6. H. M. Wisniewski and A. B. Keith, Ann. Neurol. 1, 144 (1977). 7. A. Ridley, Z. Immun. Allergieforsch. 125, 173 (1963); M. B. A. Oldstone and F. S. Dixon, Am. J. Pathol. 52, 251 (1968). 8. I. Grundke-Iqbal, H. Lassmann, H. M. Wis- niewski, Arch. Neurol. (Chicago) 37, 651 (1980). 9. E. C. Reinherz, C. Moretta, M. Roper, J. M. Breard, M. S. Mingari, M. D. Cooper, S. F. Schlossman, J. Exp. Med. 151, 969 (1980). 10. U. Traugott, C. S. Raine, S. H. Stone, J. Chiba, E. Shevach, J. Neuropathol. Exp. Neurol. 43, 320 (Abstr.) (1981). 11. J. S. Greenspan, G. A. Gutmann, I. L. Weiss- man, N. Talal, J. Clin. Immunol. Immuno- pathol. 3, 16 (1974). 12. J. L. Turk, Immunology 5, 478 (1962). 13. R. J. Scheper, T. C. M. Th. Van Maarsseveen, Recent research has lent credence to the hypothesis that sleep and in particu- lar slow-wave sleep (SWS) (1), is a re- covery period for daily metabolism (2, 3). Evidence in support of this theory includes the synchrony of growth hor- mone release with SWS in humans (4), the suggestion that optimum conditions for anabolism prevail during sleep (5), and studies showing SWS duration to be proportional to preceding wakefulness (6). Although many other studies (7) have yielded supportive evidence for the theory, the prediction that daytime exer- cise would increase SWS has produced equivocal results (8). Possible reasons Slow-wave sleep M JEZUEJVA Awake Stage Stage Stage Stage REM 1 2 3 4 sleep 480 Eg 360 0:22 240 Control 1 2 3 4 Nights after extended marathon Fig. 1. Total sleep time (mean ± standard error of the mean) and mean number of min- utes spent in each sleep stage on control nights, after the 92-km marathon, and on three subsequent nights. A. C. H. M. Van Dintehr-Janssen, Cell. Im- munol. 53, 19 (1980). 14. C. M. Elboim, C. L. Reinisch, S. F. Schloss- man, J. Immunol. 118, 1042 (1977). 15. J. W. Prineas and R. G. Wright, Lab. Invest. 38, 409 (1978); D. H. Snyder, M. P. Valsamis, S. H. Stone, C. S. Raine, J. Neuropathol. Exp. Neurol. 34, 209 (1975). 16. C. S. Raine, L. C. Scheinberg, J. M. Waltz, Lab. Invest., in press. 17. We thank S. Swartwout and E. Swanson for expert technical assistance and A. Geoghan and M. Palumbo for typing the manuscript. Support- ed in part by National Multiple Sclerosis Society grant RG 1001-D-4 and by Public Health Service grants NS 08952, NS 07098, and NS 11920. 17 July 1981; revised 21 September 1981 for these conflicting findings include the variable fitness of the subjects tested (9), the time during the day at which the exercise is performed (10), and the abso- lute amount of exercise (11). The abso- lute amount of exercise is relevant since it is the increase in energy expenditure during exercise over and above basal metabolism that would be expected to influence the amount of SWS. To evalu- ate the theory that SWS is a recovery process and to resolve the question of the effect of exercise on sleep, an experi- ment was carried out in which the sleep patterns of six subjects were studied after a 92-km marathon. We thought that this extreme event would highlight the effect of a large increase in energy ex- penditure on sleep. All subjects (age, 18 to 26 years; mean age, 21.7 years) slept for two nonconsec- utive nights (with two intervening nights) in the sleep laboratory 2 weeks before the marathon. The first of these was to allow for the "first-night effect" (12) and was not recorded. The second of these was used as a prerace baseline level. Sleep patterns were recorded on the night of the marathon (night 1) and for the subsequent three nights (nights 2 to 4); they were recorded again 2 weeks after the marathon as a postrace control. Of the six recorded nights, the only day on which any specific exercise had been performed was that of the extended mar- athon (13). None of the subjects were taking medication, and they did not drink alcohol or coffee on the days of the study. Sleep recordings were made in the standard manner and were scored blind by two trained scorers according to stan- dard criteria (14). All six subjects had previously completed several standard marathons over the preceding year, and three had in previous years completed this extended marathon. Five of the six subjects were tested for treadmill maxi- mum aerobic power (V0, ma) and degree of fitness 3 weeks before the marathon. Lactic acid turn point (15) was over 70 percent of VO, max for all but one of these subjects, indicating -a high state of fit- ness. The range of Vo, mlx for these five subjects was 3.56 to 4.07 liters (55.8 ± 2.2 ml per kilogram of body weight per minute, mean ± standard er- ror of the mean). The marathon started at 0600, and the subjects required be- tween 8'/2 and 103/4 hours to complete the 92 km (average speed, 10.7 to 8.6 km/ hour). Body mass of the runners de- creased (despite considerable fluid in- take during the marathon) from 70.2 ± 2.6 to 68.1 ± 2.7 kg [intrasubject com- parisons, t (5) = 4.20, P < .01], and rec- tal temperature increased from 37.6° ± 0.15° to 39.10 ± 0.160C [t (5) = 13.33, P < .0011. Environmental wet- and dry-bulb tem- perature ranges during the race were 10.2° to 19.2°C and 10.40 to 26.20C, re- spectively, and wind velocity ranged from 0.2 to 3.5 m/sec. The analysis of the sleep records showed no significant differences be- tween the 2-week pre- and 2-week post- marathon nights; therefore, the mean of- these two recordings is used as the base- line sleep values. The baseline values were similar to those obtained in norma- tive studies of males of similar age (16). Total sleep time increased significant- ly over control times on each of the four nights after the marathon [F(4, 20) = 21.3, P < .05] (Fig. 1). Wakeful- ness was greatest on the night of the marathon, perhaps because of muscle and blister pains; this result could ex- plain why the longest sleep period occurs on the second night after the marathon. Subjective sleep ratings for the seven laboratory nights showed that four of the six subjects reported having slept best on the night 2. 50' - 40 I 30 E 20 10 I + I I OI Stage 4 U Stage 3 ii Control Night 1 Night 2 Night 3 Night 4 Nights after extended marathon Fig. 2. Slow-wave sleep as a percentage of the total night's sleep on control nights, the night after the 92-km marathon, and on three subse- quent nights. 0036-8075/81/121 1-1253$01 .00/0 Copyright © 1981 AAAS Slow-Wave Sleep: A Recovery Period After Exercise Abstract. Sleep recordings were carried out on athletes on four successive nights after completing a 92-kilometer road race. Significant increases in total sleep time and slow-wave sleep were found after this metabolic stress. The results show a definite exercise effect on sleep and support sleep-restoration hypotheses. 1 253 SCIENCE, VOL. 214, 11 DECEMBER 1981 on August 22, 2014 www.sciencemag.org Downloaded from on August 22, 2014 www.sciencemag.org Downloaded from