Perceptual and Motor Skills, 1991, 73, 243-252.
O Perceptual and Motor
Skills 1991
D O E S TRAMPOLINING A N D ANAEROBIC PHYSICAL FITNESS AFFECT SLEEP? ' JOACHIM BUCHEGGER, REINER FRITSCH, ALFRED MEIER-KOLL AND HARTMUT RIEHLE2 Universitat Konstanz Summary.-The structure of nocturnal sleep of 16 volunteers, participating in the anaerobic sports of trampolining, dancing, and soccer, was monitored by means of polygraphic recordings. Since trampolining requires the acquisition of unfamiliar patterns of motor coordination, it can be considered as a special form of motor learning, whereas the acquisition of motor skills specific for dancing and soccer can be linked with motor patterns of normal biped locomotion. According to this view, an experimental group of 8 volunteers was formed; they participated in a training course of trampolining. In addition, a control group of 8 subjects was recruited, who engaged in one of the other two anaerobic sports. Subjects who had acquired new motor skills during a 13-wk. program in trampolining showed a statistically significant increase in REM-sleep. By contrast, the 8 subjects of the control group showed no considerable changes in REM-sleep. This suggests that efforts in acquiring new and complex motor patterns activate processes specifically involved in the generation of REM stage during nocturnal sleep.
There is a considerable body of evidence that sleep can be affected by physical fitness and exercise. I n their study of the effects of physical exercise on human sleep Baekeland and Lasky (1966) commented upon the unusually high levels of slow-wave sleep in physically fit athletes. This corresponds to the observation reported by Griffin and Trinder (1978) that young aerobically fit athletes showed more slow-wave sleep than unfit sedentary controls. Corresponding augmentations of slow-wave sleep, however, could not be observed in slightly older subjects (Trinder, Bruck, Paxton, Montgomery, & Bowling, 1982), while Horne and Porter (1975) observed little or no significant alteration in total ,slow-wave sleep following daytime exercise. Torsvall, Akestedt, and Lindbeck (1984) reported that extreme racing condition caused both delay of the first REM phase and decrease of REM duration. Correspondingly, they found an increase of Stage 2 and a weakly shortened latency of slow-wave sleep. I n addition to differences in slow-wave sleep, aerobically fit individuals have been reported to have longer sleep durations and shorter sleep-onset latencies (Montgomery, Trinder, & Paxton, 1982). However, the difference in slow-wave sleep between fit and unfit subjects might not be due to aerobic fitness but to constitutional factors particularly related to athletic activities and characteristics of athletic individuals. 'Requests for reprints should be addressed to Prof. Dr. Alfred Meier-KoU, Universitat Konstanz, 5stEach 5560, 7750 Konstanz, West German The authors thank Prof. Dr. R. B. Freeman Ydr his comments on the manuscript.
244
J. BUCHEGGER, ETAL.
Further studies of exhaustive exercise and sleep in both male and female subjects have demonstrated significant changes with respect to the quantity and temporal distribution of slow-wave sleep on the night following the exercise (Bunnell, Bevier, & Horvarth, 1983). Other experiments (Horne & Pettitt, 1984) indicated that sleep deprivation had no statistically significant effects on several physiological parameters (heart rate, 0 2 / C 0 2capacity, respiratory quotient) and physical work capacity. The study of Paxton, Trinder, Shapiro, Adam, Oswald, and Graf (1984) assessed the effect of physical fitness and body composition on sleep and the nighttime secretion of hormones (hGH, prolactin, and cortisol). While physical fitness had no significant effect on either sleep or hormone levels, body composition was related to both sleep and growth hormone. Furthermore, anaerobic exercises involving familiar and well-established motor patterns such as sprint running and sprint swimming, weightlifting, and bodybuilding caused no significant alterations in any sleep variables (Trinder, Paxton, Montgomery, & Fraser, 1985). By contrast, influences on sleep-cycle periods have been observed only in aerobic athletes. Shorter sleep-cycle periods were found in athletes with high physical fitness rather than in unfit control subjects. In summary, significantly prolonged slow-wave sleep was exclusively found as a consequence of aerobic exercises, whereas alterations of REM sleep were not observed. In contrast to these findings, Buchegger and Meier-Koll (1988) reported a significant increase in the duration of REM sleep per sleep cycle as well as prolongation of sleep-cycle periods corresponding to an increased index of performance after participating in training units of trampolining. Since trampolining represents an anaerobic form of physical training but requires the acquisition of novel motor patterns, the results were discussed in terms of processes of consolidation linked to neuronal mechanisms involved in REM sleep. I t was hypothesized that skillful trampolining, requiring unfamiliar movements such as translation and rotation of the whole body in the three-dimensional space, affects the architecture of nocturnal sleep, specifically the phases of REM sleep. Further support for the hypothesis of the involvement of sleep in learning and memory comes from the fact that patterns of neuroendocrinal activities are related to specific stages of sleep. As emphasized by Stern and Morgane (1977), the enhancement of protein synthesis during REM sleep might be involved in consolidation of long-term memory.
Subjects
Sixteen subjects (8 men, 8 women), ranging in age from 21 to 24 years, were recruited from undergraduates of the University of Konstanz and were
SPORTS AND NOCTURNAL SLEEP STRUCTURE
245
randomly assigned to an experimental and a control group. Both groups included 4 men and 4 women, respectively. They participated in basic training courses either in trampolining (experimental group) or in dancing and soccer (control group) during 13 consecutive weeks. None of them had earlier experience with acrobatic forms of sport but some of them were experienced in modern rhythmic gymnastics and soccer. AU subjects were of good physical fitness and mental health. They abstained from any medication before and throughout the experiment.
Training Program The training programs consisted of 13 units, one unit per week. During each training unit in trampoliring (every Tuesday from 1 p.m. to 3 p.m.), 8 volunteers (experimental group) had to acquire specific basic movements such as tuck-, piked-, straddle-, and twist-jump as well as several types of somersaults. The performances of trampoliners during each training unit were recorded on videotape. The videorecordings were analyzed with respect to individual performance of each subject by one of the authors (R. F.). For quantification an index of performance was defined as the ratio of successfully performed movements to the number of trials. Individual learning curves were generated by plotting indices of performance against training units. Four subjects of the control group participated in modern rhythmic gymnastic dancing, while the other four were trained in soccer techniques. The training units for controls were held every Thursday (from 1 p.m. to 3 p.m.).
Polygraphic Recordzngs and Sleep-stage Scoring After each of the two weekly training units (about 3 p.m.), a couple of subjects, one from the experimental and one from the control group, came to the laboratory and stayed until the next morning. They engaged in several activities like reading and writing and were allowed to eat and drink ad libitum, but had to abstain from coffee and alcoholic beverages. Also, they abstained from activities which could evoke mental arousal or physical stress. Polygraphic sleep recording started not earlier than 12 a.m. and not later than 12:30 a.m. and lasted till the morning (about 8 a.m.). The polygraphic recordings included two bilateral EEG leadings (C,A,, C,A,) and unipolar leadings of eye movements. Before the beginning of the training program each subject of the experimental group had participated twice in baseline sleep recording, while only one baseline sleep recording was available for each subject of the control group. Since the whole training program of both trampolining and the other anaerobic sports lasted 13 weeks, each subject could participate in sleep recordings four or five times, at least one or two times in baseline and three times during the whole training period. Sleep re-
246
J. BUCHEGGER, ET AL.
cords were blind scored by two of us (J. B. and A. M-K.) according to Rechtschaffen and Kales's instructions (1968). Finally, the duration of REM and nonREM stages observed during the individual recordings as well as their percents of total sleep time were determined.
REM Sleep of Experimental Group (Baselines us Exercise Data) The trampolining program of 13 consecutive weeks was divided into three training blocks, each characterized by the acquisition of special motor skills: Block 1: 1st-5th weekly unit (basic movements) Block 2: 6th-9th weekly unit (somersaults backwards) Block 3: 10th-13th weekly unit (somersaults forwards)
For each training unit individual indices of performance were determined. Subsequently a mean group index of performance was calculated for each week. As expected, the mean group index of performance increased continuously during consecutive training units (Fig. 1). The percents of REM-sleep determined for individual subjects during baseline nights and three training blocks were averaged. These mean values of R E M sleep (percents of total sleep duration) show a significant increase during the training blocks with respect to the baselines (Fig. 1 and Table 1).
Week
FIG.1. REM Sleep (histogram) augmentation (baseline vs exercise data) and increasing index of performance (0)in the experimental group ( n = 8)
SPORTS AND NOCTURNAL SLEEP STRUCTURE TABLE 1 MEAND U R A ~ O N OFS SLEEPSTAGE PER TOTALSLEEP(%) AND STANDARD DEVIATIONS AS A FUNCTION OF BASELINES AND TRAININGBLOCKS ~-
Experimental Group
n
Stage 1
M
SD
Stage 2 M SD
SWS
M
REM
SD
7.8 7.4 25.2 2.2 46.0 8 5.4 9.6 7.0 24.6 45.7 8 7.2 4.0 6.7 4.5 20.2 3.1 44.6 8 5.0 7.8 5.2 23.4 3.0 43.7 8 4.2 4.0 21.8 5.0 1.9 46.7 8 4.1 Note.-For statistical analysis the Wilcoxon matched-pairs test (signed-rank) error robability (p) was estimated for one-sided formulation of the question. ~i~rd!cantly different from both baselines: +p< . D l . $p< ,005. 1.Baseline 2.Baseline Block 1 Block 2 Block 3
SD
M 22.8 22.6 30.2$ 28.7t 27.4
2.9 3.8 4.3 3.5 3.4
was used; the
REM Sleep of Control Group @uselines us Exercise Data) During all training units no significant alteration of REM sleep was detected (Fig. 2). The mean values of REM sleep (percents of total sleep time) determined for the three training blocks were not significantly different from those determined for baseline nights.
Blocks
FIG. 2. Mean durations of REM sleep (%) in the experimental (ns = 8) (No second baseline measure for control group.)
( 0 ) and
control
(m)
groups
NonREM Sleep of Experimental Group (Baseline us Exercise Data) REM-sleep increased significantly during the trarnpolining course with
248
J. BUCHEGGER, ET AL.
respect to baseline condition. Consequently nonREM sleep must decrease considerably. When stages of nonREM sleep were separately tested, no significant differences in baseline and experimental conditions could be detected, neither in Stages 1 and 2 nor in slow-wave sleep (Stages 3 and 4). However, all nonREM stages showed a tendency to decrease slightly from baselines to training units. The augmentation of REM-sleep was obviously not correlated with a decrease of a single one of the nonREM stage, but they all contributed to the considerable decrease of nonREM sleep.
NonREM Sleep of Control Group (Baseline us Exercixe Data) In the control group, no significant differences could be found in sleep Stages 1 and 2 and slow-wave sleep determined for baseline and exercise conditions, respectively. Correspondingly values were fairly constant during the 13-wk. training program (Table 2). TABLE 2 MEANDURATIONS OF SLEEPSTAGE PER TOTALSLEEP(%) AND STANDARD DEVIATIONS AS A FUNCTIONOF BASELINES AND TRAINING BLOCKS Control Group
Stage 1
n
M
SD
Stage 2
SD
M
SWS
M
REM
SD
M
SD
1.Baseline 8 7.9 5.0 52.6 9.2 17.9 5.8 21.6 4.0 8 2. Baseline Block 1 8 7.5 5.7 50.6 6.7 20.8 6.5 21.2 3.2 Block 2 8 7.9 5.6 49.0 8.0 18.9 7.9 24.2 3.6 Block 3 8 8.5 5.0 52.3 8.0 18.3 6.3 20.9 3.7 Note.-For statistical analysis the Wicoxon matched-pairs test (signed-rank) was used; the error probability (P) was estimated for one-sided formulation of the question: nonsignificant differences for ill ;leep stages. TABLE 3
MEANDURATIONS OF REM-SLEEPPER REM/NREM-CYCLE (MIN.) AS A
Experimental vs Control Group
FUNCTION OF BASELINES A N D TRAINING BLOCKS
REM per REWNREM-cycle Exoerimental Control
M
SD
M
SD
1. Baseline 26.7 5.1 25.6 6.8 2. Baseline 26.2 4.1 6.0 8.4 25.9 Block 1 34.9* 5.3 5.5 26.2 Block 2 32.2* 6.4 4.0 26.3 Block 3h 29.5 Note.-For statistical analysis the Wilcoxon matched- airs error orobabilitv (o) was estimated for one-sided formuition
Duration of REMINREM-cycle Experimental Control
M
SD
M
SD
92 92 96 96 91
11 11 10 13 6
102
21
97 94 87
17 14 18
test (signed-rank) was used; the of the question: significantly d ~ f -
REM Sleep per REMINREM Sleep-cycle Period (Both Groups) In the experimental group, REM sleep per REM/NREM sleep-cycle
249
SPORTS AND NOCTURNAL SLEEP STRUCTURE
period was significantly increased during Block 1 and Block 2 of the training program with respect to baseline values. In the control group, no significant change was observed. Also, there were no significant variations in sleep-cycle length of either group (Fig. 3 and Table 3).
Blocks
FIG.3. Alterations of mean duration of total REM sleep (min.) per REMINREM-cycle in experimental ( * ) and control (m) groups (ns = 8) TABLE 4 MEANINDEX
AND
Experimental vs Control Group 1. Baseline 2. Baseline Block 1 Block 2 Block 3
STANDARDDEVIATIONS OF EM-INTENSITY (REMsIMIN.) Rapid Eye Movements Per Minute Exuerimental Control M SD M
SD
0.6 1.1 0.6 0.4 0.7 1.0 0.6 0.7 1.5 0.8 0.5 1.6 0.9 Note.-For statistical analysis the Wilcoxon matched-pairs test (signed-rank) was used; the error probability ( p ) was estimated for one-sided formulation of the question: nonsigdicant differences between baselines and exercise data. 1.4 1.1 1.0 1.2 1.8
REM Intensi~&pid Eye Movements/Min.) of Both Groups In both groups the mean REM intensity (REMs/min.) showed only
250
J. BUCHEGGER, ET AL
slight but no significant alteration from baseline to exercise conditions (Table 4).
DISCUSSION The results support earlier studies of both animals and humans showing changes in sleep stage organization, especially in REM sleep, as a consequence of learning (Grieser, Greenberg, & Harrison, 1972; Zimmerman, Stoyva, & Metcalf, 1971; Horne & McGrath, 1984). There is evidence that sleep might be involved in neuronal and biochemical mechanisms of learning (LeConte, Hennevin, & Bloch, 1974; Lucero, 1970; Fishbein & Gutwein, 1977). More specifically, the acquisition of motor experience seems to be correlated with an increase in REM sleep. Among human studies our observations are most similar to those reported by Paul and Dittrichova (1975) who analyzed sleep patterns following learning in early infancy. Their results have shown an increase in REM sleep during the first sleep-cycle after successful learning in an operant paradigm. There are several studies demonstrating the influence of aerobic exercise on sleep, especially on sleep Stage 2 and slow-wave sleep (Griffin & Trinder, 1978; Buguet, Roussel, Angus, Sabiston, & Radomski, 1980; Trinder, Stevenson, Paxton, & Montgomery, 1982; Bunnell, et al., 1983; Paxton, Trinder, & Montgomery, 1983; Trinder, et al., 1985; Torsvall, et al., 1984). However, sleep studies in aerobically well-trained athletes have not detected an increase in slow-wave sleep (Walker, Floyd, Fein, Cavness, Lualhati, & Feinberg, 1978; Trinder, et al., 1982; Paxton, Montgomery, Trinder, Newman, & Bowling, 1982). With respect to anaerobic exercises such as sprint running, sprint swimming, weightlifting, and bodybuilding, no significant alterations were detected in any sleep variables (Trinder, et al., 1985). In summary, these studies showed that sleep Stage 2 and slow-wave sleep were significantly prolonged exclusively as a consequence of aerobic exercises, while significant alteration of REM sleep was not observed. These findings correspond to our own observations that REM sleep of our control group participating in anaerobic exercises was not prolonged significantly. Correspondingly, sleep Stage 1, Stage 2, and slow-wave sleep showed no considerable variations. In our previous study (Buchegger & Meier-Koll, 1988), we reported significant increase in the REM sleep per sleep-cycle and prolonged periods of REMINREM cycle as a consequence of trampolining exercises. While the increase in REM sleep per sleep-cycle found in the present study is similar to our previous findings (Fig. 3), the statistical analysis of sleep-cycle periods did not replicate a prolongation of sleep-cycle period (Table 3). From this point of view, the results of the previous and present study are contradictory, since no significant alterations in the sleep-cycle period of the experimental and control groups could be observed.
SPORTS AND NOCTURNAL SLEEP STRUCTURE
251
In summary, the results of the present study provide evidence that organization of sleep stages can be changed by processes in motor learning or by effects related to the acquisition of unfamiliar motor patterns. These processes might influence an REM-sleep oscillator system (McCarley & Massaquoi, 1986) by decreasing the trigger level of REM phase, whereas both REM intensity and period of the REM cycle remain unchanged. REFERENCES BAEKELAND, F., & LASKY,R. (1966) Exercise and sleep patterns in college athletes. Perceptual and Motor Skills, 23, 1203-1207. BUCHEGGER, J., & MEEK-KOLL,A. (1988) Motor learning and ultradian sleep cycle: an encephalographic scudy of trampoliners. Perceptual and Motor Skills, 67, 635-1545, BUNNELL, D. E., BEVIER,W., & HORVARTH, S. M. (1983) Effects of exhaustive exercise on sleep of men and women. Psychobiology, 20, 50-58. BUGLET, A,, ROUSSEL,B., ANGUS,R., SABISTON, B., & RADOMSKI, M. (1980) Human sleep and adrenal individual reactions to exercise. Electroencephalography and Clinical Neuropsychology, 49, 515-523. FISHBEIN,W., & GUTWEIN,B. M. (1977) Paradoxical sleep and memory storage processes. Behavioral Biology, 19, 425-464. GRIESER,C., GREENBERG, R., & HARRISON,R. H . (1972) The adaptive function of sleep: the differential effects of sleep and dreaming on recall. Journal of Abnormal Psychology, 80, 280-286. GRIFFIN,S. J., & TRINDER, J. (1978) Physical fitness, exercise and human sleep. Psychobiology, 15, 447-450. HORNE,J. A., & MCGRATH,M. (1984) The consolidation hypothesis for REM sleep function: stress and other c o n ~ u n d i n gfacton-a review. Biological Psychology, 19, 165-184. HORNE, J. A,, & P E ~ I T T ,A. N. (1984) Sleep deprivation and the physiological response to exercise under steady-state conditions in untrained subjects. Sleep, 7, 168-179. HORNE,J. A,, & PORTER,J. M. (1975) Exercise and human sleep. Nature, 256, 573-575. LECONTE,l?, HENNEVIN, E., & BLOCH,V. (1974) Duration of paradoxical slee necessary for the acquisition of conditioned avoidance in the rat. Physiology a n 8 Behavior, 13, 675-681. LUCERO,M. A. (1970) Lengthening of REM sleep duration consecutive to learning in the rat. Brain Research, 20, 319-322. MCCARLEY, R. W., & MASSAQUOI, S. G . (1986) A limit cycle mathematical model of the REM sleep oscillator system. American Journal of Physiology, 251, R 1011-R 1029. I., TRINDER,J., & PAXTON,S. J. (1982) Energy expenditure and total sleep MONTGOMERY, time: the effect of physical exercise. Sleep, 5, 159-168. PAUL,K., & D I ~ I C H O VJ.A(1975) , Sleep patterns following learning in infants. In P. Levin & W. D. Koella (Eds.), Sleep 1974. Second European Congress of Sleep Research, Basel: Karger. Pp. 338-390. PAXTON,S., MONTGOMERY, I., T ~ D E RJ.,, NEWMAN. J.: & BOWLING, A. (1982) Slee after exercise of variable intensity in fit and unflt subjects. Aus&alian Journal of ~ s ~ c ~ o l o g ~ , 34, 289-296. PAXTON,S. ,TRINDER,J., & MONTGOMERY, I. (1983) Does aerobic fitness affect sleep? Psyc obrology, 20, 320-324. PAXTON,S. J., TRINDER,, SHAPIRO,C. M., ADAM,K., OSWALD, I., & GRLF, K. J. (1984) Effect of physical kness and body composition on sleep and sleep-relaad hormone concentrations. Sleep, 7, 339-346. RECHTSCI-IAFFEN, A., & KALES, A. (1968) A manual of standardized terminology, techniques and scoring system for sleep stoges of human subjects. (Public Health Service) Washington, DC: US Government Printing Office. P. J. (1977) Sleep and memory: effects of growth hormone on STERN,W. C., & MORGANE, sleep, brain neurochemistry and behavior. In R. R. Dmcker-Colin & J. L. McGaugh (Eds.), Neurobiology of sleep and memory. New York: Academic Press. Pp. 373-400.
d.,
252
J. BUCHEGGER, E T A L .
TORSVALL, L., AKERSTEDT, T.,& LINDBECK, G . (1984) Effects on sleep stages and E E G power density of different degrees of exercise in fit subjects. Electroencephalography and Clinical Neuropbysiology, 57, 347-353. TRINDER, J., BRUCK,D., PAXTON,S., MONTGOMERY, I., & BOWLING, A. (1982) Physical fitness, exercise, age and human sleep. Australian Journal of Psychology, 32, 131-138. TRMDER,J., PAXTON,S. J., MONTGOMERY, I., & FRASER,G . (1985) Endurance as opposed to power training: their effect on sleep. Psychobiology, 22, 668-673. TRINDER, J., STEVENSON, J., PAXTON,S. J., & MONTGOMERY, I. (1982) Physical fitness, exercise and REM sleep cycle length. Psychobiology, 19, 89-93. C., L u w n , R., & FEINBERG, I. (1978) WALKER,J. M., FLOYD,T. C., FEIN, G., CAVNESS, Effects of exercise on sleep. Journal of Applied Psychology, 44, 945-951. ZIMMERMAN, J., STOWA,J., & METCALF,0. (1971) Distorted visual feedback and augmented REM sleep. Psychophysiology, 7, 298.
Accepted July 2, 1991.
O Perceptual and Motor
Skills 1991
D O E S TRAMPOLINING A N D ANAEROBIC PHYSICAL FITNESS AFFECT SLEEP? ' JOACHIM BUCHEGGER, REINER FRITSCH, ALFRED MEIER-KOLL AND HARTMUT RIEHLE2 Universitat Konstanz Summary.-The structure of nocturnal sleep of 16 volunteers, participating in the anaerobic sports of trampolining, dancing, and soccer, was monitored by means of polygraphic recordings. Since trampolining requires the acquisition of unfamiliar patterns of motor coordination, it can be considered as a special form of motor learning, whereas the acquisition of motor skills specific for dancing and soccer can be linked with motor patterns of normal biped locomotion. According to this view, an experimental group of 8 volunteers was formed; they participated in a training course of trampolining. In addition, a control group of 8 subjects was recruited, who engaged in one of the other two anaerobic sports. Subjects who had acquired new motor skills during a 13-wk. program in trampolining showed a statistically significant increase in REM-sleep. By contrast, the 8 subjects of the control group showed no considerable changes in REM-sleep. This suggests that efforts in acquiring new and complex motor patterns activate processes specifically involved in the generation of REM stage during nocturnal sleep.
There is a considerable body of evidence that sleep can be affected by physical fitness and exercise. I n their study of the effects of physical exercise on human sleep Baekeland and Lasky (1966) commented upon the unusually high levels of slow-wave sleep in physically fit athletes. This corresponds to the observation reported by Griffin and Trinder (1978) that young aerobically fit athletes showed more slow-wave sleep than unfit sedentary controls. Corresponding augmentations of slow-wave sleep, however, could not be observed in slightly older subjects (Trinder, Bruck, Paxton, Montgomery, & Bowling, 1982), while Horne and Porter (1975) observed little or no significant alteration in total ,slow-wave sleep following daytime exercise. Torsvall, Akestedt, and Lindbeck (1984) reported that extreme racing condition caused both delay of the first REM phase and decrease of REM duration. Correspondingly, they found an increase of Stage 2 and a weakly shortened latency of slow-wave sleep. I n addition to differences in slow-wave sleep, aerobically fit individuals have been reported to have longer sleep durations and shorter sleep-onset latencies (Montgomery, Trinder, & Paxton, 1982). However, the difference in slow-wave sleep between fit and unfit subjects might not be due to aerobic fitness but to constitutional factors particularly related to athletic activities and characteristics of athletic individuals. 'Requests for reprints should be addressed to Prof. Dr. Alfred Meier-KoU, Universitat Konstanz, 5stEach 5560, 7750 Konstanz, West German The authors thank Prof. Dr. R. B. Freeman Ydr his comments on the manuscript.
244
J. BUCHEGGER, ETAL.
Further studies of exhaustive exercise and sleep in both male and female subjects have demonstrated significant changes with respect to the quantity and temporal distribution of slow-wave sleep on the night following the exercise (Bunnell, Bevier, & Horvarth, 1983). Other experiments (Horne & Pettitt, 1984) indicated that sleep deprivation had no statistically significant effects on several physiological parameters (heart rate, 0 2 / C 0 2capacity, respiratory quotient) and physical work capacity. The study of Paxton, Trinder, Shapiro, Adam, Oswald, and Graf (1984) assessed the effect of physical fitness and body composition on sleep and the nighttime secretion of hormones (hGH, prolactin, and cortisol). While physical fitness had no significant effect on either sleep or hormone levels, body composition was related to both sleep and growth hormone. Furthermore, anaerobic exercises involving familiar and well-established motor patterns such as sprint running and sprint swimming, weightlifting, and bodybuilding caused no significant alterations in any sleep variables (Trinder, Paxton, Montgomery, & Fraser, 1985). By contrast, influences on sleep-cycle periods have been observed only in aerobic athletes. Shorter sleep-cycle periods were found in athletes with high physical fitness rather than in unfit control subjects. In summary, significantly prolonged slow-wave sleep was exclusively found as a consequence of aerobic exercises, whereas alterations of REM sleep were not observed. In contrast to these findings, Buchegger and Meier-Koll (1988) reported a significant increase in the duration of REM sleep per sleep cycle as well as prolongation of sleep-cycle periods corresponding to an increased index of performance after participating in training units of trampolining. Since trampolining represents an anaerobic form of physical training but requires the acquisition of novel motor patterns, the results were discussed in terms of processes of consolidation linked to neuronal mechanisms involved in REM sleep. I t was hypothesized that skillful trampolining, requiring unfamiliar movements such as translation and rotation of the whole body in the three-dimensional space, affects the architecture of nocturnal sleep, specifically the phases of REM sleep. Further support for the hypothesis of the involvement of sleep in learning and memory comes from the fact that patterns of neuroendocrinal activities are related to specific stages of sleep. As emphasized by Stern and Morgane (1977), the enhancement of protein synthesis during REM sleep might be involved in consolidation of long-term memory.
Subjects
Sixteen subjects (8 men, 8 women), ranging in age from 21 to 24 years, were recruited from undergraduates of the University of Konstanz and were
SPORTS AND NOCTURNAL SLEEP STRUCTURE
245
randomly assigned to an experimental and a control group. Both groups included 4 men and 4 women, respectively. They participated in basic training courses either in trampolining (experimental group) or in dancing and soccer (control group) during 13 consecutive weeks. None of them had earlier experience with acrobatic forms of sport but some of them were experienced in modern rhythmic gymnastics and soccer. AU subjects were of good physical fitness and mental health. They abstained from any medication before and throughout the experiment.
Training Program The training programs consisted of 13 units, one unit per week. During each training unit in trampoliring (every Tuesday from 1 p.m. to 3 p.m.), 8 volunteers (experimental group) had to acquire specific basic movements such as tuck-, piked-, straddle-, and twist-jump as well as several types of somersaults. The performances of trampoliners during each training unit were recorded on videotape. The videorecordings were analyzed with respect to individual performance of each subject by one of the authors (R. F.). For quantification an index of performance was defined as the ratio of successfully performed movements to the number of trials. Individual learning curves were generated by plotting indices of performance against training units. Four subjects of the control group participated in modern rhythmic gymnastic dancing, while the other four were trained in soccer techniques. The training units for controls were held every Thursday (from 1 p.m. to 3 p.m.).
Polygraphic Recordzngs and Sleep-stage Scoring After each of the two weekly training units (about 3 p.m.), a couple of subjects, one from the experimental and one from the control group, came to the laboratory and stayed until the next morning. They engaged in several activities like reading and writing and were allowed to eat and drink ad libitum, but had to abstain from coffee and alcoholic beverages. Also, they abstained from activities which could evoke mental arousal or physical stress. Polygraphic sleep recording started not earlier than 12 a.m. and not later than 12:30 a.m. and lasted till the morning (about 8 a.m.). The polygraphic recordings included two bilateral EEG leadings (C,A,, C,A,) and unipolar leadings of eye movements. Before the beginning of the training program each subject of the experimental group had participated twice in baseline sleep recording, while only one baseline sleep recording was available for each subject of the control group. Since the whole training program of both trampolining and the other anaerobic sports lasted 13 weeks, each subject could participate in sleep recordings four or five times, at least one or two times in baseline and three times during the whole training period. Sleep re-
246
J. BUCHEGGER, ET AL.
cords were blind scored by two of us (J. B. and A. M-K.) according to Rechtschaffen and Kales's instructions (1968). Finally, the duration of REM and nonREM stages observed during the individual recordings as well as their percents of total sleep time were determined.
REM Sleep of Experimental Group (Baselines us Exercise Data) The trampolining program of 13 consecutive weeks was divided into three training blocks, each characterized by the acquisition of special motor skills: Block 1: 1st-5th weekly unit (basic movements) Block 2: 6th-9th weekly unit (somersaults backwards) Block 3: 10th-13th weekly unit (somersaults forwards)
For each training unit individual indices of performance were determined. Subsequently a mean group index of performance was calculated for each week. As expected, the mean group index of performance increased continuously during consecutive training units (Fig. 1). The percents of REM-sleep determined for individual subjects during baseline nights and three training blocks were averaged. These mean values of R E M sleep (percents of total sleep duration) show a significant increase during the training blocks with respect to the baselines (Fig. 1 and Table 1).
Week
FIG.1. REM Sleep (histogram) augmentation (baseline vs exercise data) and increasing index of performance (0)in the experimental group ( n = 8)
SPORTS AND NOCTURNAL SLEEP STRUCTURE TABLE 1 MEAND U R A ~ O N OFS SLEEPSTAGE PER TOTALSLEEP(%) AND STANDARD DEVIATIONS AS A FUNCTION OF BASELINES AND TRAININGBLOCKS ~-
Experimental Group
n
Stage 1
M
SD
Stage 2 M SD
SWS
M
REM
SD
7.8 7.4 25.2 2.2 46.0 8 5.4 9.6 7.0 24.6 45.7 8 7.2 4.0 6.7 4.5 20.2 3.1 44.6 8 5.0 7.8 5.2 23.4 3.0 43.7 8 4.2 4.0 21.8 5.0 1.9 46.7 8 4.1 Note.-For statistical analysis the Wilcoxon matched-pairs test (signed-rank) error robability (p) was estimated for one-sided formulation of the question. ~i~rd!cantly different from both baselines: +p< . D l . $p< ,005. 1.Baseline 2.Baseline Block 1 Block 2 Block 3
SD
M 22.8 22.6 30.2$ 28.7t 27.4
2.9 3.8 4.3 3.5 3.4
was used; the
REM Sleep of Control Group @uselines us Exercise Data) During all training units no significant alteration of REM sleep was detected (Fig. 2). The mean values of REM sleep (percents of total sleep time) determined for the three training blocks were not significantly different from those determined for baseline nights.
Blocks
FIG. 2. Mean durations of REM sleep (%) in the experimental (ns = 8) (No second baseline measure for control group.)
( 0 ) and
control
(m)
groups
NonREM Sleep of Experimental Group (Baseline us Exercise Data) REM-sleep increased significantly during the trarnpolining course with
248
J. BUCHEGGER, ET AL.
respect to baseline condition. Consequently nonREM sleep must decrease considerably. When stages of nonREM sleep were separately tested, no significant differences in baseline and experimental conditions could be detected, neither in Stages 1 and 2 nor in slow-wave sleep (Stages 3 and 4). However, all nonREM stages showed a tendency to decrease slightly from baselines to training units. The augmentation of REM-sleep was obviously not correlated with a decrease of a single one of the nonREM stage, but they all contributed to the considerable decrease of nonREM sleep.
NonREM Sleep of Control Group (Baseline us Exercixe Data) In the control group, no significant differences could be found in sleep Stages 1 and 2 and slow-wave sleep determined for baseline and exercise conditions, respectively. Correspondingly values were fairly constant during the 13-wk. training program (Table 2). TABLE 2 MEANDURATIONS OF SLEEPSTAGE PER TOTALSLEEP(%) AND STANDARD DEVIATIONS AS A FUNCTIONOF BASELINES AND TRAINING BLOCKS Control Group
Stage 1
n
M
SD
Stage 2
SD
M
SWS
M
REM
SD
M
SD
1.Baseline 8 7.9 5.0 52.6 9.2 17.9 5.8 21.6 4.0 8 2. Baseline Block 1 8 7.5 5.7 50.6 6.7 20.8 6.5 21.2 3.2 Block 2 8 7.9 5.6 49.0 8.0 18.9 7.9 24.2 3.6 Block 3 8 8.5 5.0 52.3 8.0 18.3 6.3 20.9 3.7 Note.-For statistical analysis the Wicoxon matched-pairs test (signed-rank) was used; the error probability (P) was estimated for one-sided formulation of the question: nonsignificant differences for ill ;leep stages. TABLE 3
MEANDURATIONS OF REM-SLEEPPER REM/NREM-CYCLE (MIN.) AS A
Experimental vs Control Group
FUNCTION OF BASELINES A N D TRAINING BLOCKS
REM per REWNREM-cycle Exoerimental Control
M
SD
M
SD
1. Baseline 26.7 5.1 25.6 6.8 2. Baseline 26.2 4.1 6.0 8.4 25.9 Block 1 34.9* 5.3 5.5 26.2 Block 2 32.2* 6.4 4.0 26.3 Block 3h 29.5 Note.-For statistical analysis the Wilcoxon matched- airs error orobabilitv (o) was estimated for one-sided formuition
Duration of REMINREM-cycle Experimental Control
M
SD
M
SD
92 92 96 96 91
11 11 10 13 6
102
21
97 94 87
17 14 18
test (signed-rank) was used; the of the question: significantly d ~ f -
REM Sleep per REMINREM Sleep-cycle Period (Both Groups) In the experimental group, REM sleep per REM/NREM sleep-cycle
249
SPORTS AND NOCTURNAL SLEEP STRUCTURE
period was significantly increased during Block 1 and Block 2 of the training program with respect to baseline values. In the control group, no significant change was observed. Also, there were no significant variations in sleep-cycle length of either group (Fig. 3 and Table 3).
Blocks
FIG.3. Alterations of mean duration of total REM sleep (min.) per REMINREM-cycle in experimental ( * ) and control (m) groups (ns = 8) TABLE 4 MEANINDEX
AND
Experimental vs Control Group 1. Baseline 2. Baseline Block 1 Block 2 Block 3
STANDARDDEVIATIONS OF EM-INTENSITY (REMsIMIN.) Rapid Eye Movements Per Minute Exuerimental Control M SD M
SD
0.6 1.1 0.6 0.4 0.7 1.0 0.6 0.7 1.5 0.8 0.5 1.6 0.9 Note.-For statistical analysis the Wilcoxon matched-pairs test (signed-rank) was used; the error probability ( p ) was estimated for one-sided formulation of the question: nonsigdicant differences between baselines and exercise data. 1.4 1.1 1.0 1.2 1.8
REM Intensi~&pid Eye Movements/Min.) of Both Groups In both groups the mean REM intensity (REMs/min.) showed only
250
J. BUCHEGGER, ET AL
slight but no significant alteration from baseline to exercise conditions (Table 4).
DISCUSSION The results support earlier studies of both animals and humans showing changes in sleep stage organization, especially in REM sleep, as a consequence of learning (Grieser, Greenberg, & Harrison, 1972; Zimmerman, Stoyva, & Metcalf, 1971; Horne & McGrath, 1984). There is evidence that sleep might be involved in neuronal and biochemical mechanisms of learning (LeConte, Hennevin, & Bloch, 1974; Lucero, 1970; Fishbein & Gutwein, 1977). More specifically, the acquisition of motor experience seems to be correlated with an increase in REM sleep. Among human studies our observations are most similar to those reported by Paul and Dittrichova (1975) who analyzed sleep patterns following learning in early infancy. Their results have shown an increase in REM sleep during the first sleep-cycle after successful learning in an operant paradigm. There are several studies demonstrating the influence of aerobic exercise on sleep, especially on sleep Stage 2 and slow-wave sleep (Griffin & Trinder, 1978; Buguet, Roussel, Angus, Sabiston, & Radomski, 1980; Trinder, Stevenson, Paxton, & Montgomery, 1982; Bunnell, et al., 1983; Paxton, Trinder, & Montgomery, 1983; Trinder, et al., 1985; Torsvall, et al., 1984). However, sleep studies in aerobically well-trained athletes have not detected an increase in slow-wave sleep (Walker, Floyd, Fein, Cavness, Lualhati, & Feinberg, 1978; Trinder, et al., 1982; Paxton, Montgomery, Trinder, Newman, & Bowling, 1982). With respect to anaerobic exercises such as sprint running, sprint swimming, weightlifting, and bodybuilding, no significant alterations were detected in any sleep variables (Trinder, et al., 1985). In summary, these studies showed that sleep Stage 2 and slow-wave sleep were significantly prolonged exclusively as a consequence of aerobic exercises, while significant alteration of REM sleep was not observed. These findings correspond to our own observations that REM sleep of our control group participating in anaerobic exercises was not prolonged significantly. Correspondingly, sleep Stage 1, Stage 2, and slow-wave sleep showed no considerable variations. In our previous study (Buchegger & Meier-Koll, 1988), we reported significant increase in the REM sleep per sleep-cycle and prolonged periods of REMINREM cycle as a consequence of trampolining exercises. While the increase in REM sleep per sleep-cycle found in the present study is similar to our previous findings (Fig. 3), the statistical analysis of sleep-cycle periods did not replicate a prolongation of sleep-cycle period (Table 3). From this point of view, the results of the previous and present study are contradictory, since no significant alterations in the sleep-cycle period of the experimental and control groups could be observed.
SPORTS AND NOCTURNAL SLEEP STRUCTURE
251
In summary, the results of the present study provide evidence that organization of sleep stages can be changed by processes in motor learning or by effects related to the acquisition of unfamiliar motor patterns. These processes might influence an REM-sleep oscillator system (McCarley & Massaquoi, 1986) by decreasing the trigger level of REM phase, whereas both REM intensity and period of the REM cycle remain unchanged. REFERENCES BAEKELAND, F., & LASKY,R. (1966) Exercise and sleep patterns in college athletes. Perceptual and Motor Skills, 23, 1203-1207. BUCHEGGER, J., & MEEK-KOLL,A. (1988) Motor learning and ultradian sleep cycle: an encephalographic scudy of trampoliners. Perceptual and Motor Skills, 67, 635-1545, BUNNELL, D. E., BEVIER,W., & HORVARTH, S. M. (1983) Effects of exhaustive exercise on sleep of men and women. Psychobiology, 20, 50-58. BUGLET, A,, ROUSSEL,B., ANGUS,R., SABISTON, B., & RADOMSKI, M. (1980) Human sleep and adrenal individual reactions to exercise. Electroencephalography and Clinical Neuropsychology, 49, 515-523. FISHBEIN,W., & GUTWEIN,B. M. (1977) Paradoxical sleep and memory storage processes. Behavioral Biology, 19, 425-464. GRIESER,C., GREENBERG, R., & HARRISON,R. H . (1972) The adaptive function of sleep: the differential effects of sleep and dreaming on recall. Journal of Abnormal Psychology, 80, 280-286. GRIFFIN,S. J., & TRINDER, J. (1978) Physical fitness, exercise and human sleep. Psychobiology, 15, 447-450. HORNE,J. A., & MCGRATH,M. (1984) The consolidation hypothesis for REM sleep function: stress and other c o n ~ u n d i n gfacton-a review. Biological Psychology, 19, 165-184. HORNE, J. A,, & P E ~ I T T ,A. N. (1984) Sleep deprivation and the physiological response to exercise under steady-state conditions in untrained subjects. Sleep, 7, 168-179. HORNE,J. A,, & PORTER,J. M. (1975) Exercise and human sleep. Nature, 256, 573-575. LECONTE,l?, HENNEVIN, E., & BLOCH,V. (1974) Duration of paradoxical slee necessary for the acquisition of conditioned avoidance in the rat. Physiology a n 8 Behavior, 13, 675-681. LUCERO,M. A. (1970) Lengthening of REM sleep duration consecutive to learning in the rat. Brain Research, 20, 319-322. MCCARLEY, R. W., & MASSAQUOI, S. G . (1986) A limit cycle mathematical model of the REM sleep oscillator system. American Journal of Physiology, 251, R 1011-R 1029. I., TRINDER,J., & PAXTON,S. J. (1982) Energy expenditure and total sleep MONTGOMERY, time: the effect of physical exercise. Sleep, 5, 159-168. PAUL,K., & D I ~ I C H O VJ.A(1975) , Sleep patterns following learning in infants. In P. Levin & W. D. Koella (Eds.), Sleep 1974. Second European Congress of Sleep Research, Basel: Karger. Pp. 338-390. PAXTON,S., MONTGOMERY, I., T ~ D E RJ.,, NEWMAN. J.: & BOWLING, A. (1982) Slee after exercise of variable intensity in fit and unflt subjects. Aus&alian Journal of ~ s ~ c ~ o l o g ~ , 34, 289-296. PAXTON,S. ,TRINDER,J., & MONTGOMERY, I. (1983) Does aerobic fitness affect sleep? Psyc obrology, 20, 320-324. PAXTON,S. J., TRINDER,, SHAPIRO,C. M., ADAM,K., OSWALD, I., & GRLF, K. J. (1984) Effect of physical kness and body composition on sleep and sleep-relaad hormone concentrations. Sleep, 7, 339-346. RECHTSCI-IAFFEN, A., & KALES, A. (1968) A manual of standardized terminology, techniques and scoring system for sleep stoges of human subjects. (Public Health Service) Washington, DC: US Government Printing Office. P. J. (1977) Sleep and memory: effects of growth hormone on STERN,W. C., & MORGANE, sleep, brain neurochemistry and behavior. In R. R. Dmcker-Colin & J. L. McGaugh (Eds.), Neurobiology of sleep and memory. New York: Academic Press. Pp. 373-400.
d.,
252
J. BUCHEGGER, E T A L .
TORSVALL, L., AKERSTEDT, T.,& LINDBECK, G . (1984) Effects on sleep stages and E E G power density of different degrees of exercise in fit subjects. Electroencephalography and Clinical Neuropbysiology, 57, 347-353. TRINDER, J., BRUCK,D., PAXTON,S., MONTGOMERY, I., & BOWLING, A. (1982) Physical fitness, exercise, age and human sleep. Australian Journal of Psychology, 32, 131-138. TRMDER,J., PAXTON,S. J., MONTGOMERY, I., & FRASER,G . (1985) Endurance as opposed to power training: their effect on sleep. Psychobiology, 22, 668-673. TRINDER, J., STEVENSON, J., PAXTON,S. J., & MONTGOMERY, I. (1982) Physical fitness, exercise and REM sleep cycle length. Psychobiology, 19, 89-93. C., L u w n , R., & FEINBERG, I. (1978) WALKER,J. M., FLOYD,T. C., FEIN, G., CAVNESS, Effects of exercise on sleep. Journal of Applied Psychology, 44, 945-951. ZIMMERMAN, J., STOWA,J., & METCALF,0. (1971) Distorted visual feedback and augmented REM sleep. Psychophysiology, 7, 298.
Accepted July 2, 1991.
