The calculation of the correlation dimension (D2) was applied to the study of the profiles of EEG slow-wave activity in nine narcoleptic subjects and nine sex- and age-matched control subjects who, following a baseline night recording, were maintained on 16 h of diurnal sleep deprivation and, thereafter, submitted to a-32-h bed rest protocol. The reversibility test allowed us to reject the null hypothesis that the time series considered in our study were generated by a static transformation of a linear Gaussian random process. Similarly, all profiles showed a positive largest Lyapunov exponent. Finally, the computation of D2 showed an average value of 5.27 (0.68 S.D.) in normal controls and 4.05 (1.49 S.D.) in narcoleptic patients (p=0.067). Four of the narcoleptic patients showed values of D2 lower than 4, this was never observed in the normal controls (p=0.0294). This study indicates that the mechanism of sleep–wake regulation in narcolepsy shows a somewhat lower degree of complexity as compared to normal controls. In particular, these data seems to confirm the already suggested different and simpler coupling between the homeostatic process of sleep regulation and the circadian and ultradian drives to sleep that occurs in bed rest condition in this disease.
The two-process model of sleep regulation was proposed by Borbély (1982) which postulates the existence of a sleep-dependent process, termed Process S, and of a sleep-independent circadian process, termed Process C. In this model, the level of Process S, as reflected by the EEG slow-wave activity, corresponds to the sleep-related aspect of sleep propensity. Sleep onset is triggered when S approaches an upper threshold and awakenings occurs when S reaches a lower threshold (Daan et al., 1984, Borbély et al., 1989 and Ackermann et al., 1993). Finally, in this model a simple additive interaction between the two processes is postulated.
In recent studies, the calculation of the correlation dimension (D2) of the profiles of EEG slow-wave activity during sleep, in normal subjects, allowed us to conclude that sleep regulation might be considered as a deterministic non-linear process with an average dimension above three. Moreover D2 does not show significant changes across consecutive nights, in the same subject, and does not seem to change significantly with age in children and young adults (Ferri et al., 1996 and Ferri et al., 1998). The implications of these results are that a number of variables higher than three with non-linear interactions are needed in order to model sleep regulation and the time course of the EEG slow-wave activity during sleep.
This kind of approach has not yet been applied to pathological sleep; for this reason, we decided to analyze the profiles of EEG slow-wave activity in a group of narcoleptic subjects who participated in a previous study in which their sleep was recorded for 32 h under bed rest conditions, following 16 h of diurnal sleep deprivation (Nobili et al., 1995). The main conclusion of such a first study was that EEG slow-wave activity profiles of narcoleptic patients show in this experimental condition a reduction of the strength of circadian and circasemidian mechanisms while the ultradian drives to sleep seem to be more active thus resulting in a less complex wake-sleep distribution (Nobili et al., 1996); thus, the hypothesis tested in the present study is that such profiles show a lower degree of complexity, compared to those of normal controls, as reflected by their D2.
2. Subjects and method
2.1. Subjects
In this study we reanalized the profiles of EEG slow-wave activity during sleep of 9 narcoleptic patients who participated in a previous study (Nobili et al., 1995). The patients (seven males and two females) were aged 20–55 years and had been chosen on the basis of the presence of at least two sleep onset REM–sleep episodes at the Multiple Sleep Latency test (MSLT), cataplectic episodes and excessive diurnal sleepiness, reflected by a mean sleep latency at the MSLT of 5 min or less. Also their genetic status was studied and showed a DR2 positivity in the HLA region. None of the patients was taking drugs at the moment of the recording and in the three previous weeks. Nine healthy age- and sex-matched paid volunteers were also included in this study after having obtained their informed consent.
Fig. 1 shows, as examples, the profile of the EEG slow-wave activity during the 32-h period considered in this study in a normal control (top) and a narcoleptic patient (bottom). Even the simple visual inspection of these profiles allows us to detect important differences between the two subjects; in fact, the normal profile seems to be characterized by two major delta band pulses during the two night periods. On the contrary, the narcoleptic patient shows only a major peak at the beginning of the first night which is followed by a regular sequence of peaks smaller than the first and without important differences between them. This aspect was somewhat repetitive in each corresponding group (patients and controls); a more detailed statistical analysis of the delta peaks can be found in the previous work (Nobili et al., 1995).
