Effects of exercise on sleep J. M. WALKER, T. C. FLOYD, G. FEIN, C. CAVNESS, R. LUALHATI, AND I. FEINBERG Veterans Administration Hospital, and University of California, San Francisco, California 94121 WALKER, J. M., T. C. FLOYD, G. FEIN, C. CAVNESS, R. LUALHATI, AND I. FEINBERG. Effects of exercise on sleep. J. Appl. Physiol. : Respirat. Environ. Exercise Physiol. 44(6): 945-951, 1978. -We tested the hypothesis that EEG sleep stages 3 and 4 (slow-wave sleep, SWS) would be increased as a function of either acute or chronic exercise. Ten distance runners were matched with 10 nonrunners, and their sleep was recorded under both habitual (runners running and nonrunners not running, 3 nights) and abruptly changed (runners not running and nonrunners running, 1 night) con- ditions. Analyses of both visually scored SWS and computer measures of delta activity during non-rapid eye-movement (NREM) sleep failed to support the SWS-exercise hypothesis. The runners showed a significantly higher proportion and a greater absolute amount of NREM sleep than the nonrunners. The runners showed less rapid eye-movement activity during sleep than the nonrunners under both experimental condi- tions, indicating a strong and unexpected effect of physical fitness on this measure. Modest afternoon exercise in nonrun- ners was associated with a strong trend toward elevated heart rate during sleep. Mood tests and personality profiles revealed few differences, either between groups or within groups, as a function of exercise. distance running; slow-wave sleep; computer analysis; REM density; mood; personality OSWALD (33) and Roffwarg et al. (35) were among the first to propose that the phase of sleep characterized by dense, high-voltage, slow EEG activity (stages 3 and 4 in the Dement and Kleitman (14) classification system) would be increased with exercise. Empirical studies of this question have produced conflicting results. In as- sessing this literature, one must consider several meth- odological issues. These include the nature and quanti- fication of the exercise, whether it represents a drastic change in activity level (which might produce a stress effect) and, on a more prosaic level, whether the sleep records have been scored without knowledge of experi- mental conditions. Table 1 summarizes the investigations of the effects of exercise on human stage 3-4 sleep (also called slow- wave sleep, SWS) reported thus far. Six studies reported negative findings and two of the three studies that reported positive results suffered from serious method- ological defects, lacking both “blind scoring” and quan- tification of exercise. Of the negative investigations, those of Horne and Porter (29, 30), Webb and Agnew (43), and Zir et al. (45) seem to provide particularly strong evidence, because each employed a repeated- measures design involving quantified exercise with an adequate number of subjects. The strongest study with positive findings is that of Shapiro et al. (38), but it was based upon only two subjects. These subjects showed graduated increases in SWS after daily progressive increases in exercise calibrated according to estimated vo 2max. Two studies in animals are often cited as supporting the SWS-exercise hypothesis (28, 31). However, neither of these studies showed change in amount of SWS as a result of exercise; rather they suggested that exercise may alter the temporal distributions of non-rapid eye- movement (NREM) and rapid eye-movement (REM) sleep. More recently, Boland and Dewsbury (9) reported increased amounts of NREM sleep in 4-h periods after forced wheelrunning in the rat. Both theoretical and practical considerations led us to carry out a further test of the SWS-exercise hypothe- sis. It is of considerable theoretical interest to determine whether sleep primarily functions to promote energy- conservation (2, 6, 41) or somatic restoration (1) or whether it acts to reverse certain of the effects of wake- fulness on the brain (16, 20). It is known that SWS is increased as wakefulness is extended (7, 42); this in- crease could result from prolongation of waking brain activity, or instead be caused by the physical activity associated with waking. The SWS-exercise hypothesis is also of considerable practical importance. Sleep studies permit nonintru- sive, empirical investigations of hitherto unexplored aspects of brain electrophysiology. In many such stud- ies, the sleep patterns of institutionalized patients with functional or organic disorders of the central nervous system are compared to those of ambulatory normal subjects. Differences in activity level would confound these comparisons. Studies of the effects of psychoactive drugs, to which the sleep EEG is remarkably sensitive, raise similar problems since changes in physical activ- ity are often produced by the drugs being studied. Because our laboratory is involved in both kinds of investigation, and because we are especially interested in whether sleep serves mainly the brain or the body, we studied sleep patterns in long-distance runners and matched control subjects. Our main interest was in comparing the effects of chronically different levels of activity on sleep. In addition, we studied the effect of an acute change in exercise level on a single night’s sleep. METHOD Subjects Following a screening procedure, 10 male cross-coun- try (or distance) runners were selected from a commu- by 10.220.33.6 on July 9, 2017 http://jap.physiology.org/ Downloaded from