The Effects of a Short Daytime Nap After Restricted Night Sleep

Transcription

1 Sleep. 19(7): American Sleep Disorders Association and Sleep Research Society The Effects of a Short Daytime Nap After Restricted Night Sleep Mats Gillberg, Garan Kecklund, John Axelsson and Torbjarn Akerstedt Karolinska Institute and National Institute for Psychosocial Factors and Health, Stockholm, Sweden Summary: Eight subjects participated on three occasions in a study investigating the effect of a 30-minute daytime nap opportunity on alertness/sleepiness. The baseline condition was a normal home sleep (7.5 hours, with bedtime at 2300 hours). Sleep during the other two conditions was between 2400 hours and 0400 hours. During one of the two 4-hour conditions, a short nap was allowed (between 1045 hours and 1115 hours). Self-ratings of sleepiness/alertness (Karolinska Sleepiness Scale) were recorded every hour. At 10, 12, and 15 hours, the subjects performed a 28-minute visual vigilance task. Electroencephalograms (EEG) and electrooculograms (EOG) were recorded continuously, including during a lo-minute standardized recording procedure at the beginning of each day. Mean total sleep time during the nap was 19.8 (standard error 2.4) minutes. Compared to baseline, EEGIEOG sleepiness and subjective sleepiness were significantly higher and vigilance performance at 10 hours lower, respectively, after the two short sleeps. The nap brought performance to baseline levels, and subjective sleepiness decreased significantly. It was concluded that the short nap had a clear positive effect on alertness. Key Words: Daytime nap--restricted sleep--performance-eegieog-subjective sleepiness. Sleepiness is an unavoidable consequence of shift work and irregular work hours and severely increases the risk for accidents and mistakes. One countermeasure against sleepiness might be to allow naps during work. Naps have been shown to have positive effects on alertness in several different settings. Naps taken during the day or evening in advance of expected sleepiness, so-called prophylactic naps, have efficiently. counteracted the loss of alertness during the ensuing night whether they have been 1 (1),2 (2-5), or 3 hours (6,7) long. It has also been demonstrated that naps can counteract the effects of sleep deprivation. In a study of our own (1), alertness measured during the early morning was increased by a I-hour nap at 0430 hours. Rodgers et al. (8), however, found only a limited effect of a I-hour nap on overnight performance. Dinges et al. (2,3) have shown that 2-hour naps efficiently counteracted sleepiness after 30, 42, and 54 hours awake, respectively. Positive effects were also demonstrated by Naitoh (9) in an experiment with naps scheduled after 45 or 54 hours awake. Accepted for publication April Address correspondence and reprint requests to Dr. Mats Gillberg, Department of Public Health Sciences, Division for Stress Research, Karolinska Institute, Box 220, S Stockholm, Sweden. 570 Although a nap of several hours' duration seems unrealistic in work situations, a short nap that is, of no more than 30 minutes' duration, might be acceptable. There are studies of repeated short naps, for example, 20-minute naps repeated every 6 hours (10), but very few studies have investigated the effects of single short naps. Lumley et al. (11) compared the effects of naps that were 0, 15, 30, 60, and 120 minutes long, respectively. The naps were allowed at 0900 hours after 1 night without sleep. Alertness was assessed with sleep latency tests at 2-hour intervals for 8 hours following the nap. Three findings emerged: first, all naps except the I5-minute nap were beneficial; second, increasing the nap duration from 60 to 120 minutes did not further increase alertness; and third, the positive effects did not appear until 4 hours after the nap. The purpose of the present experiment was to investigate the effects of a short nap on performance, electrophysiologically defined sleepiness, and subjective sleepiness. Furthermore, the purpose was to simulate the moderate levels of alertness encountered in connection with, for example, shift work, rather than sleep deprivation. In the present case we chose to simulate the daytime sleepiness appearing during an early morning shift.

2 EFFECTS OF A SHORT DAYTIME NAP 571 Subjects and design METHODS Eight male subjects participated (mean age 32 years, range 24-45) in a cross-over repeated measures design with three conditions appearing in a counterbalanced order. The baseline condition was a normal 7- to 7.5- hour (approximate bedtime at 2300 hours) sleep at home, followed by an 8-hour (0800 hours to 1600 hours) test period in the laboratory. The other two conditions involved 4-hour sleeps (between 2400 hours and 0400 hours). Both 4-hour conditions were followed by the same 8-hour test period as for the baseline condition. However, during one of the two 4-hour conditions a 30-minute nap was allowed (between 1045 hours and 1115 hours). In the following the conditions are referred to as BL (baseline), Nap, and NN (no nap). To ascertain that the subjects were in good health and had no sleep problems, they were given a questionnaire. This questionnaire also included the diurnal type scale (DTS) (12). The mean rating of general health ("how has your general health been the last year?"; 5-point scale with 5 = very good) was 4.87 [standard error (SE) 0.12]. Four subjects took no medication at all; three subjects took aspirin occasionally, and one subject used hypnotics "once, or a couple of times per year". The general sleep quality rating ("how is on the average your sleep quality?"; 5-point scale with 5 = very good) was 4.37 (SE 0.18). Mean bedtime during weekends was 2430 hours (SE 30 minutes), and mean sleep duration was 9.2 hours (SE 0.7). The corresponding values for work days were hours (SE 18 minutes) and 7.8 hours (SE 0.4). None of the subjects took naps. The mean coffee consumption was 1.6 cups per day (SE 0.6); there was no consumption of other caffeinated beverages. The subjects varied between 1.14 and 2.71, with a mean of 1.93 and an SE of 0.23, on the Diurnal Type Scale. This scale ranges from 1 to 4; higher values mean more "morningness" and lower values more "eveningness". Procedure For the BL condition the subjects were instructed to go to bed at their habitual bedtime, but not later than 2300 hours, and to rise between 0600 hours and 0630 hours. For the other two conditions the subjects were instructed to go to bed at 2400 hours and to rise at 0400 hours. The home sleeps were monitored by means of subjective sleep diaries (13). The subjects were told not to drink caffeinated beverages after rising, and to spend their time, after awakening and before traveling to the laboratory, reading or doing similar sedentary activities. The participation of the subjects was carefully planned on each of the three occasions so that prior work or leisure activities should not interfere with the experiment. In addition, the subjects were asked to report any deviations from the instructions, for example, failure to fall asleep and sleep according to the above instructions, for which there was a risk, especially in connection with the restricted sleeps. None of the subjects reported such deviations from the instructions. In connection with all three conditions, the subjects arrived at the laboratory at 0800 hours for instructions and application of electrodes. The experiment proper started at 0900 hours with a 1O-minute (5 minutes with open eyes, followed by 5 minutes with closed eyes) standardized electroencephalographic (EEG)/electrooculographic (EOG) recording procedure, the Karolinska alertness test (KAT, see below), and ended at 1600 hours with removal of electrodes. Subjective sleep quality The sleep quality measure extracted from the sleep diaries was an index calculated as a mean of the following 5-point items (5 = best quality): "how was your sleep?", "unrestful sleep", "did you get enough sleep", "difficulties falling asleep", "ease of awakening", and "well rested". Test blocks and nap Three identical 45-minute test blocks were presented. They were scheduled at 0945 hours, 1145 hours, and 1500 hours, respectively. The test blocks contained a 1O-minute reaction time (RT) task and a 28-minute visual vigilance task. Subjective ratings were obtained before and after the RT task and after the vigilance task. The nap during the Nap condition was scheduled between 1045 hours and 1115 hours. There was exactly 30 minutes between lights out and awakening. A 30-minute period was then allowed for sleep inertia (14) to dissipate before the start of the next test block (at 1145 hours). Reaction time task The RT task was 10 minutes long and was presented on a handheld computer. Sixteen signals per minute were presented at random intervals. Unfortunately, approximately half of the recordings were lost due to technical failure and therefore the data could not be analyzed. Sleep. Vol. 19. No.7, 1996

3 572 M. GILLBERG ET AL. Vigilance task The 28-minute vigilance task was presented on a computer display and contained 60 signals, occurring at random time intervals. The temporal distribution of signals was similar for each. quarter of the test. The task was described in more detail in a previous paper (15). Subjective ratings of sleepiness The scale used was the 9-point (9 = highest level of sleepiness) Karolinska sleepiness scale (KSS) (16). In addition to the ratings obtained in connection with the test blocks, ratings were obtained every hour, beginning at 0800 hours and ending at 2200 hours. Ratings from 1700 hours and onward were given outside the laboratory. Electroencephalography and EOG Electroencephalography (leads: C3-A2 and 02-P4) and EOG (left eye) were recorded continuously during each experimental condition. These recordings were scored visually in order to derive an electrophysiological measure of sleepiness during both the KAT and the vigilance tests. During the KAT the subjects sat in a reclining position and were instructed to relax, but try to remain awake, and first look at a black dot in the ceiling for 5 minutes and then to close their eyes for 5 minutes. The instructions were given via an intercom. The EEGIEOG recordings were considered to show signs of sleepiness if one, or more, of the following activities was present: alpha activity (8-12 Hz), theta activity (4-8 Hz), slow rolling eye movements (defined as > I-second slow, rolling excursions of the EOG of at least 100 f.lv amplitude), or regular sleep (stage 1 activity or "deeper"). Each 30-second epoch was visually scored, could contain from 0 to 100% signs of sleepiness, and was scored in six intervals (0, 10, 30, 50, 70, and 90%). If no EEGIEOG signs of sleepiness were present, 0 was scored. If signs of sleepiness were present, but for less than 20% of an epoch, 10% was scored. Thirty percent was scored for signs of sleepiness in the range of 20-40% of an epoch, and so on, for the remaining 20% intervals, up to 100%. In addition, the recordings outside the nap and the tests were scanned to ascertain the amount of unauthorized sleep. The nap sleep was scored according to standard procedures (17). The subjects were studied in pairs but spent their time in separate, sound-insulated rooms containing a bed and a reclining chair in which they sat during the KAT and while performing the tasks. The two rooms were parts of a sound-insulated small apartment also containing a living roomlkitchen and a toilet. The subjects were served a light meal (sandwiches and mineral water) on three occasions: before the KAT, after the nap, and before the third test block. During these meals the subjects sat together. Between tests the subjects were allowed to read and listen to music. Clocks and other time cues were not allowed. Statistical analyses If not otherwise stated, repeated measures analyses of variance (ANOV As) were used to evaluate the experimental effects. Because such analyses are sensitive to violations of the requirement of equal variances, p values are given after the Huynh-Feldt epsilon (E) correction (18). E is a measure of the degree to which the requirements of equal variances are violated; it is used to adjust (downwards) the degrees of freedom (df) before performing the F test. Although the F tests were performed with adjusted df, the original df are maintained in the ANOV A presentations. The 5% alpha level was adopted. Post-hoc comparisons were made with the Newman-Keuls test. Prior night sleep RESULTS Data extracted from the sleep diaries show that the subjective sleep durations (calculated as time in bed minus time to fall asleep and time awake in connection with awakenings) varied in the expected manner (means and SEs in hours): 6.73 ± 0.16 for BL, 3.75 ± 0.06 for Nap, and 3.71 ± 0.06 for NN. The times of awakening varied significantly (means and SEs: 0615 hours ± 7 minutes for BL, 0411 hours ± 8 minutes for Nap, and 0412 hours ± 9 minutes for NN). Finally, the subjective sleep qu,ality of the two curtailed sleeps was significantly lower compared to the BL sleep [means and SEs: BL = 4.0 ± 0.25, Nap = 3.2 ± 0.22, and NN = 2.8 ± 0.29; p < 0.002; F(2,14) = 10.20; E = 0.98]. Effect of sleep curtailment/extended wake time on alertness Subjective alertness rated after arrival to the laboratory but before the start of the experiment (the mean of ratings at 0800 hours and 0900 hours) showed no significant differences between conditions [BL = 4.56, Nap = 4.75, and NN = 5.13; P < 0.160; F(2,14) = 2.285; E = 0.859]. Results from the KAT with open eyes showed low levels of sleepiness and no differences between conditions. KAT with closed eyes, on Sleep. Vol. 19. No

4 EFFECTS OF A SHORT DAYTIME NAP 573 TABLE 1. Percentage of time with EEGIEOG sleepiness during the Karolinska alertness test (KAT) BL Nap NN F(2,14) P (E) Open eyes 0 (0) 3.7 (2.3) 6.3 (3.7) 1.9 ns (1.0) Closed eyes 6.5 (4.2) 48.2 (12.6) 44.3 (15.2) (1.0) Data shown are means and SEs. Results are from one-way AN OVAs. the other hand, showed markedly higher levels of sleepiness when prior sleep was curtailed (Table O. Sleep parameter TABLE 2. TST Stage 1 Stage 2 Stage 3 Stage 4 REM Latency stage 1 Nap sleep parameters Duration 19.8 (2.4) 7.1 (1.2) 9.0 (1.5) 2.4 (1.3) 1.3 (0.7) (2.4) Range Data (means and SEs) are in minutes. TST, total sleep time; REM, rapid eye movement. Nap sleep parameters Table 2 shows that all subjects fell asleep during the nap opportunity and that the sleep obtained consisted mainly of stage 1 and stage 2 sleep. Effects of the nap on performance and alertness There were clear significant effects of conditions on vigilance performance [F(2,14) = 7.4; P < 0.008; E = 0.91]. Neither changes across the three tests nor the interaction were significant [F(2,14) = 0.83 and F(4,28) = 1.75, respectively] (Fig. 1). Pairwise comparisons revealed that there were significantly (p < 0.05) fewer detected signals (hits) during the NN condition compared to both of the other two conditions. Comparing results from the first test across conditions shows significantly fewer hits after both the Nap and NN conditions compared to BL. Pairwise comparisons of the results on the second test, that is, the one performed after the nap, showed significantly fewer hits during the NN condition. compared to the other conditions. During the third test, only the NN and BL conditions differed from each other. Because the main issue concerned the Nap and NN conditions and the changes from before to after the nap, we carried out an additional, restricted ANO VA using only this 2 X 2 design. The analysis shows a significant effect of condition [F(I,7) = 9.2; P < 0.019; E = 1.00] and a significant interaction [F(l,7) = 10.9; P < 0.013; E = 1.00], indicating a differential, positive effect of the nap. ElectroencephalographiclEOG signs of sleepiness during tests Only the effect of condition was significant (Table 3). Pairwise comparisons showed that the BL and NN conditions differed significantly (p < 0.05) during the first test as well as for all tests combined. Although data suggest that the nap decreased sleepiness, this is not supported by the statistical analyses. As for the vigilance results, a restricted 2 X 2 ANOV A was ap- plied to the data in Table 3. Neither the main effects nor the interaction, however, were significant [F(l,7) conditions = 3.6; F(I,7) test number = 1.8; F(1,7) interaction = 1.4]. Subjective sleepiness In order to obtain an integrated measure of subjective sleepiness before and after the nap (and the corresponding period for the other two conditions), individual averages of all KSS ratings recorded from 0800 hours until the nap and of ratings recorded after the nap until after the third vigilance test, respectively, were calculated. The data and results from ANOV A are given in Table 4. The analysis showed a significant variation across conditions and a significant interaction between condition and time (i.e. before and after the nap). A closer inspection by means of pairwise comparisons showed that sleepiness before the nap (or the corresponding period) was significantly (p < 0.05) lower during the BL condition compared to both of the other conditions. Comparing sleepiness after this period revealed sig- ~ s:. '0 Z E ::J c:: Baseline ' '.'.,8.,.,.,. ill-,.,.,.,.,.,.,. '.'.'.'.'.'.'.' 86 No nap I range of s.e. 30+L~~~-r~~~-'--~r-~ time of day FIG. 1. The number of hits on the vigilance task. Shown are the means of eight subjects. The range of the SE: (i.e. the largest and the smallest SE) is shown to the left. Sleep, Vol. 19, No.7, 1996

5 574 M. GILLBERG ET AL. TABLE 3. Percentage of time with EEGIEOG signs of sleepiness during the vigilance task Test Condition First Second Third Results from two-way ANOV A BL 0.7 (0.5) 5.0 (4.9) 9.3 (7.9) F(2,14) conditions = 5.2; P < 0.04; E = 0.74 Nap 20.2 (8.2) 11.2 (6.1) 11.4 (6.7) F(2,14) test no. = 0.3; ns; E = 0.91 NN 25.4 (11.5) 22.3 (8.4) 22.6 (8.8) F(4,28) interaction = 0.9; ns; E = 0.48 Data shown are means and SEs. nificantly (p < 0.05) higher levels during the NN condition compared to both the BL and the Nap conditions. The subjective ratings obtained outside the laboratory during the afternoon and evening (between 1700 hours and 2200 hours) were averaged to demonstrate the long-term effects of prior sleep manipulation and of the nap. The mean level of sleepiness was lower during the BL condition compared to the other two conditions [means and SEs: BL = 4.7 (0.3), Nap = 5.3 (0.3), NN = 5.3 (0.3)]. However, the variation between conditions was not significant (ns) (ANOV A: F(2,14) = 2.53, ns, E = 0.61). DISCUSSION Restricting the period of sleep to the first 4 hours of the night resulted in subjective total sleep times that were approximately 3 hours and 45 minutes, compared to 6 hours and 45 minutes for the baseline night. The two shorter nights also showed lower subjective quality. The restricted sleeps both had clear negative effects on daytime alertness, as measured by performance and electrophysiological variables. Vigilance performance on the first vigilance test was approximately 20% lower after the two shorter sleep periods compared to that after a normal sleep duration. When the subjects were asked to relax and close their eyes for 5 minutes, there were electrophysiological signs of sleepiness for about 50% of this period if prior night sleep had been 3 hours and 45 minutes, whereas such signs were present for only 10% of the recording period after the baseline night. Subjective alertness was also lower when prior sleep was curtailed. EEGIEOG signs of sleepiness during the vigilance test also tended to be higher after the restricted sleeps. Interestingly, there were no differences in subjective alertness between conditions immediately after arrival to the laboratory (0800 hours to 0900 hours), but such were revealed later, presumably when the subjects were placed in the relaxing and monotonous experimental situation. The effects of sleep curtailment observed in the present experiment are in agreement with similar findings in other studies (19-22). When discussing the effects of sleep curtailment one must, however, keep in mind that sleep was not polygraphically recorded but assessed by subjective methods. Hence, we do not know the exact duration or architecture of sleep. In our experience there is, however, a fair agreement between sleep diary data and polygraphically recorded sleep (13). Also, the lower morning alertness after the two curtailed nights clearly indicates that subjects actually had shorter sleeps. All subjects managed to fall asleep during the allotted time, and the effects on vigilance performance and subjective alertness were clear and positive. The nap brought vigilance performance to levels nearly identical to those observed during the baseline condition. Similar results were observed for subjective sleepiness. EEGIEOG signs of sleepiness during the vigilance test, on the other hand, were not as clearly affected by the nap. As indicated by the vigilance results, the positive effects of the nap were observed starting 30 minutes after the nap. It is obvious from Fig. 1 that the positive effect on performance had partly dissipated by the third vigilance test. However, because the nap condition did not differ significantly, either from the baseline or from the no-nap condition during the third test, it seems that some of the effect remained. Also, the subjective ratings integrated for the post-nap period suggest that the positive effect persisted until the subjects left the laboratory. Finally, the integrated ratings obtained outside the laboratory (1700 hours to 2200 hours) revealed no significant dif- TABLE 4. Mean levels Karolinska sleepiness scale (KSS) ratings obtained before and after the nap (or the corresponding period for the other conditions). Condition BL Nap NN Data shown are means and SEs. Before nap 5.3 (0.5) 6.2 (0.5) 6.3 (0.6) After nap 5.8 (0.2) 5.7 (0.5) 6.7 (0.6) Results from two-way ANOV A F(2,14) condition = 6.0; P < 0.01; E = 1.0 F(l,7) before/after = 0.6; ns; E = 1.0 F(2,14) interaction = 4.0; P < 0.04; E = 0.99 Sleep, Vol. 19, No.7, 1996

6 EFFECTS OF A SHORT DAYTIME NAP 575 ferences, although there was a trend towards lower alertness after the two restricted nights. The alerting effects of the nap were obtained despite the fact that the mean nap duration was only 20 minutes and that it consisted mainly of stage 1 and 2 sleep. All subjects increased their performance from before to after the nap. Interestingly, because the minimum nap duration was 11 minutes (with no slow wave sleep), this suggests that even such short naps may have recuperative effects. As mentioned earlier, the positive effect of the nap was most prominent when measured 30 minutes after it ended. Lumley et al. (11), on the other hand, concluded that the positive effect of their 30-minute nap did not occur until alertness was measured 4 hours after the end of the nap. In their case, however, the nap was given at 0900 hours after 1 night without sleep (11). At that time the subjects were very sleepy, as indicated by the 1.8-minute mean sleep latency. In addition, because they were allowed to sleep for 30 minutes after sleep onset, a more severe sleep inertia could be expected. In conclusion, the alerting effects of a short daytime nap are obvious. A short nap will not necessarily be as successful if given during the night, when there is longer prior wakefulness and the circadian rhythm is close to its trough. In another study (23) we found no effect of a 30-minute night nap opportunity (resulting in 18.7 minutes of sleep) whether the nap was scheduled at 0100 hours or 0330 hours. Similarly, Saito and Sasaki (24) could not demonstrate positive effects of a 30-minute nap opportunity (resulting in 20.5 minutes of sleep) scheduled at 0300 hours. A nighttime nap may require longer duration for effect, as evidenced by our earlier study (1) showing alerting effects of a I-hour nighttime nap. Hence, the effect of a short nap may depend on the circadian phase and on the amount of prior wakefulness. Acknowledgement: This study was made possible through a grant from the Swedish Work Environment Fund. REFERENCES 1. Gillberg M. The effects of a I-hour nap on early morning performance. Bioi Psychol 1984;19: Dinges DF, Orne MT, Whitehouse WG, Orne EC. Temporal placement of a nap for alertness: contributions of circadian phase and prior wakefulness. Sleep 1987;10: Dinges DF, Whitehouse WG, Orne EC, Orne MT. The benefits of a nap during prolonged work and wakefulness. Work Stress 1988;2: Lavie P, Weier B. Timing of naps: effects on post-nap sleepiness levels. Electroencephalogr Clin Neurophysiol 1989;72: Bonnet MH. The effect of varying prophylactic naps on performance, alertness and mood throughout a 52-hour continuous operation. Sleep 1991;14: Sugerman JL, Walsh JK. Physiological sleep tendency and ability to maintain alertness at night. Sleep 1989;12: Schweitzer PK, Muehlbach MJ, Walsh JK. Countermeasures for night performance deficits: the effect of napping or caffeine on continuous performance at night. Work Stress 1992;6: Rogers AS, Spencer MB, Stone BM, Nicholson AN. The influence of a 1 h nap on performance overnight. Ergonomics 1989;32: Naitoh P. Circadian cycles and restorative power of naps. In: Johnson LC, Tepas DI, Colquhoun Wp, Colligan MJ, eds. Biological rhythms: sleep and shift work. New York: Spectrum Publications, 1981 : Naitoh P, Kelly TL, Babkoff H. Napping, stimulant, and fourchoice performance. In: Broughton RI, Ogilvie RD, eds. Sleep, arousal, and performance. Boston: Birkhiiuser, Lumley M, Roehrs T, Zorick F, Lamphere J, Roth T. The alerting effects of naps in sleep-deprived subjects. Psychophysiology 1986;23: Torsvall L, Akerstedt T. A diurnal type scale. Scand J Work Environ Health 1980;6: Akerstedt T, Hume K, Minors D, Waterhouse J. The meaning of good sleep: a longitudinal study of polysomnography and subjective sleep quality. J Sleep Res 1994; Dinges DF, Orne EC, Evans FJ, Orne MT. Performance after naps in sleep-conducive and alerting environments. In: Johnson LC, Tepas DI, Colquhoun WP, Colligan MJ, eds. Biological rhythms: sleep and shift work. New York: Spectrum Publications, 1981: Gillberg M, Kecklund G, Akerstedt T. Relations between performance and subjective ratings of sleepiness during a night awake. Sleep 1994;17: Akerstedt T, Gillberg M. Subjective and objective sleepiness in the active individual. Int J Neurosci 1990;52: Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Bethesda, MD: U.S. Department of Health, Education and Welfare, Public Health Servi,\e, Huynh, H, Feldt LS. Estimation at the box 'correction for degrees of freedom for sample data in randomised block and split-plot designs. J Educ Stat 1976;1: Carskadon MA, Dement WC. Cumulative effects of sleep restriction on daytime sleepiness. Psychophysiology 1981;18: Carskadon MA, Dement WC. Nocturnal determinants of daytime sleepiness. Sleep 1982;5(Suppl):S Tilley AI, Wilkinson RT. The effects of a restricted sleep regime on the composition of sleep and on performance. Psychophysiology 1984;21: Gillberg M, Akerstedt T. Sleep restriction and SWS suppression: effects on day time alertness and night time recovery. J Sleep Res 1994;3: Gillberg M, Kecklund G, Akerstedt T. Sleepiness and performance of professional drivers in a truck simulator--comparisons between day and night driving. J Sleep Res 1996;5: Saito Y, Sasaki T. Sleepiness and fatigue during early morning hours after a 30 minute nocturnal nap. Shiftwork Inti Newsl 1995;12:53. Sleep, Vol. 19, No.7, 1996