INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

RAPID EYE MOVEMENT (REM) and slow wave sleep (SWS) in adult mammals are controlled by distinct regulatory mechanisms (3, 6). A circadian pacemaker organizes sleep and wakefulness across the 24-h day, and an ultradian mechanism controls the alternation of REM sleep and SWS in each sleep period (3, 6). Mammalian sleep also appears to be homeostatically regulated. Prolonged waking increases the need for sleep in a wide variety of mammalian species (32), and sleep deprivation in mammals produces compensatory increases in REM sleep and SWS time and SWS electroencephalograph (EEG) delta activity (0.5-4.0 Hz) during recovery (6, 11, 17). These changes in sleep need and sleep expression have been hypothesized to reflect the accumulation of sleep need during enforced waking and a homeostatic discharge of sleep need during recovery (6).

The majority of studies characterizing sleep homeostatic mechanisms have been conducted in adult mammals. Consequently, comparatively little is known regarding the maturation of sleep homeostatic mechanisms in developing mammals. Previous studies of neonatal mammals suggest that neonatal sleep is not homeostatically regulated. Pharmacological suppression of active sleep (AS), a putative neonatal homologue to adult REM sleep, does not increase AS amounts during recovery in kittens (28) or rats (22) in the preweaning period. Two days of instrumental sleep deprivation in preweanling monkeys also fail to produce compensatory increases in SWS or AS time during recovery (5). The interpretation of these studies, however, is complicated by several factors. Psychoactive compounds induce a number of behavioral changes in neonates, which may mask subtle changes in neonatal sleep expression following sleep deprivation (see Ref. 16 for discussion). The very stressful forms of sleep deprivation (electric shock and loud tones) used in the neonatal monkey may have contributed to the results of this study because stress disrupts normal neonatal sleep patterns (20). Moreover, in no previous study of neonatal sleep were quantitative assessments made of potential changes in EEG slow wave activity following sleep deprivation.

In this study, we investigated the maturation of sleep homeostasis in the postnatal period using gentle forms of sleep deprivation and protocols designed to minimize stress in neonatal rats. We have previously demonstrated that electrographically distinct sleep states emerge from a background of relatively undifferentiated EEG and motor activity at approximately postnatal days 11-12 (P11-P12) in the rat (15), and these states have diurnal organization by P20 (14). Whether or not these early occurring sleep states are homeostatically regulated is unknown. In addition, there are presently no studies on the effects of sleep deprivation in developing mammals that span the preweaning and postweaning period. The ages we selected began ~24-48 h after the first appearance of electrographically determined sleep and concluded in the postweaning period (15, 19). The primary objective of this study was to determine when homeostatic regulation of REM sleep and SWS first appears during development.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Neonatal rat EEG and electromyograph implantation. Pregnant Long-Evans dams in our colony were maintained in a 12:12-h light-dark (LD) cycle, at an ambient temperature (T_a) of 22°C, and provided food and water ad libitum. The day of birth was designated as postnatal day 0 (P0). Each litter was culled to 10 pups, and the home cage was moved to the sleep recording room (identical 12:12-h LD cycle), where the animals were left undisturbed until surgery. Two days before the first sleep recording (at P10, P14, P18, or P21), equal numbers of male and female pups (8 each, randomly selected from 14 different litters) were randomly assigned to sleep deprivation by drum (Dep-D) or control (Con) group. Each Con pup was age and sex matched with a corresponding Dep-D pup. Two additional sets of randomly selected, age- and sex-matched pups from our colony formed the gentle-handling sleep deprivation (Dep-H) groups (P10, n = 6; P21, n = 4). In this manner, each group of pups was formed from a population of animals randomly distributed with respect to litter and maternal factors, but matched with respect to body mass, sex, and date of birth. Each pup was placed on a heating pad and anesthetized with Metofane (methoxyflurane) inhalant. Miniaturized EEG electrodes (no. 000 stainless steel screws beveled to roughly 1.5 mm) were implanted bilaterally over frontal and parietal cortex. Electromyographic (EMG) electrodes were bilaterally inserted into nuchal muscles. The electrodes were attached with stainless steel Teflon-coated wire to a seven-pin electrical socket. The entire assembly was fixed in place with dental acrylic. The pups were injected with gentamicin antibiotic (subcutaneous) immediately after surgery. On recovery from anesthesia, the pups were returned to the home nest. One day before the first recording, preweanling pups (P11-P15) were fitted with stainless steel cheek cannulas made from 20-gauge hypodermic needles. The needles were cut into 1.5- to 2.0-cm tubes, the ends were smoothed, and a washer cut from P100 Teflon tubing was placed on one end. The pups were lightly anesthetized, and the feeding cannulas were threaded through their lower cheeks. Each cannula was fixed in place with a drop of cyanoacrylate glue on the outside of the cheek. All surgical procedures were approved by the Stanford University Administrative Panel for Laboratory Animal Care.

Neonatal rat sleep recording procedures. Twelve hours before the sleep deprivation, Con and Dep-D and Dep-H pups were briefly anesthetized (Metofane: <5 min) and attached to flexible recording cables, which in turn were connected to slip-ring commutators. The anesthesia was necessary to prevent the struggling animal from damaging the EEG/EMG implant, and at no other times were the pups anesthetized. The Dep-D groups were fitted with partitioned cables that allowed for quick disconnection and reconnection of the recording cable 2-3 cm from the EEG/EMG implant. These were electrical cables that had an extra electrical socket interposed 2-3 cm from the skull implant. These partitioned cables allowed us to disconnect Dep-D pups from their electrical cables without anesthesia and without restraint. Feeding tubes were attached to the cheek cannulas and connected to slip-ring injection ports. Pups were then placed in acrylic incubators (14 cm²) containing home bedding material and warmed by a self-enclosed water bath housed in a grounded Faraday cage. The incubator temperatures were reduced at each age to compensate for increasing thermogenesis in the rat pups (T_a, P12: 33-34°C; P16: 29-30°C; P20-P24: 22°C) (15). P12 and P16 rat pups were periodically groomed and fed an enriched milk formula according to schedules previously shown to provide for normal growth and to control for the effects of maternal separation on sleep organization (15, 20). P20 and P24 rat pups were provided with water and rat chow softened in milk formula ad libitum. Con, Dep-D, and Dep-H pups were placed in their incubators the night before the sleep-deprivation experiment at zeitgeber time (ZT) 19. All sleep deprivations began the next day at ZT 0 (lights on) at the beginning of the rest phase.

Data collection procedure. EEGs were recorded on a Grass 7 polygraph. Unihemispheric, frontal-parietal EEG potentials were filtered at 0.3 and 35 Hz (1/2 max, 6 dB/octave), digitized at 100 Hz, and stored in 10-s epochs on a personal computer. EMG signals were full-wave rectified, integrated, and stored as one value (0-100) per epoch. The EEG was then Fourier analyzed in 10-s epochs using a fast Hartley transform. Data collection began at lights out (2000), 12 h before the beginning of the sleep deprivation, and continued for 13 h, at which time the experiment was terminated. These were acute studies, and each group of pups was used only once for a given time point (at either P12, P16, P20, or P24).

Deprivation technique. The Dep-D rats were sleep deprived using a modified form of forced locomotion. The sleep-deprivation chamber was a revolving acrylic drum (30 cm diameter) lined with an absorbent pad. The drum was sealed with a perforated, acrylic lid and warmed to incubator temperatures by a lamp. At P12 (n = 8), P16 (n = 8), P20 (n = 7), and P24 (n = 7), 2-3 Dep-D pups (or single Dep-D pups with 1-2 littermates) were disconnected from their partitioned EEG/EMG recording cables and feeding tubes and placed in the center of the drum. Con pups were left undisturbed in the incubators with the milk flow turned off for the duration of the sleep deprivation. The drum was then slowly rotated (2-3 rpm), with the number of rotations gradually increasing up to a maximum of 6-7 rpm near the end of the sleep-deprivation period. The drum's rotation forced the P12-P16 pups to climb over each other, and right themselves while the older rats (P20+) walked or hopped during the drum's rotation. All sleep deprivations began at lights-on (ZT 0) and continued for 3 h. After the deprivation, Dep-D pups were returned to their incubators, milk flow was resumed for both Con and Dep-D preweaned groups, and sleep data were collected for 13 h.

The Dep-H pups were sleep deprived using gentle handling to provide comparison data for the forced locomotion technique and also to verify that EEGs were desynchronized during the sleep deprivation. The Dep-H pups remained in their warmed incubators (with the milk flow turned off) and were gently prodded or made to crawl once sleep was detected. Sleep was detected by the presence of EEG slow waves in the polygraphic record and/or behaviorally observed phasic motor activity coupled with REMs during behavioral quiescence. Milk flow was resumed at the termination of the sleep deprivation, and sleep data were collected for 13 h.

Vigilance state determination. A computer algorithm validated for use with neonatal rats was used to determine vigilance states (14). For each pup, the EEG records were Fourier transformed, and the resulting delta (0.5-4.0 Hz) power (DP) values for each 10-s epoch were plotted with the corresponding integrated EMG signals. Epochs with high DP coupled with a low EMG signal were scored as SWS. Epochs with low DP coupled with a low EMG were scored as AS. Although AS may initially be distinct from REM sleep according to our previous work (15), we use the term AS in this report to refer to all REM sleep-like states. Epochs with low values of DP coupled with a high EMG signal were scored as wake. Epochs with movement artifacts were discarded from analyses. The algorithm used to score vigilance states in neonatal rats has been shown to agree with manual scoring (polygraph) by >90% (14). Once the sleep data were divided into AS, SWS, and wake, the following analyses were conducted.

Effects of sleep deprivation on SWS DP. Changes in mean SWS DP were determined following sleep deprivation at P12, P16, P20, and P24. SWS DP hourly values for Con and Dep-D and Dep-H groups were expressed as a percentage of the mean SWS DP value in the dark period preceding the sleep deprivation.

Effects of sleep deprivation on sleep distribution and sleep continuity. The amounts of AS, SWS, and wake as a percent of total recording time (TRT) were calculated following sleep deprivation at all ages. The vigilance state amounts were then compared with time-matched values obtained from Con pups.

We also examined the effects of sleep deprivation on sleep continuity in the recovery period at all ages. This was accomplished by determining the mean duration of AS, SWS, and wake bouts (20 s was considered the minimum bout length) in Dep-H, Dep-D, and Con groups after sleep deprivation. We also determined the number of brief arousals (BAs) occurring following sleep deprivation in all three groups. Brief arousals were defined as wake periods 10-20 s in length that occurred within ongoing sleep bouts (17). BAs were expressed as per hour of total sleep in 3-h bins (17), which in total comprised the entire recovery period following sleep deprivation.

Estimating recovery of lost sleep time. We also determined how much SWS and AS was recovered following sleep deprivation in the Dep-D and Dep-H groups. For each time point, we calculated the mean amounts of SWS and AS (in minutes) in the first 3 h of the light period in the Con group. This provided an estimate of total SWS and AS lost during the sleep deprivation in the Dep-D and Dep-H groups. We then calculated the mean difference in SWS and AS amounts across the recovery period (hours 3-12) among the Dep-D, Dep-H, and Con groups. Separate recovery values for SWS and AS were then calculated by expressing the difference in SWS and AS amounts as a percent of the estimated amount of SWS and AS lost. These sleep recovery values were then tabulated as a function of developmental age.

Changes in EEG DP during sleep deprivation and in other vigilance states. Our finding that P12 Dep-H and Dep-D animals did not intensify SWS DP following sleep deprivation was rather surprising. Therefore, to verify that EEG DP was not intensifying and discharging in AS or wakefulness, we conducted the following post hoc analyses. We first determined whether EEG DP was accumulating and being discharged during the sleep-deprivation period by measuring the level of DP in the P12 Dep-H group during sleep deprivation. The level of EEG DP in Dep-H and Con groups was expressed as a percent of the 12-h mean wake DP value obtained from the 12-h recording period before sleep deprivation. We also determined whether EEG DP was discharging in AS or wakefulness by measuring the level of DP in AS and wake in P12 and P16 rats. AS and wake EEG DP values in Dep-D and Dep-H groups were normalized to respective 12-h mean AS DP and wake DP values obtained from the 12-h recording period before sleep deprivation. These normalized DP values were then compared with similarly normalized AS DP and wake DP values derived from the Con group.

Statistics. All statistics were performed with SAS/STAT software (27). Repeated-measures ANOVAs were used to test for significance between main effects by condition (Dep-D pups and Dep-H pups vs. Con pups), time, and interactions between time and condition levels. Ryan-Einot-Gabriel-Welsch post hoc tests were used when there were significant main effects by condition, time, or significant interactions between time and condition levels (27). Student's t-tests were employed at all other times.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Sleep deprivation in P12-P24 rats: behavioral effects. It was extremely difficult to keep P12 rats awake beyond 1-1.5 h. In P12 rat pups, the frequency of attempts to enter sleep dramatically increased after 1-1.5 h of gentle handling or time in the revolving drum. Once returned to their cages, P12 rat pups rapidly entered sleep. P16 rats appeared able to withstand longer periods of sleep deprivation, but after 2 h of forced locomotion, the frequency of attempts to enter sleep rapidly increased. P20-P24 rat pups could maintain wakefulness for longer periods of time without difficulty (>2 h). However, by the third hour of total sleep deprivation, P20-P24 rats frequently attempted to enter sleep. On return to their cages, P20-P24 rats were often awake for 5 to 10 min (grooming and inspecting their cages) before falling asleep.

Sleep deprivation in P12-P24 rats: effects on SWS DP. Sleep deprivation by forced locomotion or gentle handling did not increase SWS DP until P24. In P12 (condition, time, and interaction: NS, P > 0.05) and P20 rats (condition, time, and interaction: NS, P > 0.05), SWS DP did not significantly differ from control values during recovery sleep. In P16 rats, SWS DP was suppressed relative to control values in the recovery period after sleep deprivation (condition: df = 1, F = 5.69, P < 0.032; time, interaction: NS, P > 0.05). By P24, sleep deprivation significantly increased SWS DP during recovery sleep relative to control values (condition: df = 2, F = 6.84, P < 0.02; interaction: df = 18, F = 6.23, P < 0.0001). After this initial increase, SWS DP declined to control values (see Fig. 1).

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Changes in slow wave sleep (SWS) delta power (DP: 0.5-4.0 Hz) after 3-h sleep deprivation (SD) by forced locomotion (Dep-D) or gentle handling (Dep-H) compared with non-sleep-deprived controls (Con) in postnatal day 12 (P12), P16, P20, and P24 rats. Mean (±SE) SWS DP is expressed as a percent of the 12-h mean SWS DP value before sleep deprivation at time 0-3. Hour 12 represents first hour of dark period following sleep deprivation. a Significant difference between Dep-D, Dep-H, and Con values (P < 0.05). b Significant difference between Dep-D and Con values (P < 0.05).

Sleep deprivation in P12-P24 rats: effects on vigilance states. In P12-P24 rats, sleep deprivation by forced locomotion or gentle handling significantly altered the distribution of sleep states. From P12-P16, sleep deprivation in both Dep-H and Dep-D groups increased SWS amounts as a percentage of TRT primarily in the first 2 h of recovery (P12 condition: df = 2, F = 20.92, P < 0.0001; time: df = 9, F = 4.88, P < 0.0001; interaction: df = 18, F = 2.88, P < 0.0002; P16 condition: NS; time: df = 9, F = 2.66, P < 0.007; interaction: df = 9, F = 2.72, P < 0.006). In P20-P24 rats, sleep deprivation did not significantly increase SWS time during recovery (P20 condition, time, interaction: NS, P > 0.05; P24 condition, time, interaction: NS, P > 0.05) (see Fig. 2).

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Changes in mean (±SE) SWS amounts as a percent of total recording time (TRT) after 3-h sleep deprivation by forced locomotion or gentle handling compared with non-sleep-deprived controls in P12, P16, P20, and P24 rats. Hour 12 represents first hour of dark period following sleep deprivation. a Significant difference between Dep-D, Dep-H, and Con values (P < 0.05). b Significant difference between Dep-D and Con values (P < 0.05).

Sleep deprivation did not elicit rebound increases in AS time until P16-P20. In P12 rats, AS amounts were slightly suppressed in the Dep-H group relative to the Con group primarily in the first recovery hour (condition: df = 2, F = 6.85, P < 0.002; time: df = 9, F = 2.91, P < 0.03; interaction: NS, P > 0.05). At ages P16-P24, AS amounts in the sleep-deprived groups were initially suppressed relative to control values and then significantly increased over control values (P16 condition: NS, P > 0.05; time: df = 9, F = 2.89, P < 0.004; interaction: df = 9, F = 3.1, P < 0.002; P20 condition: df = 1, F = 6.98, P < 0.009; time: df = 9, F = 4.92, P < 0.0001; interaction: df = 9, F = 2.99, P < 0.003; P24 condition: df = 2, F = 4.7, P < 0.01; time: df = 9, F = 7.25, P < 0.0001; interaction: df = 18, F = 1.68, P < 0.05) (see Fig. 3).

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Changes in mean (±SE) active sleep (AS) amounts as a percent of TRT after 3-h sleep deprivation by forced locomotion or gentle handling compared with non-sleep-deprived controls in P12, P16, P20, and P24 rats. b Significant difference between Dep-D and Con values (P < 0.05). c Significant difference between Dep-H and control values (P < 0.05). Hour 12 represents the first hour of the dark period following sleep deprivation.

Changes in wake time after sleep deprivation were less consistent than changes in sleep time (data not shown). Sleep deprivation decreased wake time during recovery in P12 Dep-D and Dep-H groups (condition: df = 2, F = 7.88, P < 0.0005; time: NS, P > 0.05; interaction: df = 18, F = 1.83, P < 0.03), primarily in the first 2 h of recovery. In P20 rats, wake time was primarily suppressed in the first and sixth through eighth hours of recovery (condition: df = 1, F = 6.28, P < 0.01; time: NS, P > 0.05; interaction: NS, P > 0.05). Sleep deprivation had no effects on wake amounts in P16 (condition, time, interaction: NS, P > 0.05) and P24 rats (condition, interaction: NS, P > 0.05; time: df = 9, F = 3.53, P < 0.001).

Changes in sleep continuity after sleep deprivation. Sleep deprivation had no effect on either AS or SWS bout durations or on BAs in the recovery period until P20 (see Figs. 4 and 5). In P12 rats, sleep deprivation had no effect on AS (condition: NS, P > 0.05) or SWS (condition: NS, P > 0.05) bout durations. However, in both Dep-H and Dep-D groups, there was a significant reduction in wake bout duration compared with Con group: Dep-H, 0.93 ± 0.1 min; Dep-D, 0.98 ± 0.11 minutes; and Con 1.3 ± 0.1 min (means ± SE; condition: df = 2, F = 5.43, P < 0.005). The BAs per hour of total sleep did not differ between Dep-H, Dep-D, or Con groups in P12 rats after sleep deprivation (condition, time, interaction: NS, P > 0.05). In P16 rats, there were no differences between Dep-D and Con groups in AS (condition, time, interaction: NS, P > 0.05), SWS (condition, time, interaction: NS, P > 0.05), or wake (condition, time, interaction: NS, P > 0.05) bout durations. Nor were there any differences between P16 Dep-H and Con groups in BAs during sleep following sleep deprivation (condition, time, interaction: NS, P > 0.05). In P20 and P24 rats, however, sleep deprivation significantly increased sleep bout durations and reduced BAs during recovery sleep. In P20 rats, sleep deprivation increased the duration of AS bouts (condition: df = 1, F = 9.4, P < 0.003) and SWS bouts (condition: df = 1, F = 4.77, P < 0.03) during recovery. There were no significant effects on wake bout duration (condition: NS, P > 0.05). Sleep deprivation in P20 rats also reduced BAs expressed per hour total sleep (condition: df = 1, F = 13.01, P < 0.0009; interaction, time: NS, P > 0.05) during recovery. In P24 rats, sleep deprivation increased the duration of SWS bouts in the Dep-D group (condition: df = 2, F = 3.67, P < 0.03) compared with the Con group. The P24 Dep-H group, however, showed only a significant increase in AS bout duration compared with the Con group (condition: df = 2, F = 4.1, P < 0.02). There were no significant effects on wake bout durations in either Dep-H or Dep-D P24 rats after sleep deprivation (condition: NS, P > 0.05). The analysis of BAs was restricted to the Dep-D and Con groups because only two animals in the Dep-H group were available for analysis, but in both Dep-D and Dep-H groups sleep deprivation reduced BAs during recovery sleep. In the Dep-D group, BAs expressed per hour total sleep (condition: df = 1, F = 9.2, P < 0.005; time, interaction: NS, P > 0.05) were lower than in the Con group.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Changes in mean (±SE) amounts of brief arousals (BAs) after 3-h sleep deprivation by forced locomotion or gentle handling compared with non-sleep-deprived controls in P12, P16, P20, and P24 rats. BAs are expressed per hour of total sleep in 3-h bins comprising the entire recovery period. * Significant difference between Con and Dep group (P < 0.05).

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Changes in mean (±SE) SWS and AS bout durations after 3-h sleep deprivation by forced locomotion or gentle handling compared with non-sleep-deprived controls in P12, P16, P20, and P24 rats. b Significant difference between Dep-D and Con values (P < 0.05). c Significant difference between Dep-H and Con values (P < 0.05).

Estimating the amount of SWS and AS time recovered after sleep deprivation. The recovery of sleep lost during the sleep deprivation period changed dramatically during postnatal development. There was nearly a complete recovery of lost SWS time in the P12 rats (90-96%) during the post-sleep-deprivation period. By P24, however, the recovery of lost SWS time had declined to very small values (8-9.3%). In contrast, AS time was not recovered at all in P12-P16 rats. By P20 and P24, ~80-95% of lost AS time was recovered after sleep deprivation (see Fig. 6).

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Recovery of SWS and AS time after 3-h sleep deprivation by forced locomotion in P12, P16, P20, and P24 rats. Data from the post-sleep-deprivation period are expressed as percentages of SWS and AS estimated lost (based on control group values) during sleep deprivation (see METHODS). Similar recovery values were obtained from P12 and P24 rats sleep deprived by gentle handling (P12: SWS, 96%, AS, 0%; P24: SWS, 9.3%, AS, 95%).

Changes in EEG DP during sleep deprivation and in other vigilance states. We were surprised to find that rats younger than P24 did not increase SWS DP after sleep deprivation. To determine whether sleep need, as reflected by increases in EEG DP, was accumulating and being discharged during wakefulness or in AS, we first measured the level of waking EEG DP in Dep-H and Con groups. EEG DP in P12 rats did not appear to be accumulating and discharging during the sleep deprivation. Although wake EEG DP (as a percentage of wake baseline values) tended to rise near the end of the deprivation period, this increase was not significantly greater than wake EEG DP values observed in normally sleeping control pups (condition, time, interaction: NS, P > 0.05) (see Fig. 7). Nor were there significant increases in EEG DP in waking periods after sleep deprivation (condition, time, interaction: NS, P > 0.05, data not shown). There were increases in AS EEG DP after sleep deprivation (compared with AS baseline) in P12 and P16 rats (P12 condition: NS, P > 0.05; time: df = 9, F = 2.5, P < 0.01; interaction: df = 18, F = 3.44, P < 0.0001; P16 condition: NS, P > 0.05; time: df = 9, F = 3.61, P < 0.01; interaction: df = 9, F = 4.39, P < 0.002) (see Fig. 8). However, the levels of AS EEG DP were considerably lower than EEG DP levels observed in SWS and at no time exceeded baseline SWS DP values. For example, mean AS DP values in the first 3 h of recovery (when AS DP increases were maximal) in the P16 Dep-D group were 31.4% of the DP values observed in baseline SWS. Similar results were obtained at P12 in both the Dep-D and Dep-H groups. In P12 rats, AS DP values in the first 2 h of the recovery period were 40-50% of the baseline SWS DP values. Therefore, the small increase of AS DP did not appear to compensate for the absence of SWS DP increases after sleep deprivation.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Changes in electroencephalograph (EEG) DP (0.5-4.0 Hz) during sleep deprivation by gentle handling in P12 rats. Thirty-minute mean (±SE) EEG DP values are expressed as a percent of the 12-h mean wake DP value before sleep deprivation at time 0-3 in Dep-H and Con rats.

View larger version (22K): [in this window] [in a new window]   Fig. 8.   Changes in AS DP (0.5-4.0 Hz) after 3-h sleep deprivation by forced locomotion or gentle handling compared with non-sleep-deprived controls in P12, P16, P20, and P24 rats. Mean (±SE) AS DP is expressed as a percent of the 12-h mean AS DP value before sleep deprivation at time 0-3. Hour 12 represents first hour of dark period after sleep deprivation. AS DP increases were 30-50% of baseline SWS DP values (see RESULTS for discussion). a Significant difference among Dep-D, Dep-H, and Con values (P < 0.05). b Significant difference between Dep-D and Con values (P < 0.05). c Significant difference between Dep-H and Con values (P < 0.05).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

This is the first study of sleep homeostasis in developing mammals that spans the preweaning and postweaning period. We find that 3-h sleep deprivation produces dramatically different compensatory responses in developing rats. Before P24, neonatal rats do not increase SWS DP in response to sleep deprivation. However, P12-P20 rats do show evidence of sleep homeostasis because they increase sleep time and sleep continuity during recovery sleep. These increases in sleep time were most marked in the youngest rats. P12 rats recovered between 90 and 96% of the total amounts of SWS lost during the sleep deprivation, strongly suggesting that neonatal rats carefully regulate SWS time.

The fact that 3-h sleep deprivation does not increase SWS DP in P12-P20 rats is surprising given previous reports of SWS DP increases in P24 rats after 2-h sleep deprivation (1). We also find large increases in SWS DP at P24, which indicates that the P12-P20 responses reflect developmental processes and not differences between rat strains or experimental protocols. Moreover, we have recently demonstrated that the normal decline in SWS DP during the rest phase typical of adult mammals and reflective of adult sleep homeostasis is absent in rats younger than P24 (14). Overall, these findings suggest that sleep is regulated during early development and that the accumulation and discharge of sleep need changes dramatically between the third and fourth postnatal week.

Sleep deprivation and stress responses in neonates. Sleep deprivation in adult mammals can transiently increase circulating levels of stress hormones (33). These hormones have complex effects on the amounts of sleep and wake time (as a percent of total sleep, or recording time) and lesser effects on the sleep EEG (29, 34). Neonatal sleep also appears to be influenced by elevations in stress hormones. Maternal separation, for example, increases stress hormone levels (31) and disrupts normal sleep/wake patterns in neonatal rats (20). It is, therefore, reasonable to ask to what extent the stress associated with sleep deprivation contributed to the findings of the present study.

We feel that stress was not a significant factor in our results. Maternal factors important in preserving the stress-hyporesponsive period (SHRP) were included in the present study (15). The SHRP extends from P4 to approximately P14 in the rat and represents a period of time when the infant rat is hyporesponsive to stresses that induce large stress hormone release in adult rats (9). The SHRP can be maintained in neonatal rats by providing the grooming, feeding, and heating protocols included in the present study (15, 20, 31). We also find that gentle forms of sleep deprivation (gentle handling) produce changes in sleep expression similar to forced locomotion in P12 and P24 rats. In addition, more stressful forms of sleep deprivation in newborn monkeys do not produce changes in sleep comparable to those reported in the present study (5). Therefore, although stress cannot be completely excluded as a factor in the present study, we believe that factors other than stress are responsible for the changes in neonatal sleep patterns following sleep deprivation.

Do compensatory sleep changes in P12-P20 rats reflect sleep homeostasis? Total sleep deprivation in adult mammals decreases sleep latency and produces compensatory increases in sleep time, sleep continuity, and SWS DP during recovery sleep (6, 11, 17). These changes in sleep latency and sleep expression are thought to reflect the accumulation of sleep need during enforced waking and a homeostatic discharge of sleep need during recovery (6).

There are several findings that suggest that neonatal sleep is also homeostatically regulated. Sleep deprivation increases the number of attempts to enter sleep in human neonates (2). We and others (1) find that short periods of sleep deprivation easily tolerated by adult rats produce very high levels of sleepiness in neonatal (P20) and juvenile (P24+) rats. These findings suggest that sleep deprivation produces an accumulation of sleep need in neonatal mammals. After sleep deprivation, human infants increase subsequent SWS time in a manner that suggests a compensatory response to sleep deprivation (2). In the present study, we also find what may be compensatory increases in SWS time after sleep deprivation. These increases were most evident at P12 and P16, when sleep deprivation elicited a elevation of SWS time followed by a decline. This result was strikingly similar to the initial enhancement and subsequent decline of SWS DP observed in adult mammals after sleep deprivation (6, 11, 17). The increase in SWS time in P12 rats almost fully compensated for the absolute amounts of SWS lost during the sleep deprivation. By P20, the increase in SWS time following sleep deprivation had diminished, large recoveries of lost AS time were observed, and both SWS and AS became more consolidated. These findings indicate that sleep homeostasis may be present in neonatal rats but its manifestation as increased SWS DP is absent until the fourth postnatal week.

What is the nature of the change in sleep homeostatic mechanisms? Why do neonatal and juvenile rats respond so differently to sleep deprivation? One possible explanation for this difference is that the neuronal circuits and/or membrane properties necessary for EEG delta wave activity are very immature in rats younger than P24. The EEG delta waves of SWS are generated by oscillatory interactions between large populations of reticular thalamic, thalamocortical, and cortical neurons (30). The oscillations within these intrathalamic and thalamocortical circuits are in turn dependent on several membrane properties intrinsic to thalamic and cortical neurons, most notably a low-threshold calcium current (I_t) and a hyperpolarization-activated cation current (I_h) (30). The initial increase in EEG delta activity observed during sleep onset is associated with increasing hyperpolarization in individual thalamic and cortical neurons and a progressive synchronization of oscillatory-burst firing in thalamocortical assemblies (30). The enhancement of EEG delta activity after sleep deprivation in adult mammals is presumably mediated by similar mechanisms.

Ontogenetic changes in neuronal membrane properties and circuitry do not appear to be related to the dramatic developmental switch in homeostatic responses to sleep deprivation found in the present study. In mice, which have a time course for SWS development comparable to the rat (8), the morphology and function of intrathalamic connections necessary for EEG delta waves are largely mature by P20 (35). Oscillations within the delta range (2-4 Hz) in intrathalamic circuits are observed by P12 and appear to be generated by I_t and I_h currents (35). A similar time course for the maturation of adult membrane properties necessary for EEG delta waves is also reported for rat thalamic and cortical neurons (23, 36). The basal levels of EEG delta activity during SWS are also at adult levels by P16-P18 in the rat (15, 19). These findings suggest that the maturation of the neuronal circuitry and membrane properties necessary for EEG delta wave activity is not a critical factor in the development of adultlike responses to sleep deprivation in the rat.

A second possible explanation for why neonates and adult mammals respond so differently to sleep loss is that the release and metabolism of sleep-promoting substances in the developing brain are different than in the adult brain. Adenosine, for example, has received increased attention in recent years as an endogenous sleep-promoting molecule that mediates the accumulation and discharge of sleep need (4, 5). It has been further hypothesized that the release of extracellular adenosine primarily occurs when cerebral glycogen stores are reduced during waking (4). Adenosine's sleep-promoting effects are thought to permit the restoration of this supplemental source of fuel to the brain (4). If this hypothesis is correct, then one might expect that the differences in neonatal and adult responses to sleep deprivation can be explained in terms of developmental changes in adenosinergic systems and brain metabolism.

There are a number of important developmental changes in adenosinergic systems and brain metabolism in the rat that may account for the developmental changes in how neonatal rats respond to sleep deprivation. The number of mammalian brain adenosine-1 receptors increase during the first three postnatal weeks, reaching adult levels between P20 and P30 (18). Therefore, the developmental increase in SWS DP following sleep deprivation between P20 and P24 may partially represent the development of neuronal responsiveness to extracellular adenosine.

Alternatively, differences in neuronal fuel sources available to developing and adult brains may explain why neonatal rats fail to increase EEG DP following sleep deprivation. In contrast to adult neurons, whose energy is provided almost entirely by glucose from cerebral blood and astrocytic glycogen, the preweanling brain obtains up to 72% of its energy from ketone bodies and other products of lipid metabolism obtained from the fat-rich diet of the neonate (10). The switch to full glucose metabolism does not occur until the fourth postnatal week (10), when glycogen deposition in astrocytes and the concentrations of enzymes required for glycogen mobilization reach values comparable to those reported for the adult brain (13, 24). Before the fourth postnatal week, neonatal rats are unable to quickly mobilize brain glycogen stores even under extreme metabolic demands (12, 13). These findings suggest that enforced waking depletes different fuel sources in the neonatal and adult brain. Consequently, it is possible that the adenosine release associated with glucose metabolism, and perhaps necessary for adultlike responses to sleep deprivation (4), is not present in the neonatal brain.

Does SWS regulation precede AS regulation? REM sleep (often called AS during infancy) has traditionally been thought of as a precociously developed sleep state. We have recently demonstrated that the maturation of REM sleep may be much slower than previously believed (15). AS expression, for example, is not suppressed by cholinergic blockade until P14, which suggests that cholinergic mediation of REM sleep develops relatively slowly in the postnatal period (16). We now show that sleep deprivation does not increase AS amounts during recovery in P12 rats. In P16-P24 rats, however, sleep deprivation produces small increases in AS amounts during the latter one-half of the recovery period. By P20-P24, increases in AS time begin to compensate for AS time lost during the sleep-deprivation period. One potential explanation for these findings is that a 3-h sleep deprivation is too short to induce AS homeostatic responses in P12 rats. An alternative explanation for the relatively late appearance of AS regulation is that AS regulation, like other aspects of AS, appears relatively late in the postnatal period. The findings of the present study are consistent with a more protracted development of AS because the first signs of AS regulation appeared several days after similar regulation was observed in SWS.

Perspectives

The possibility that neonatal sleep is not simply a passive withdrawal from the external environment but an actively regulated state supports previous suggestions that sleep serves important functions in developing mammals (22, 26). REM sleep, for example, has received considerable attention as a state that facilitates brain development (22, 26). REM sleep may provide endogenous brain stimulation at times when wake amounts are very low (22, 26). In fact, neonatal mammals seem to need more REM sleep than adults because REM sleep amounts progressively decrease as mammals mature (15).

The observation that neonatal rats regulate SWS time suggests that SWS may also be important for developing mammals. The maturation of SWS coincides with the formation of thalamocortical and intracortical patterns of innervation and periods of heightened synaptogenesis (21). SWS is associated with two processes important in synaptic remodeling: elevations of intracellular Ca²⁺ and synchronized firing in neuronal networks (7). It is therefore possible that neonatal SWS contributes to the maturation of neuronal networks by providing an endogenous source of repetitive, synchronized activity in the developing brain.