
of deprivation to be modulated by changing the ambient temperature, but also allows sleep expression to be studied while deprivation is occurring.
The results of this study show that the impairment of both
NREM sleep and REM sleep occurrence is dependent on the
degree of thermoregulatory activation. However, while NREM
sleep is affected more in its “intensity,” as shown by the decrease
in delta power that is most likely due to the extreme shortening
of episodes, the effects on REM sleep are reflected in the amount
of REM sleep that the animal is allowed to make. In the recovery, the main changes characterizing the sleep stages occur during
the first day and appear as a mirror image of those occurring at
low ambient temperature, since the delta power increases in both
NREM-sleep and REM-sleep amount in proportion to the thermal
load.
Moreover, we have found that theta power during REM sleep
is high in the few episodes that occur in the cold and throughout
the first day of recovery. This suggests the existence of a process
that regulates the “intensity” of REM sleep that is also operant
during deprivation and might possibly reflect an effort to minimize the debt while it occurs.
703
Cold Exposure and Sleep in the Rat—Cerri et al
21. Feinberg I. Delta homeostasis, stress, and sleep deprivation in the
rat:a comment on Rechtschaffen et al. Sleep 1999;22:1021-4.
22. Rechtschaffen A, Bergmann BM. Sleep stage priorities in
rebounds from sleep deprivation:a response to Feinberg. Sleep
1999;22:1025-30.
23. Rechtschaffen A, Bergmann B.M. Sleep rebounds and their implications for sleep stage substrates:a response to Benington and
Heller. Sleep 1999;22:1038-43.
24. Rechtschaffen A, Bergmann BM, Gilliand MA, Bauer K. Effects
of method, duration, and sleep stage on rebounds from sleep deprivation in the rat. Sleep 1999;22:11-31.
25. Borbély AA, Achermann P. Sleep homeostasis and models of sleep
regulation. In: Kryger MH, Roth T, Dement WE, eds. Principles
and Practice of Sleep Medicine. Philadelphia: WB Saunders;
2000:377-90.
26. Horne JA. REM sleep-by default? Neurosci Biobehav Rev
2000;24:777-97.
27. Franken P. Long-term vs. short-term processes regulating REM
sleep. J Sleep Res 2002;11:17-28.
28. Amici R, Cerri M, Domeniconi R, et al. Dynamics of REM sleep
rebound in the rat following short term deprivation at low ambient
temperature. J Sleep Res 2000;9(suppl1):3.
29. Amici R, Cerri M, Jones CA, et al. Sleep regulation in the
rat exposed to changes in ambient temperature. J Sleep Res
2002;11(suppl1):4.
30. Parmeggiani PL. Brain cooling across wake-sleep behavioural
states in homeotermic species: an analysis of the underlying physiological mechanisms. Rev Neurosci 1995;6:352-63.
31. Wallenstein S, Zucker CL, Fleiss JL. Some statistical methods useful in circulation research. Circ Res 1980;47:1-9.
32. Parmeggiani PL. The autonomic nervous system in sleep. In:
Kryger MH, Roth T, Dement WE, eds. Principles and Practice of
Sleep Medicine. Philadelphia: WB Saunders; 1994:194-203.
33. Parmeggiani PL. Physiological regulation in sleep. In: Kryger MH,
Roth T, Dement WE, eds. Principles and Practice of Sleep Medicine. Philadelphia: WB Saunders; 2000:169-78.
34. Satinoff E. Behavioral thermoregulation in the cold. In: Fregly MJ,
Blatteis CM, eds. Handbook of Physiology, Section 4: environmental physiology, Vol. 1. New York: Oxford University Press;
1996:481-505.
35. Trachsel L, Tobler I, Borbély A.A. Electroencephalogram analysis
of non-rapid eye movement sleep in rats. Am J Physiol 1988;24:
R27-37.
36. Benington JH, Kodali SK, Heller HC. Scoring transitions to
REM sleep based on the EEG phenomena of pre-REM sleep: an
improved analysis of sleep structure. Sleep 1994;17:28-36.
37. Gottesman C. The transition from slow-wave sleep to paradoxical sleep: evolving facts and concepts of the neurophysiological
processes underlying the intermediate stage of sleep. Neurosci
Biobehav Rev 1996;20:367-87.
38. Ocampo-Garcés A, Vivaldi EA. Short-term homeostasis of REM
sleep assessed in an intermittent REM sleep deprivation protocol
in the rat. J Sleep Res 2002;11:81-9.
39. Schwierin B, Borbély AA, Tobler I. Prolonged effects of 24-h total
sleep deprivation on sleep and sleep EEG in the rat. Neurosci Lett
1999;261:61-4.
40. Zamboni G, Perez E, Amici R, Jones CA, Parmeggiani PL. Control of REM sleep: an aspect of the regulation of physiological
homeostasis. Arch Ital Biol 1999;137:249-62.
41. Meerlo P, Pragt B, Daan S. Social stress induces high intensity
sleep in rats. Neurosci Lett 1997;225:41-4.
42. Benington JH, Kodali SK, Heller HC. Stimulation of A1 adenosine receptors mimics the electroencephalographic effects of sleep
deprivation. Brain Res 1995;692:79-85.
43. Schwierin B, Borbély AA, Tobler I. Effects of N6-cyclopentyladenosine and caffeine on sleep regulation in the rat. Eur J Pharmacol 1996;300:163-71.
ACKNOWLEDGEMENTS
The authors would like to thank: Mr. G. Mancinelli and Mr. L.
Sabattini for the wiring and the mechanical work needed for the
adaptation of both the recording apparatus and room.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Parmeggiani PL, Rabini C, Cattalani M. Sleep phases at low environmental temperature. Arch Sci Biol 1969;53:277-90.
Parmeggiani PL, Rabini C. Sleep and environmental temperature.
Arch Ital Biol 1970;108:369-87.
Glotzbach SF, Heller HC. Temperature regulation. In: Kryger MH,
Roth T, Dement WE, eds. Principles and Practice of Sleep Medicine. Philadelphia: WB Saunders; 2000:289-304.
Parmeggiani PL. Temperature regulation during sleep: a study in
homeostasis. In: Orem J, Barnes CD, eds. Physiology in Sleep.
New York: Academic Press; 1980:97-143.
Parmeggiani PL, Rabini C. Shivering and panting during sleep.
Brain Res 1967;6:789-91.
Parmeggiani PL. Interaction between sleep and thermoregulation:
an aspect of the control of behavioral states. Sleep 1987;10:426-35.
Parmeggiani PL. Interaction between sleep and thermoregulation.
Waking Sleeping 1977;1:123-32.
Schmidek WR, Hoshino K, Schmidek M, Timo-Iaria C. Influence
of environmental temperature on the sleep-wakefulness cycle in
the rat. Physiol Behav 1972;8:363-71.
Buguet AGC, Roussel BHE, Watson WJ, Radomski MW. Cold-induced diminution of paradoxical sleep in man. Electroencephalogr
Clin Neurophysiol 1979;46:29-32.
Sakaguchi S, Glotzbach SF, Heller HC. Influence of hypothalamic
and ambient temperatures on sleep in kangaroo rats. Am J Physiol
1979;237:R80-8.
Roussel B, Turrillot P, Kitahama K. Effect of ambient temperature on the sleep-waking cycle in two strains of mice. Brain Res
1984;294:67-73.
Sichieri R, Schmidek WR. Influence of ambient temperature on
the sleep-wakefulness cycle in the golden hamster. Physiol Behav
1984;33:871-7.
Alföldi P, Rubicsek G, Cserni G, Obál F. Jr. Brain and core
temperatures and peripheral vasomotion during sleep and wakefulness at various ambient temperatures in the rat. Pflügers Arch
1990;417:336-41.
Franken P, Tobler I, Borbély AA. Effects of 12-h sleep deprivation and of 12-h cold exposure on sleep regulation and cortical
temperature in the rat. Physiol Behav 1993;54:885-94.
Thannickal CT, Mohan Kumar V. Effect of ambient temperature
on sleep-wakefulness in normal and medial preoptic area lesioned
rats. Sleep Res Online 2000;3:141-5.
Parmeggiani PL, Cianci T, Calasso M, Zamboni G, Perez E.
Quantitative analysis of short term deprivation and recovery of desynchronized sleep in cats. Electroencephalogr Clin Neurophysiol
1980;50:293-302.
Amici R, Zamboni G, Perez E, et al. Pattern of desynchronized
sleep during deprivation and recovery induced in the rat by
changes in ambient temperature. J Sleep Res 1994;3:250-6.
Amici R, Zamboni G, Perez E, Jones CA, Parmeggiani PL. The
influence of a heavy thermal load on REM sleep in the rat. Brain
Res 1998;781:252-8.
Zamboni G, Amici R, Perez E, Jones CA, Parmeggiani PL. Pattern
of REM sleep occurrence in continuous darkness following the
exposure to low ambient temperature in the rat. Behav Brain Res
2001;122:25-32.
Benington JH, Heller HC. Implications of sleep deprivation
experiments for our understanding of sleep homeostasis. Sleep
1999;22:1033-7.
SLEEP, Vol. 28, No. 6, 2005
704
Cold Exposure and Sleep in the Rat—Cerri et al
44. Deboer T. Brain temperature dependent changes in the electroencephalogram power spectrum of humans and animals. J Sleep Res
1998;7:254-62.
45. Gaztelu JM, Romero-Vives M, Abraira V, Garcia-Austt E. Hippocampal EEG theta power density is similar during slow-wave sleep
and paradoxical sleep. A long-term study in rats. Neurosci Lett
1994;172:31-4.
46. Bland BH. The physiology and pharmacology of hippocampal
formation theta rhythms. Prog Neurobiol 1986;26:1-54.
47. Borbély AA, Tobler I, Hanagasioglu M. Effect of sleep deprivation on sleep and EEG power spectra in the rat. Behav Brain Res
1984;14:171-82.
48. Franken P, Dijk DJ, Tobler I, Borbély A.A. Sleep deprivation in
rats: effects on EEG power spectra, vigilance states, and cortical
temperature. Am J Physiol 1991;261:R198-208.
49. Deboer T. Electroencephalogram theta frequency changes in parallel with euthermic brain temperature. Brain Res 2002;930:212-5.
SLEEP, Vol. 28, No. 6, 2005
705
Cold Exposure and Sleep in the Rat—Cerri et al