SLEEP, SLEEP DEPRIVATION, AND PERFORMANCE

Transcription

1 30722 Sleep 5.qxd 6/27/2003 2:16 PM Page 567 SLEEP, SLEEP DEPRIVATION, AND PERFORMANCE Sleep Regulation in the Djungarian Hamster: Comparison of the Dynamics Leading to the Slow-Wave Activity Increase After Sleep Deprivation and Daily Torpor Tom Deboer PhD 1,2 and Irene Tobler PhD 2 1 Department of Neurophysiology, LUMC, Leiden, The Netherlands; 2 Institute of Pharmacology and Toxicology, University of Zürich, Zürich, Switzerland. Study Objectives: Emerging from daily torpor, Djungarian hamsters (Phodopus sungorus) show an initial increase in electroencephalographic slow-wave activity (power density between 0.75 and 4.0 Hz) during sleep that gradually declines. This feature is typical for sleep following prolonged waking and supports the hypothesis that sleep pressure increases during daily torpor. After hamsters were subjected to sleep deprivation or partial non-rapid eye movement sleep deprivation immediately following torpor, slow-wave activity remained high and decreased only when sleep was allowed. An analysis of the dynamics of the process underlying the build-up of sleep pressure during episodes of waking and torpor may provide insights into the regulation of normal sleep and wakefulness. We have analyzed in more detail the timecourse of the process that is common for waking and daily torpor and that could account for the subsequent increase in slow-wave activity. Design: Continuous 24-hour recordings of electroencephalography, electromyography, cortical temperature, and electroencephalographic spectral analysis were performed. Torpor data of 28 hamsters and sleep-deprivation data of diverse durations collected previously in 15 hamsters were analyzed. Setting: N/A Patients or Participants: N/A Interventions: Sleep deprivation. Measurements and Results: Slow-wave activity invariably increased as a function of the duration of both prior waking and torpor. However, the time constant of the build-up of slow-wave activity was approximately 2.75 times slower during torpor compared to sleep deprivation. Brain temperature recorded during the torpor bouts was 10º to 12ºC below euthermic brain temperature. Therefore, the temperature coefficient of the time constant for the slow wave-activity increase is between 2.3 and 2.8, a range typical for biochemical processes. Conclusions: We conclude that the processes occurring during daily torpor in the Djungarian hamster are similar to those occurring during sleep deprivation, but the build-up of sleep pressure during torpor appears to be slowed down by the lower brain temperature. Key Words: Daily torpor, Djungarian hamster, EEG spectral analysis, hibernation, sleep regulation Citation: Deboer T; Tobler I. Sleep Regulation in the Djungarian Hamster: comparison of the dynamics leading to the slow-wave activity increase after sleep deprivation and daily torpor. SLEEP 2003;26(5): INTRODUCTION UNTIL A FEW YEARS AGO, DAILY TORPOR AND HIBERNATION WERE THOUGHT TO HAVE EVOLVED FROM SLEEP BECAUSE, BEHAVIORALLY AND PHYSIOLOGICALLY SLEEP, DAILY TOR- POR, AND HIBERNATION SEEM TO FORM A CONTINUUM. 1,2 However, this view was challenged by the finding that hibernating ground squirrels after emerging from torpor 3-5 and Djungarian hamsters after daily torpor 6 sleep as if they had been sleep deprived. Invariably, the animals sleep and electroencephalographic (EEG) slow-wave activity (SWA) (mean EEG power density between Hz) in non-rapid eye movement (NREM) sleep is initially high and thereafter gradually declines. Both features are typical for recovery after sleep deprivation (SD). 7 A comparison of the dynamics of the processes underlying the build-up of sleep pressure during waking and torpor may provide new insights into the regulation and function of normal sleep and wakefulness. Disclosure Statement This research was supported by the Swiss National Science Foundation Grant and and now Grant Submitted for publication November 2002 Accepted for publication March 2003 Address correspondence to: Dr. T. de Boer, Department of Neurophysiology, Leiden University Medical Centre, PO Box 9604, 2300 RC Leiden, The Netherlands; Tel: ; Fax: ; Tom.de_Boer@lumc.nl SLEEP, Vol. 26, No. 5, Slow-wave activity in NREM sleep is an indicator of NREM sleep intensity and is thought to be an electrophysiologic correlate of a sleep homeostatic process S that reflects the prior history of sleep and wakefulness. 8,9 Studies in many species have shown enhanced levels of SWA after SD. 7 Moreover, a relationship between the changes in SWA and the duration of prior waking has been established in several mammals (rat 10 ; human 11 ; cat 12 ; ground squirrel 13 ; mouse 14 ). Interestingly, the initial value of SWA is positively correlated with torpor bout duration in the Djungarian hamster 15 and in hibernating ground squirrels, 4 the 2 species that were investigated. In both species, this relationship is similar to the one observed after different durations of enforced waking. Moreover, the hibernators spend at least 60% of the euthermic periods occurring between hibernation bouts in sleep. 3,5 The data indicate that the increase in need for sleep and the enhanced sleep intensity immediately following the emergence from torpor is a consequence of a SD incurred during the torpid state. Studies in Djungarian hamsters have shown that sleep is necessary to dissipate the initial increase in SWA. 16 When the animals are sleep deprived immediately after emerging from the torpid state, SWA is still enhanced and it is significantly higher compared to a control SD of the same duration. 16 Recently this finding has been confirmed by a partial NREM sleep deprivation experiment in the Djungarian hamster. 17 A similar experiment in ground squirrels did not lead to an additional increase in SWA, and the level of SWA did not differ from SWA after a control SD of the same duration. 13 It has been proposed that the prolonged electrical silence in the brain during its very low temperatures (<15 C) during hibernation might reduce synaptic connections 13 or other structures, 18 which would cause

2 30722 Sleep 5.qxd 6/27/2003 2:16 PM Page 568 the cortex to start oscillating in a slow-wave mode after reaching euthermia. This interpretation does not apply to the findings in Djungarian hamsters, since their minimum brain temperature does not drop below 20 C during daily torpor. 6 The aim of the present study was to compare the dynamics of the SWA build-up during SD and torpor, allowing the comparison of the hypothetical time constants of the increase of the homeostatic process S during the 2 states. Torpor data obtained previously 6,15,16 were categorized according to the duration of the torpor bouts, and the increase of SWA was compared with its level after 1.5 hours and 4 hours of SD. 16,19 of the baseline light period. Data Acquisition and Analysis The EEG and EMG signals were amplified (amplification factor approximately 2000), conditioned by analog filters (high-pass filter: - 3dB at Hz; low-pass filter: -3dB at 40 Hz; less than 35 db at 128 Hz), sampled with 256 or 512 Hz, digitally filtered (EEG: low-pass FIR METHODS The present work was performed after approval of the governmental institution for animal experimentation. Animals The environmental living conditions and surgical procedures have been described previously. 6,15,16,19-21 In short, the experiments were performed in adult male Djungarian hamsters (Phodopus sungorus) raised under natural photoperiod. They were kept in Macrolon cages in an individual isolated chamber at approximately 16 C ambient temperature (T a ) and a short photoperiod (L:D 8:16 hours). Food and water were available ad libitum, and once every week fresh apple slices were provided. Animals were selected for sleep recordings when episodes of daily torpor were recognized and when the change in weight and fur color indicated a strong adaptation to the short photoperiod. At the age of 5 to 6 months, the animals were implanted with electrodes under deep anesthesia. Two gold-plated miniature screws served as EEG electrodes and were screwed through the skull onto the dura over the right parietal cortex and the cerebellum. To record the electromyogram (EMG), two gold wires were inserted into the neck muscle tissue. To measure cortical temperature (T CRT ), a thermistor was inserted between the skull and the dura through a hole over the left cortex. 15,20 Experimental Protocol Torpor episodes obtained in 28 hamsters were used. In brief, EEG, EMG, and T CRT were recorded continuously until torpor occurred. 6,15,16 A day on which no torpor occurred served as baseline within an individual hamster. For the 4-hour SD data, a baseline day was followed by 4 hours of SD, starting at lights on (N=7). 19 For the 1.5-hour SD, the animals were submitted to SD at approximately the same time of day when torpor had previously terminated spontaneously (N=8). 16 Sleep deprivation was attained by tapping on the cage and by introducing objects (e.g., nesting material) into the cage whenever the animal appeared drowsy or assumed a sleeping position. Since sleep and waking and changes in EEG power density in the slow-wave range are distributed evenly over the day in animals well adapted to a short photoperiod and 16 C ambient temperature, 6,19,21 minor differences in the phase angle between offset of torpor and SD were not expected to influence the results in a significant way. All spectral data were expressed relative to the mean value Table 1 Mean torpor variables (SEM) Torpor duration group Short Medium Long N Duration (hours) * 2.1 (0.3) 4.7 (0.3) 7.3 (0.4) T CRT, C 24.8 (0.3) 22.5 (0.4) 20.8 (0.6) T CRT min, C 23.2 (0.2) 20.6 (0.6) 19.1 (0.6) Time of entrance (hours after lights on) 2.5 (0.1) 1.4 (0.3) 0.5 (0.3) Time of emergence * (hours after lights on) 4.6 (0.4) 6.1 (0.3) 7.7 (0.5) Cortical temperature (T CRT ) below 27ºC is defined as torpor. *, duration and time of emergence differ significantly between the 3 groups (P<0.05, 2-tailed t-test after significant ANOVA). SLEEP, Vol. 26, No. 5, Figure 1 Timecourse of slow-wave activity (SWA) (mean EEG power density between Hz) in non-rapid eye movement (NREM) sleep, the amount of NREM sleep and rapid eye movement (REM) sleep, and cortical temperature after 3 categories of torpor duration (short: <3 hours, n=6; medium: 3-6 hours n=15; long: > 6 hours, N=7). Curves connect 30-minute mean values for the first 150 minutes after emergence from torpor (0 min corresponds to T CRT = 27 C, the end of the individual torpor bouts). SWA is expressed relative to the 24-hour value of the baseline day (=100%), and NREM sleep and REM sleep as percentage of recording time. The asterisks indicate significant differences between the shorttorpor group and the other 2 groups in the first 30-minute interval (P<0.005, 2-tailed t-test after significant ANOVA for factor group ). The open circles indicate a significant difference in SWA between the short-torpor group and the long-torpor group in the second 30- minute interval (P<0.05, 2-tailed t-test after significant ANOVA for factor duration ).

3 30722 Sleep 5.qxd 6/27/2003 2:16 PM Page 569 filter 25 Hz; EMG: band-pass FIR filter Hz), and stored with a resolution of 128 Hz. The EEG power spectra were computed for consecutive 4-second epochs by Fast Fourier Transform routine within the frequency range of 0.25 to 25.0 Hz. Between 0.25 and 5.0 Hz, the values were collapsed into 0.5-Hz bins, and between 5.25 and 25.0 Hz, into 1- Hz bins. The EMG signals were integrated over 4 seconds, and T CRT and T a inside the cage were recorded at 4-second intervals. All data were recorded simultaneously and stored on optical disc. Before the start of each recording, the EEG and EMG channels were calibrated with a 10- Hz sine wave, 300µV peak-to-peak signal. After the experiment, the vigilance states were determined for 4-second epochs by visual scoring. 20 Epochs containing EEG artifacts were omitted from further analysis of the power spectra (<4 % of total recording time, N=28), but vigilance states could always be determined. The 28 torpors were divided into 3 categories based on torpor-bout duration (short: torpor<3 hours, N=6; medium: torpor 3-6 hours, N=15; long: torpor>6 hours, N=7), enabling a statistical analysis between the groups. The shorter category is based on a previous analysis indicating that torpors lasting less than 3 hours are followed by lower SWA levels than after the longer torpor bouts, 15 whereas the other 2 groups might produce similar results. The data for torpor duration, minimum T CRT reached, and time of day of entrance and emergence from torpor are listed in Table 1. Animals were considered in torpor when T CRT decreased below 27 C. This definition of torpor is based on earlier observations that showed that slow waves in the NREM sleep EEG become completely suppressed below 27 C. 15 Analysis of variance (ANOVA) served to determine overall differences between the 3 torpor categories and between the 2 SD-duration groups. Whenever significant effects were present (P<0.05), further differences were evaluated by 2-tailed t-tests. In addition, for some computations, torpor duration was considered to be a continuous variable, and regression analysis was performed. The time constants for the increase of process S during waking and torpor were estimated by fitting the function S t =A-(1-S 0 ) e -t/ti through the baseline mean value (mean over first 30 minutes after lights on), the initial 30-minute mean value after SD (=highest value), or the highest 30-minute mean value attained after torpor (t = duration of SD or torpor; S 0 = level of S before the start of SD or torpor; S t = level of process S after SD or torpor duration t; A = upper asymptote of S; T i = time constant of increase) 22,14. The S 0 and the upper asymptote A were optimized by calculating the time constants after SD, and the same values were applied on the calculation of the time constant after torpor. RESULTS Torpor All 3 torpor categories reached the maximum SWA value during recovery only after a delay of at least 30 minutes (Figure 1A). Moreover, the short-torpor group exhibited the lowest value in SWA in the first 30 minutes compared to the other 2 groups. The second post-torpor 30-minutes interval still reflected a dose-response relationship between torporbout duration and the level of SWA increase. Thereafter, no differences remained between the torpor categories. For the medium- and long-torpor group, SWA was above baseline in the first four 30-minutes intervals. For the short-torpor group, only the second and third intervals were above baseline (P<0.05, 2-tailed paired t-test after significant ANOVA). The amount of NREM sleep was similar after all torpors (Figure 1B), while the amount of REM sleep was significantly larger in the short-torpor group in the first 30 minutes (Figure 1C). The amount of REM sleep was below baseline values in the first 30-minute interval after torpor for the longer 2 torpor groups (P<0.05, 2-tailed paired t-test). T CRT was similar for all torpors (Figure 1 D) and differed significantly from baseline in the first 30-minute interval (P<0.05, 2-tailed paired t-test). The changes in the EEG spectrum supported the SWA findings (Figure 2). In addition, it is evident that the lower power density after the short torpor was restricted to the slow-wave range. In the first 30 minutes after torpor, no significant difference in power density was found between the 3 groups in frequencies above 4 Hz (Figure 2). When mean theta activity (5-7 Hz) and high-frequency activity (10-25 Hz) were calculated, no differences between the 3 torpor groups were obtained at any of the time points. A significant positive correlation was found between torpor-bout duration and mean SWA in NREM sleep in the second 30 minutes after emergence from torpor (r=0.42, P<0.05, N=28), when the highest SWA values were reached. Thereafter, 60 to 90 minutes after emergence from torpor, only minor differences remained between the 3 groups in singlefrequency bins. The mean (T mean ) and minimum (T min ) T CRT reached during torpor did not correlate significantly with SWA in the first or second 30-minute interval after torpor for the entire group (P>0.15, N=28) or for the medium-torpor length only (P>0.3, N=15). The T mean and T min both correlated significantly with torpor-bout duration (T mean : r=-0.808, P<0.001; T min : r=-0.748, P<0.001, N=28). Sleep Deprivation Figure 2 Electroencephalographic (EEG) power density in non-rapid eye movement (NREM) sleep (computed for 0.5-Hz bins in the range between Hz, and 1.0-Hz bins between Hz) for the first three 30-minute intervals after emergence from 3 categories of torpor duration [Short (S): <3 hours N=6; Medium (M): 3-6 hours N =15; Long (L): > 6 hours, N =7]. Power density is expressed relative to the mean power density of the baseline light period (=100 %). Values are plotted at the upper limit of each bin. Lines above the abscissa indicate frequency bands, which differ significantly between the 3 torpor duration groups (P<0.05, 2-tailed t-test after significant ANOVA factor duration ). Sleep deprivation induced a significant SWA increase above the baseline level. The magnitude of the SWA increase in NREM sleep depended on the duration of SD. Thus, the initial value after 4-hour SD was significantly higher than after 1.5-hour SD (Figure 3). No relationship was found for the amount of NREM sleep, despite its higher amount in the initial recovery hour after 4-hour SD compared to 1.5- hour SD and its lowest value in the baseline condition (NS, Figure 3, middle panel). The SD had no effect on REM sleep (Figure 3, lower panel). Baseline levels of SWA were restored within 1 hour after 1.5-hour SD, whereas after 4-hour SD, SWA remained above baseline for 1.5 hours. The EEG power spectra computed for three 30-minute intervals after SD also showed an increase in frequencies above 5 Hz (Figure 4). However, differences between the 2 SD conditions occurred only in the slow-wave range. Comparison Between SD and Torpor A qualitative comparison of the EEG spec- SLEEP, Vol. 26, No. 5,

4 30722 Sleep 5.qxd 6/27/2003 2:16 PM Page 570 tra after SD and daily torpor showed that the maximum increase in SWA after a torpor of longer than 6 hours was lower than its increase after 4 hours of SD (Figure 2, middle panel, and Figure 4, left panel). Moreover, an increase in the higher frequencies was found only after SD (Figure 4). In Figure 5, SWA in NREM sleep during recovery is plotted as a function of time after torpor or SD onset, together with the mean baseline value at the time of onset (value at 0.25 hours). The time constant for the increase of process S estimated on the basis of the baseline value and the initial value after 2 SDs of differing durations is approximately 4 hours (upper thin curve), whereas the time constant estimated from the initial baseline value and the highest values attained after the 3 different torpor durations is approximately 11 hours (lower thin curve), which is 2.75 times slower than during SD. The S 0 was the baseline value at 0.25 hours and the upper asymptote (A) was 250% for both conditions (see methods section). After SD, SWA immediately showed a gradual continuous decrease in the first 2 to 3 hours. In contrast, after torpor, SWA increased in the first 30 to 60 minutes before it started to decline. The speed of decline after torpor, once the reversal took place, is similar to the speed after SD (P>0.1. F=1.867, ANOVA for the interaction of factor condition and interval over 4 intervals). DISCUSSION Figure 3 Timecourse of slow-wave activity (SWA) (mean electroencephalographic power density between 0.75 and 4.0 Hz) in non-rapid eye movement (NREM) sleep, and the amount of NREM sleep and rapid eye movement (REM) sleep during baseline and after 1.5 hours and 4 hours of sleep deprivation (SD) (N=8 in the 1.5 hour-group, N=7 in the 4-hour group). Curves connect 30-minute mean values for the first 150 minutes of recovery after SD. SWA is expressed relative to the 24-hour value of the baseline day (=100%). The asterisks indicate significant differences between the 3 conditions (P<0.05, 2-tailed t-test after significant ANOVA factor SD duration ). SLEEP, Vol. 26, No. 5, In this experiment, daily torpor induced a qualitatively similar increase in NREM sleep pressure and timecourse of SWA during recovery sleep after emergence from torpor and after SD in the Djungarian hamster. Moreover, a significant positive correlation was found between torpor-bout duration and the subsequent increase in SWA, confirming previous results obtained in a smaller data set. 15 It took the animals approximately 30 minutes to attain the maximum SWA after torpor. However, SWA showed a torpor-duration-dependent increase in the first and second 30-minute interval after torpor. SWA after torpor was similar to the level observed after 1.5 hours of SD but was lower than after 4 hours SD (Figure 5). It is important to note that the increase in SWA after torpor and its subsequent decline were sleep related and reflect sleep regulation because SWA only dissipated when sleep was allowed. 16,17 Moreover, the timecourse of the decline of SWA after torpor did not differ significantly from the timecourse after SD. Together, the data indicate that there is no qualitative difference in sleep after SD or daily torpor. The group with the shortest torpor had normal amounts of REM sleep and less SWA in the initial 30-minute interval, whereas in the other two torpor groups, REM sleep was suppressed and slow-waves enhanced. This was probably not related to a difference in brain temperature, since T CRT did not differ between the 3 groups (Figure 1D). Also, differences in time of day will hardly have had an influence on the amount of REM sleep, since REM sleep is equally distributed across the day in animals displaying torpor in a short photoperiod. 21 The data suggest that there is a balance between REM sleep and slow waves. REM sleep can be expressed normally after short torpors, and slow waves after longer torpor bouts. This again is in accordance with the hypothesis that NREMsleep pressure increases with torpor-bout duration. The pressure for deep NREM sleep is low after a short torpor bout, allowing for normal amounts of REM sleep, whereas after a long torpor bout, NREM sleep pressure is high, leaving no room for the expression of REM sleep in the first 30 minutes. Despite the known dependence of SWA on prior waking that has been demonstrated in several species, it is an important finding that SWA in NREM sleep also increases as a function of prior SD duration in the Djungarian hamster. This result is in accordance with the homeostatic process S of the 2-process model of sleep regulation 8,9 and underscores the significance of the relationship between torpor and SWA obtained after different torpor durations. The time constant of the increase of process S obtained during the SD (i.e., approximately 4 hours) is similar to the time constants obtained in 2 mouse strains that have a body and brain size similar to the hamsters (129/SvJ: 3.6 h, C57BL/6J: 4.9 h). 14 This result supports the notion that the time constant for SWA increase, which is largest for humans, 23,9 intermediate in rats, 22 and smallest in mice 14 and Djungarian hamsters, may be related to the species-specific duration of the sleep cycle. 14 The time constant for SWA increase seems to be related to body or brain size 20 and may be a physiologic parameter involved in the positive correlation between brain weight and sleepcycle length. 24 The estimated time constant for the increase of process S during torpor was approximately 11 hours (i.e., 2.75 times slower than during

5 Sleep occurring during torpor that still counteracts the build up of sleep pressure, possibly causes the slower build-up. Since sleep pressure increases, this sleep would be less efficient than is euthermic sleep. Previous analysis has shown that the EEG slows down when the animal enters torpor 25 and that euthermic slow waves are virtually absent below 27 C. 15 Therefore, sleep during torpor would counteract the build-up of sleep pressure, while the occurrence of slow waves in the sleep EEG is suppressed. The absence of slow waves during torpor and its behavior during the initial 30 minutes after torpor show that SWA is not a reliable measure for the depth of sleep during that part of the torpor cycle. Other measures, such as arousal thresholds for sound or touch applied throughout a torpor cycle, could provide a Figure 4 Electroencephalographic (EEG) power density in non-rapid eye movement (NREM) sleep (computed for 0.5-Hz-bins in the range between 0.25 and 5.0 Hz and 1.0-Hz bins between 5.25 and 25.0 Hz) for the first three 30-minute intervals after 1.5 measure of sleep depth during the hypothermic state. hours (N=8) and 4 hours of sleep deprivation (SD; N=7). Power density is expressed relative to the mean power density during the baseline (BL) light period (=100 %). Values are plotted at the upper limit of each bin. Lines above the abscissa indicate frequency bands, which differ significantly from BL (=100%) or between the two SD durations (P<0.05, 2-tailed t-test after signif- Another possibility is that the lower T CRT icant ANOVA factor duration ). influences the build-up speed of sleep pressure. Previous research has shown an increase of slow-wave sleep or SWA in humans 26 and cats 27 after an increase in body or brain temperature. However, in rodents, results of warming experiments are contradictory. 19,28-30 In our previous work, we showed that mean T CRT during torpor was approximately 10 C below the corresponding time of day during baseline 15 and 12 C below the mean temperature during a 4-hour SD. 16 The temperature coefficient of the time constant for the increase of process S would therefore be between 2.3 and 2.8, a range typical for biochemical processes. This hypothesis would predict that sleep pressure increases faster at higher torpor temperatures, in contrast to results obtained in hibernating species. 31,32 The T mean did not correlate significantly with subsequent SWA. However, this may be due to the relatively small variability in T mean (Table 1), which did not differ significantly between the 3 groups. Alternatively, the build-up of sleep pressure during torpor may also be linked to the lower metabolic rate during torpor. Metabolic rate is actively lowered before the start of torpor 33 and was not measured in this study. Future studies combining metabolic-rate measurements with T CRT and EEG may resolve this issue. Recent developments in narcolepsy and orexin/hypocretin research integrating sleep and metabolism 34 suggest that changes in metabolic rate may indeed influence sleep regulation. Despite the similarity between the effects of torpor and SD on the slow-wave range of the spectrum, clear differences could be seen in Figure 5 Timecourse of slow-wave activity (SWA) (mean electroencephalographic [EEG] power density between Hz) in non-rapid eye movement (NREM) sleep after sleep higher frequencies. The differences in frequencies above 5 Hz between deprivation (SD) of 2 different durations (open circles; 1.5 hours, N=8 and 4 hours, N=7) the NREM sleep spectra after torpor and after SD are similar to the differences obtained previously. 16 It has been hypothesized that changes in and 3 categories of torpor duration (closed circles; <3 hours, N=6; 3-6 hours, N=15; > 6 hours, N=7). Curves connect 30-minutes mean values for the first 2.5 to 3 hours of recovery (0 hour corresponds to the start of SD or torpor; T CRT < 27 C is defined as torpor). SWA brain temperature can cause changes in EEG power density in faster frequencies of the EEG. 35 When correcting for temperature in our previous is expressed relative to the 24-hour value of the baseline day (=100%). The 2 thin curves indicate the estimated increase of the hypothetical process S during SD or torpor (see methods). The upper thin curve for the increase during SD, with an estimated time constant of 4 experiments, the differences between torpor and SD in EEG power density above 7 Hz disappeared. 16 This renders it likely that differences in hours, and the lower curve for torpor, with an estimated time constant of 11 hours. EEG power density in faster frequencies between torpor and SD are euthermic waking). It is unclear what causes this slower increase during mainly caused by differences in brain temperature and are not related to torpor. The NREM sleep and slow waves produced during the time T differences in regulatory mechanisms between the two conditions. CRT The present analysis indicates that the main increase in SWA after increased from 27 C to euthermia are unlikely to influence subsequent emergence from torpor in the Djungarian hamster is related to prior torpor-bout duration and that the build-up rate of SWA during torpor is SWA in a significant way. In a recent study, we performed an additional SD of 1.5 hours after the animals emerged from torpor. This additional approximately 2.75 times slower than during waking. Investigating the SD did not increase SWA significantly above the SWA after torpor alone. 16 effect of daily torpor on subsequent sleep may also shed light on the processes responsible for the drive to sleep in humans after the daily pro- The results indicate that subsequent SWA would not change significantly if the hamsters were kept awake during the initial 30 minutes longed episode of wakefulness. after torpor instead of being allowed to sleep. Therefore, the slower time constant in the torpor group is probably the result of processes occurring during torpor. SLEEP, Vol. 26, No. 5,

6 30722 Sleep 5.qxd 6/27/2003 2:16 PM Page 572 ACKNOWLEDGEMENT spectrum of humans and animals. J Sleep Res 1998; 7: We would like to thank Dr. G. Heldmaier for providing the animals. This research was supported by the Swiss National Science Foundation Grant and and NWO Grant REFERENCES 1. Berger RJ. Cooling down to hibernate: sleep and hibernation constitute a physiological continuum of energy conservation. Neurosci Lett 1993;154: Walker JM, Garber A, Berger RJ, Heller HC. Sleep and estivation (shallow torpor): continuous processes of energy conservation. Science 1979;204: Daan S, Barnes BM, Strijkstra AM. Warming up for sleep? Ground squirrels sleep during arousals from hibernation. Neurosci Lett 1991;128: Strijkstra AM, Daan S. Sleep during arousal episodes as a function of prior torpor duration in hibernating European ground squirrels. J Sleep Res 1997;6: Trachsel L, Edgar DM, Heller HC. Are ground squirrels sleep deprived during hibernation? Am J Physiol Regulatory Integrative Comp Physiol 1991;260:R Deboer T, Tobler I. Sleep EEG after daily torpor in the Djungarian hamster: similarity to the effect of sleep deprivation. Neurosci Lett 1994;166: Tobler I. Phylogeny of sleep regulation. In: Kryger MH, Roth T, Dement WC, eds. Principles and practice of sleep medicine, 3 rd ed. Philadelphia: WB Saunders; 2000: Borbély AA. A two-process model of sleep regulation. Hum Neurobiol 1982;1: Daan S, Beersma DGM, Borbély AA. Timing of human sleep: recovery process gated by a circadian pacemaker. Am J Physiol 1984;246:R Tobler I, Borbély AA. Sleep EEG in the rat as a function of prior waking. Electroencephalogr Clin Neurophysiol 1986;64: Dijk DJ, Beersma DGM, Daan S. EEG power density during nap sleep: reflection of an hourglass measuring the duration of wakefulness. J Biol Rhythms 1987;3: Lancel M, van Riezen H, Glatt A. Effects of circadian phase and duration of sleep deprivation on sleep and EEG power spectra in the cat. Brain Res 1991;548: Strijkstra AM, Daan S. Dissimilarity of slow-wave activity enhancement by torpor and sleep deprivation in a hibernator. Am J Physiol 1998;275:R Huber R, Deboer T, Tobler I. Effects of sleep deprivation on sleep and sleep EEG in three mouse strains: Empirical data and simulations. Brain Res 2000;857: Deboer T, Tobler I Natural hypothermia and sleep deprivation: Common effects on recovery sleep in the Djungarian hamster. Am J Physiol 1996;271:R Deboer T, Tobler I. Slow waves in the sleep electroencephalogram after daily torpor are homeostatically regulated. Neuroreport 2000;11: Palchykova S, Deboer T, Tobler I. Selective sleep deprivation after daily torpor in the Djungarian hamster. J Sleep Res 2002;11: Larkin JE, Heller HC. The disappearing slow wave activity of hibernators. Sleep Res Online 1998;1: Deboer T, Tobler I. Vigilance states episodes and cortical temperature in the Djungarian hamster: the influence of photoperiod and ambient temperature. Pflügers Arch 1997;433: Deboer T, Franken P, Tobler I. Sleep and cortical temperature in the Djungarian hamster under baseline conditions and after sleep deprivation. J Comp Physiol A 1994;174: Deboer T, Vyazovskiy VV, Tobler I. Long photoperiod restores the 24-h rhythm of sleep and EEG slow wave activity in the Djungarian hamster (Phodopus sungorus). J Biol Rhythms 2000;15: Franken P, Tobler I, Borbély AA. Sleep homeostasis in the rat: Simulation of the time course of EEG slow-wave activity. Neurosci Lett 1991;130: Achermann P, Dijk DJ, Brunner DP, Borbély AA. A model of human sleep homeostasis based on EEG slow-wave activity: quantitative comparison of data and simulations. Brain Res Bull 1993;31: Zepelin H. Mammalian sleep. In: Kryger MH, Roth T, Dement WC, eds. Principles and practice of sleep medicine, 2 nd ed. Philadelphia: WB Saunders;1995: Deboer T, Tobler I. Temperature dependence of EEG frequencies during natural hypothermia. Brain Res 1995;670: Shapiro CM, Allan M, Driver H, Mitchell D. Thermal load alters sleep. Biol Psychiatry 1989;26: McGinty D, Szymusiak R, Thomson D. Preoptic/anterior hypothalamic warming increases EEG delta frequency activity within non-rapid eye movement sleep. Brain Res 1994;667: Morairty SR, Szymusiak R, Thomson D, McGinty DJ. Selective increases in non-rapid eye movement sleep following whole body heating in rats. Brain Res 1993;617: Obál F Jr, Alföldi P, Rubicsek G. Promotion of sleep by heat in young rats. Pflügers Arch 1995;430: Tobler I, Franken P, Gao B, Jaggi K, Borbély AA. Sleep deprivation in the rat at different ambient temperatures: effects on sleep, EEG spectra and brain temperature. Arch Ital Biol 1994;132: Larkin JE, Heller HC. Temperature sensitivity of sleep homeostasis during hibernation in the golden-mantled ground squirrel. Am J Physiol 1996;270:R Strijkstra AM, Daan S. Ambient temperature during torpor affects NREM sleep EEG during arousal episodes in hibernating European ground squirrels. Neurosci Lett 1997;221: Heldmaier G, Ruf T. Body temperature and metabolic rate during natural hypothermia in endotherms. J Comp Physiol 1992;162: Taheri S, Zeitzer JM, Mignot E. The role of hypocretins (orexins) in sleep regulation and narcolepsy. Annu Rev Neurosci 2002;25: Deboer T. Brain temperature dependent changes in the electroencephalogram power SLEEP, Vol. 26, No. 5,