Two-process Model Of Sleep Regulation Research Articles

Effects of sleep deprivation on sleep and sleep EEG in three mouse strains: empirical data and simulations

Gene targeted mice can be used as models to investigate the mechanisms underlying sleep regulation. Three commonly used background strains for gene targeting (129/Ola, 129/SvJ and C57BL/6J) were subjected to 4-h and 6-h sleep deprivation (SD), and their sleep and sleep EEG were continuously recorded. The two-process model of sleep regulation has predicted the time course of slow-wave activity (SWA) in nonREM sleep after several sleep-wake manipulations in humans and the rat [3, 9]. We tested the capacity of the model to predict SWA in nonREM sleep on the basis of the temporal organization of sleep in mice. The strains differed in the amount and distribution of sleep and the time course of SWA. After spontaneous waking episodes of 10–30 min as well as after SD, SWA was invariably increased. Simulations of the time course of SWA were successful for 129/SvJ and C57BL/6J, but were not satisfactory for 129/Ola. Since the time constants are assumed to reflect the dynamics of the physiological processes involved in sleep regulation, the results provide a basis for the use of gene targeted mice to investigate the underlying mechanisms.

Sleep homeostasis and models of sleep regulation.

According to the two-process model of sleep regulation, the timing and structure of sleep are determined by the interaction of a homeostatic and a circadian process. The original qualitative model was elaborated to quantitative versions that included the ultradian dynamics of sleep in relation to the non-REM-REM sleep cycle. The time course of EEG slow-wave activity, the major marker of non-REM sleep homeostasis, as well as daytime alertness were simulated successfully for a considerable number of experimental protocols. They include sleep after partial sleep deprivation and daytime napping, sleep in habitual short and long sleepers, and alertness in a forced desynchrony protocol or during an extended photoperiod. Simulations revealed that internal desynchronization can be obtained for different shapes of the thresholds. New developments include the analysis of the waking EEG to delineate homeostatic and circadian processes, studies of REM sleep homeostasis, and recent evidence for local, use-dependent sleep processes. Moreover, nonlinear interactions between homeostatic and circadian processes were identified. In the past two decades, models have contributed considerably to conceptualizing and analyzing the major processes underlying sleep regulation, and they are likely to play an important role in future advances in the field.

Models of human sleep regulation.

Non-REM sleep deprivation and REM sleep deprivation both lead to specific rebounds, suggesting that these states fulfil physiological needs. In view of impaired performance after sleep deprivation, a recovery function of sleep seems likely. The timing of this recovery is restricted to a narrow time interval within the 24 hour day, i.e. the night. Generally, nocturnal sleep in humans is considered a consequence of the impact of the circadian pacemaker in the hypothalamus on sleep propensity. The interaction between the homeostatic recovery process and the circadian pacemaker has been modelled in the two-process model of sleep regulation. This model is used as a starting point in the present review. A series of refinements and several alternative models are discussed, both with respect to the quality of fit of theory and data, as well as with respect to the concepts behind the models.

A schizophrenic patient with an arrhythmic circadian rest-activity cycle

A haloperidol-treated patient with chronic schizophrenia had a near-arrhythmic circadian rest-activity cycle, whereas rhythms of 6-sulphatoxy-melatonin and core body temperature were of normal amplitude and phase-advanced. Sleep electroencephalography measured throughout a 31-h ‘constant-bedrest’ protocol revealed a phase-delayed sleep-wake propensity cycle, low sleep continuity (ultradian ‘bouts’), and very little slow-wave sleep and slow-wave activity (0.75–4.5 Hz). Switching treatment to the atypical neuroleptic clozapine improved both the circadian organization of the rest-activity cycle and the patient's clinical state. This observation can be conceptualized in terms of the two-process model of sleep regulation. High-dose haloperidol treatment may have lowered the circadian alertness threshold, whereas clozapine augmented circadian amplitude (perhaps through its high affinity to dopamine D4 and serotonin 5HT7 receptors in the suprachiasmatic nuclei). Measurement of the circadian rest-activity cycle may be a useful non-invasive method to follow functional consequences of neuroleptic treatment.

Chaotic behavior of EEG slow-wave activity during sleep

The study of the dynamics of non-linear systems allows the evaluation of the correlation dimension which, in turn, provides an estimate of the number of variables needed to model the process. In such a view, the correlation dimension was calculated for the profiles of the EEG slow-wave activity during sleep obtained from 7 young normal controls and in their corresponding artificial stochastic signals. It was possible to evaluate the complexity of all the real profiles which exhibited an average dimension of 3.76 (SD 0.331), but not that of the artificial control ones. This allows us to conclude that sleep regulation might be considered as a deterministic non-linear process and that the already proposed two-process model of sleep regulation needs to include additional variables with non-linear interactions.

Estimation of the time course of slow-wave sleep over the night in depressed patients: effects of clomipramine and clinical response

The distribution of slow-wave activity during sleep has been analyzed using a method related to the two-process model of sleep regulation. This method is applied to the analyses of data collected from 21 inpatients with unipolar depression who received the antidepressant clomipramine (CMI) in a pulse-loading protocol. CMI infusion was found to redistribute slow-wave activity, producing more concentration in the early part of the night, and also significantly reduced the fluctuation in slow-wave power as a function of time. These measures also distinguished clinical responders from the nonresponders. Drug responders had a significant redistribution of slow-wave activity to the earlier part of the night as compared to nonresponders. This suggests that measures of the distribution of slow-wave activity over the night may represent a good measure of clinical response to antidepressant therapy and have implications for the interpretation of the two-process model and sleep in depression.

Homeostatic sleep regulation in habitual short sleepers and long sleepers.

Homeostatic sleep regulation in habitual short sleepers (sleep episode < 6 h, n = 9) and long sleepers (> 9 h, n = 7) was investigated by studying their sleep structure and sleep electroencephalogram (EEG) during baseline conditions and after prolonging their habitual waking time by 24 h. In each sleep episode, total sleep time was > 3 h longer in the long sleepers than in the short sleepers. Sleep deprivation decreased sleep latency and rapid eye movement (REM) density in REM sleep more in long sleepers than in short sleepers. The enhancement of EEG slow-wave activity (SWA; spectral power density in the 0.75-4.5 Hz range) in non-REM sleep after sleep loss was larger in long sleepers (47%) than in short sleepers (19%). This difference in the SWA response was predicted by the two-process model of sleep regulation on the basis of the different sleep durations. The results indicate that short sleepers live under a higher "non-REM sleep pressure" than long sleepers. However, the two groups do not differ with respect to the homeostatic sleep regulatory mechanisms.

Revisiting the Ehlers and Kupfer hypothesis: The growth hormone cortisol secretion ratio during sleep is correlated with electroencephalographic slow wave activity in normal volunteers

According to the two-process model of sleep regulation, the level of slow wave activity [SWA, sleep electroencephalogram (EEG) power density in the delta range] increases as a monotonic function of waketime prior to sleep initiation (Borb61y 1982; Daan et al 1984). This is referred to as in the model. There is evidence, however, that sleep displaced to the morning hours, despite prolonged waketime prior to sleep, shows a smaller proportion of slow wave sleep (SWS, stage 3 and 4) (Webb and Agnew 1971) and SWA (Dijk et al 1990, 1991) than process S would predict. Among the possible explanations for this are: 1) the suppressive effect of rapid-eye-movement sleep (REMS) upon nonREMS EEG activity (Beersma et al 1990; Brunner et al 1993);

Chaotic behavior of EEG slow-wave activity during sleep

The study of the dynamics of non-linear systems allows the evaluation of the correlation dimension which, in turn, provides an estimate of the number of variables needed to model the process. In such a view, the correlation dimension was calculated for the profiles of the EEG slow-wave activity during sleep obtained from 7 young normal controls and in their corresponding artificial stochastic signals. It was possible to evaluate the complexity of all the real profiles which exhibited an average dimension of 3.76 (SD 0.331), but not that of the artificial control ones. This allows us to conclude that sleep regulation might be considered as a deterministic non-linear process and that the already proposed two-process model of sleep regulation needs to include additional variables with non-linear interactions.

Simulation of daytime vigilance by the additive interaction of a homeostatic and a circadian process.

The two-process model of sleep regulation postulates that a homeostatic and a circadian process underlie sleep regulation. The timing of sleep and waking is accounted for by the interaction of these two processes. The assumptions of two separate processes or of a single process resulting from their additive interaction are mathematically equivalent but conceptually different. Based on an additive interaction, subjective alertness ratings in a forced desynchrony protocol and subjective sleepiness ratings in a photoperiod experiment were simulated. The correspondence between empirical and simulated data supports the basic assumption of the model.

Sleep and cortical temperature in the Djungarian hamster under baseline conditions and after sleep deprivation.

The Djungarian hamster (Phodopus sungorus) is a markedly photoperiodic rodent which exhibits daily torpor under short photoperiod. Normative data were obtained on vigilance states, electroencephalogram (EEG) power spectra (0.25-25.0 Hz), and cortical temperature (TCRT) under a 16:8 h light-dark schedule, in 7 Djungarian hamsters for 2 baseline days, 4 h sleep deprivation (SD) and 20 h recovery. During the baseline days total sleep time amounted to 59% of recording time, 67% in the light period and 43% in the dark period. The 4 h SD induced a small increase in the amount of non-rapid eye movement (NREM) sleep and a marked increase in EEG slow-wave activity (SWA; mean power density 0.75-4.0 Hz) within NREM sleep in the first hours of recovery. TCRT was lower in the light period than in the dark period. It decreased at transitions from either waking or rapid eye movement (REM) sleep to NREM sleep, and increased at the transition from NREM sleep to waking or REM sleep. After SD, TCRT was lower in all vigilance states. In conclusion, the sleep-wake pattern, EEG spectrum, and time course of TCRT in the Djungarian hamster are similar to other nocturnal rodents. Also in the Djungarian hamster the time course of SWA seems to reflect a homeostatically regulated process as was formulated in the two-process model of sleep regulation.

Effects of 12-h sleep deprivation and of 12-h cold exposure on sleep regulation and cortical temperature in the rat

The predictions of the two-process model of sleep regulation were tested by 12-h sleep deprivation (SDEP) and 12-h cold exposure (cold, 4°C) in the rat, both carried out in the 12-h dark period. The analysis was based on recordings of vigilance states, electroencephalogram (EEG) power spectra (0.25–25.0 Hz), and cortical temperature ( T crt). During recovery from SDEP, slow-wave activity (SWA; EEG power density in the 0.75–4.0 Hz band) in non-REM sleep was higher and the number of brief awakenings (nBA) lower than in the corresponding baseline period. During recovery from cold, a moderate and delayed increase in SWA was present. During SDEP, T crt was above baseline, and in some intervals of recovery below baseline. In those intervals waking was also decreased. During recovery from cold, T crt and waking were not affected. It is concluded that 12-h SDEP causes an intensification of sleep, as indicated by the enhanced SWA and the reduced nBA, whereas 12-h cold has only marginal effects. The time course of SWA in both experiments could be accurately predicted by a computer simulation based on the assumptions of the two-process model and the parameter values determined in a previous experiment.

Repeated partial sleep deprivation progressively changes in EEG during sleep and wakefulness.

The effect of repeated partial sleep deprivation on sleep stages and electroencephalogram (EEG) power spectra during sleep and wakefulness was investigated in nine healthy young subjects. Three baseline nights of 8 hours (2300-0700 hours) were followed by four nights with 4 hours of sleep (2300-0300 hours) and three recovery nights of 8 hours (2300-0700 hours). Sleep restriction curtailed sleep stages 1 and 2 as well as rapid eye movement (REM) sleep, but left slow wave sleep largely unaffected. In the first two recovery nights, total sleep time and REM sleep were enhanced, and sleep latency was shortened. Slow wave sleep was increased only in the first recovery night. In accordance with the prediction of the two-process model of sleep regulation, slow wave activity (SWA; spectral power density in the 0.75-4.5-Hz range) in nonrapid eye movement (NREM) sleep increased by approximately 20% in the first night following sleep restriction, remained at this level in the subsequent 3 nights and decreased immediately after the first recovery night. In contrast to these immediate changes, progressive and more persistent changes were seen in the EEG activity of higher frequencies. Thus, activity in the upper delta band tended to gradually increase from night to night during the sleep restriction period, whereas after an initial increase, activity in the theta-alpha band changed in the opposite direction. The progressive changes were also present in the EEG spectra of REM sleep and wakefulness. Because the time course of these changes paralleled the cumulative deficit in REM sleep, they may represent a correlate of REM sleep pressure.

A model of human sleep homeostasis based on EEG slow-wave activity: Quantitative comparison of data and simulations

EEG slow-wave activity (SWA; spectral power in the 0.75–4.5 Hz band) is a function of the duration of prior waking and, thereby, an indicator of sleep homeostasis. We present a model that accounts for both the declining trend of SWA during sleep and for its variation within the successive nonrapid eye movement (non-REM) sleep episodes. The values of the model parameters were estimated by an optimization procedure in which empirical SWA of baseline nights (16 subjects, 26 nights) served as a reference. A sensitivity analysis revealed the model to be quite robust to small changes (±5%) of the parameter values. The estimated parameter values were used to simulate data sets from three different experimental protocols (sleep in the evening or sleep in the morning after prolonged waking, or extended sleep initiated at the habitual bedtime; n = 8 or 9). The timing of the REM trigger parameter was derived from the empirical data. A close fit was obtained between the simulated and empirical SWA data, and even the occasional late SWA peaks during extended sleep could be reproduced. Minor discrepancies suggest indirect or direct circadian influences on SWA. The simulations demonstrate that the concept of sleep homeostasis as proposed in the two-process model of sleep regulation can be refined to account in quantitative terms for empirical data and to predict the changes induced by the prolongation of waking or sleep.

Immediate Early Gene Expression in Brain During Sleep Deprivation: Preliminary Observations

The two-process model of sleep regulation posits that a homeostatic drive to sleep, referred to as Process S, increases with time spent awake. The purpose of this study was to evaluate whether immediate early gene (IEG) expression increases in the brain in proportion to time spent awake, when Process S would be expected to increase. Rats were deprived of sleep by cage tapping, cage rotation and gentle handling beginning at light onset for 45 minutes, 3 hours or 6 hours. At the end of the deprivation periods, deprived animals and an equal number of controls were decapitated, the brains dissected into subregions and frozen. Northern blots were prepared from cortex, thalamus, cerebellum, pons and hypothalamus and hybridized with cDNA probes to five IEG mRNAs; c-fos, c-jun, junB, NGFI-A and NGFI-B. Basal levels of c-fos mRNA were detectable in all brain regions from all animals. Sleep-deprived animals showed higher expression of c-fos mRNA than control animals following 45 minutes and 6 hours of sleep deprivation in all brain regions examined, with the greatest increases observed in the cerebellum. Surprisingly, only the pons and cerebellum showed clear increases at the 3-hour timepoint. In contrast to c-fos, c-jun mRNA was essentially invariant among the animals while junB mRNA was inconsistently elevated. The expression of NGFI-A and NGFI-B was similar to the c-fos pattern but of lesser magnitude.(ABSTRACT TRUNCATED AT 250 WORDS)

Sleep homeostasis in the rat: Simulation of the time course of EEG slow-wave activity

According to the two-process model of sleep regulation, a homeostatic Process S increases during waking and declines during sleep. For humans, the time course of S has been derived from the changes in EEG slow-wave activity (SWA; spectral power density in the 0.75–4.0 Hz range) during sleep. We tested the applicability of the model to sleep in the rat. The simulation was based on the vigilance states for consecutive 8-s epochs of a 96-h experiment in 9 animals. The level of S was made to decrease in epochs of non-REM sleep (NREMS), and to increase in epochs of waking or REM sleep according to exponential functions. By optimizing the initial value and the time constants of S, a close fit between the hourly values of SWA in NREMS and of S was obtained. The biphasic time course of SWA during baseline, its enhancement in the initial recovery period after 24-h sleep deprivation, and its subsequent prolonged undershoot were present in the simulation. We conclude that sleep homeostasis as conceptualized in the two-process model may be a general property of mammalian sleep.

Electroencephalogram Power Density and Slow Wave Sleep as a Function of Prior Waking and Circadian Phase

Human sleep electroencephalograms, recorded in four experiments, were subjected to spectral analysis. Waking prior to sleep varied from 12 to 36 h and sleep was initiated at different circadian phases. Power density of delta and theta frequencies in rapid-eye-movement (REM) sleep and non-REM (NREM) sleep increased monotonically as a function of prior waking. The increase of power density in the theta frequencies contrasts with the reported decrease of theta activity as detected by period-amplitude analysis. Slow wave activity (power density, 0.25-4.0 Hz) in NREM sleep during the first 3 h of sleep did not deviate significantly from the homeostatic process S of the two-process model of sleep regulation. In contrast, visually scored slow wave sleep, stages 3 and 4, deviated from this prediction at some circadian phases. It is concluded that, in accordance with the two-process model of sleep regulation, slow wave activity in NREM sleep depends on prior waking and is not significantly influenced by circadian phase.

Simulation of Human Sleep: Ultradian Dynamics of Electroencephalographic Slow-Wave Activity

The typical declining trend of electroencephalographic (EEG) slow-wave activity (SWA) within a sleep period is represented in the two-process model of sleep regulation by an exponentially decaying process (Process S). The model has been further elaborated to simulate not only the global changes of SWA, but also the dynamics within non-rapid-eye-movement (non-REM) sleep episodes. In this new model, the initial intraepisodic buildup of SWA is determined by the combined action of an exponentially increasing process and a saturation process, whereas its fall at the end of an episode is due to an exponentially decreasing process. The global declining trend of SWA over consecutive episodes results from the monotonic decay of the intraepisodic saturation level. In contrast to Process S in the two-process model, this decay is not represented by an exponential function, but is proportional to the momentary level of SWA. REM sleep episodes are triggered by an external function. The model allows one to simulate the ultradian pattern of SWA for baseline nights as well as changes induced by a prolonged waking period, a daytime nap, a partial slow-wave sleep deprivation, or an antidepressant drug.

Time course of EEG power density during long sleep in humans

In nine subjects sleep was recorded under base-line conditions with a habitual bedtime (prior wakefulness 16 h; lights off at 2300 h) and during recovery from sleep deprivation with a phase-advanced bedtime (prior wakefulness 36 h; lights off at 1900 h). The duration of phase-advanced recovery sleep was greater than 12 h in all subjects. Spectral analysis of the sleep electroencephalogram (EEG) revealed that slow-wave activity (SWA; 0.75-4.5 Hz) in non-rapid-eye-movement (NREM) sleep was significantly enhanced during the first two NREM-REM sleep cycles of displaced recovery sleep. The sleep stages 3 and 4 (slow-wave sleep) and SWA decreased monotonically over the first three and four NREM-REM cycles of, respectively, base-line and recovery sleep. The time course of SWA in base-line and recovery sleep could be adequately described by an exponentially declining function with a horizontal asymptote. The results are in accordance with the two-process model of sleep regulation in which it is assumed that SWA rises as a function of the duration of prior wakefulness and decreases exponentially as a function of prior sleep. We conclude that the present data do not provide evidence for a 12.5-h sleep-dependent rhythm of deep NREM sleep.

Variation in Process S: Effects on Sleep Continuity and Architecture

Given the two-process model of sleep regulation, and the hypothesis that the sleep disorder in depressive illness is a consequence of a deficient Process S, it was predicted that relatively high levels of S would result in enhanced sleep continuity, increased slow-wave sleep (SWS), prolonged rapid-eye-movement (REM) latency, and less REM sleep. These predictions were tested in two studies. In Study 1, the level of Process S (at 0900 h prior to a 3-h sleep episode) was varied by altering the time and duration of prior nocturnal sleep (2400-0300 h, 0300-0600 h, 2400-0600 h). In Study 2, the leve of Process S (at 2400 h prior to an 8-h sleep episode) was varied by studying subjects when they had not napped or had taken 2-h naps beginning at either 1000 or 1900 h. As predicted by the model, SWS varied reliably depending on the level of S at bedrest, as did indices of sleep continuity at night. Contrary to prediction, however, REM sleep was either increased (Study 1) or did not change reliably (Study 2). It is suggested that, contrary to the other aspects of sleep, REM sleep is strongly influenced by circadian and homeostatic processes and that Process S plays a relatively minor role in its regulation.