Sleep is classically divided into two main categories: rapid eye movement or REM sleep, and non-REM sleep, which consists of progressively deeper stages of 1, 2, 3, and 4. Stages 3 and 4 are collectively referred to as slow wave sleep (SWS). Since 1970, it has been thought that the slow wave component of sleep may be the most important component of the proposed restorative function of sleep (35). Both the restorative(1,43) and energy conservation(6) theories of sleep imply that sleep is a compensatory mechanism following catabolic processes of daytime activity, such that increases in training volume would result in increases in SWS. Evidence supporting this theory, however, is confused by a number of inter-study discrepancies, and studies examining this premise are equivocal. Some research supports this SWS: exercise relationship (5,46) while others negate it (37,48).
It has been suggested that habitual exercise in physically fit individuals or athletes may be responsible for changes in SWS following exercise(10,28,46). Although improved physical fitness increased SWS and sleep efficiency in male army recruits(46), it had no effect on the sleep of young women(27). However, athletes tend to have higher proportions of SWS than sedentary controls (5,14,38). Thus, evidence supporting a SWS:fitness relationship is also equivocal, and changes in sleep patterns in athletes may be caused by something other than habitual exercise (39).
In the absence, then, of a SWS:fitness relationship, there is the possibility of a SWS:temperature relationship. Although it has been suggested that the specific increase in brain-as opposed to body-temperature is responsible for an increase in SWS (16), no SWS:temperature relationship has been found with either cooling of the head during exercise (17) or with body cooling as examined after 36 h sleep deprivation (a 24-h routine) (3). Swimming training is associated with both head and body cooling, and while body heating tends to be lower in swimming training than in running(47), swimming remains a sport that has been excluded from the SWS:temperature debate.
Furthermore, when studies are done on the physiological and psychological effects of increased training load, swimmers are frequently used as the subjects (15,22,30,33). This is because of the nature of their training, which emphasises both high frequency and long duration, resulting in a high total volume or training“mileage.” Physiological changes that occur with excessive training load include a compromised immune system and muscular and endocrine changes (12,23). Sleep disturbance, defined as“an imbalance between training and recovery, exercise and exercise capacity, stress and stress tolerance” (23), is one of the proposed physiological effects of overtraining, and although it is sometimes used as a criterion for identifying overtraining syndrome(23), evidence for such disturbances remains anecdotal.
Apart from the physiological changes associated with excessive training, psychological changes appear to occur in a dose-dependent manner. More specifically, as training volume increases, so mood deteriorates, with mood only improving once training load is reduced(29,30,34).
Since swimmers are frequently used as subjects in studies relating to excessive training, we used this group of athletes to evaluate sleep patterns, in particular SWS proportions and sleep disturbance, across a normal swimming season when training volume varied according to competitive requirements rather than short term manipulation. Since sleep disturbance may reduce sleep quality and thereby affect mood detrimentally, psychological changes in association with sleep patterns were also studied. We tested the hypothesis that training volume directly influences the amount of SWS and the level of psychological stress.
SUBJECTS AND METHODS
Seven female swimmers, competitive at a national level and with a mean(± SD) age of 19 ± 2 yr and mean height of 172.3 ± 6 cm volunteered to participate in this study. The study was approved by the Committee for Research on Human Subjects at the University of the Witwatersrand. Informed, written consent was obtained prior to commencement of the study from each individual, or from a parent in the case of subjects under the age of 18 years. In light of the remote possibility of rectal infection and/or diarrhea, permission was not granted for the measurement of rectal temperatures.
Subjects served as their own controls since, practically-speaking, it was impossible to obtain competitive athletes who were prepared to maintain peak training volumes (even if they did not suffer overuse injuries) and forfeit precompetition taper. By using the same research group across the study period, we negated such possible confounding factors as age, gender, and time of day used for training.
The study was conducted over a 6-month spring-summer period, consequently minimizing possible seasonal effects on sleep hygiene and habits. Two of the seven women were taking oral contraceptives, but no attempt was made to control for menstrual cycle phase.
Swimmers were examined at 3-month intervals during their competitive training season, namely, at the start of the season (onset) when training was predominantly medium-intensity aerobic exercise, at peak training (peak) when training consisted of high intensity aerobic exercise plus maximal intensity anaerobic exercise, and after precompetition reduction in training (taper) when the low-volume training was predominantly high intensity, anaerobic exercise.
For each phase, the first night in the laboratory facilitated adaptation to the research conditions, with electrodes being applied but no recordings made(2). The second night was spent at home, and the women returned to the sleep laboratory on the third night when recordings were made. Standard polygraphic sleep recordings were obtained(36), including an electro-oculogram (EOG), a chin electromyogram (EMG), and electroencephalogram (EEG) recordings on a Beckman Montage-20 Electroencephalogram (Model IL60/76, Beckman Instruments Inc, Schiller Park, IL) at a paper speed of 15 mm·s-1. Subjects adhered to their customary bed times and awakening times. Normal daily routines were followed, with evening training being completed before 1800 h on each study night.
All sleep records were coded and scored blind in 20-s epochs by the same experienced scorer according to standard criteria (41). Random records obtained by a second experienced scorer were compared with the results to ensure reliability and provide quality control. Agreement for scoring was high (greater than 82%); furthermore, a second random selection of records was scored twice by the same scorer, again with good reliability between the initial and second scoring. The following sleep parameters were examined: (a) total recording time (TRT), the time from lights out to the time of lights on; (b) sleep onset latency (SOL), the time from lights out to the first appearance of at least 3 consecutive epochs of stage 2 sleep; (c) total sleep time (TST), calculated from TRT minus SOL and including any wakefulness intervening after sleep onset; (d) time spent in SWS; (e) time spent in REM sleep; (f) REM onset latency (ROL), the time from sleep onset to the first epoch of REM sleep; (g) shifts from any sleep stage to awake; and (h) total number of movements. Since we were particularly interested in sleep disturbances, however mild, a movement was defined as occupying 20% of any single epoch instead of the customary 50%. Since subjects were allowed to sleep according to individual need, total recording time between subjects was not uniform; however, TST was the same for each individual over the three phases. Therefore, to facilitate inter-individual comparisons sleep parameters are expressed as a percentage of TST where appropriate. On mornings following the recording nights, the women self rated their sleep quality on a 100-mm visual analogue scale.
At each visit, body composition was assessed using four skinfold sites(Harpenden calipers, British Indicators, Ltd, St. Albans, UK)(11). Psychological assessments were made using the POMS Questionnaire (26). In addition, diet was recorded for 7 d at each training phase and energy intake was assessed using theFoodfinder computer package (Medical Technologies, Tygerberg, South Africa) to exclude possible effects of dietary variations on sleep patterns.
Changes in individual performance were assessed at each training phase using official competition times for each swimmer's specialty event. A daily training log indicating waking heart rate, training volume, and sleep habits was maintained for the entire 6-month study period.
The nonparametric equivalent of repeated measures ANOVA, the Friedman test, was used to assess significance, with Dunn's statistical test being used forpost-hoc analysis and identification of the origin of significance(Instat, Instant Biostatistics Software, Graphpad Software Inc., San Diego, CA). Data are expressed in the form of the mean ± SD.
Training volumes at onset and peak training were not significantly different, but daily training distances (Table 1) were significantly higher at onset and peak than at taper, and performance times improved significantly across the 6-month period by 2.7% (range 1.2-4.9%). With the different training volumes, there was no difference in the swimmers' heart rates on awakening. Although the total energy intake at peak training(15172 ± 2437 kJ·d-1) was significantly higher than at taper (10821 ± 1540 kJ·d-1), the composition of the women's macronutrients remained similar at each study phase(Table 1). Body mass and body composition, expressed as percent body fat, decreased significantly in the first 3 months of the study from onset to peak training (Table 1).
Subjective ratings of sleep quality were consistent and similar for all three study phases. Table 2 provides data for all the sleep parameters measured. The TRT, TST, and time awake after sleep onset were similar for all three test nights, as were SOL, ROL, and total REM sleep.
The number of awakenings was not different over the training period; however, the total number of movements was significantly higher in the onset and peak training period by 10.3% and 11%, respectively, compared with taper. Furthermore, SWS was significantly higher at onset and peak training than at taper. The reduced SWS at taper was associated with a significant increase in stages 1 and 2 of non-REM sleep. Although no attempt was made to control for menstrual cycle phase in this study, a recent, carefully controlled study(9) found that menstrual cycle phase had no effect on SWS. The significance of our SWS results at taper compared with onset and peak are therefore unlikely to be attributable to differences in menstrual cycle.
Data for selected psychological parameters are shown inTable 3. Anger and fatigue levels remained similar across the training season. The decrease in training volume associated with precompetition taper resulted in significantly lower vigor scores than those obtained at either peak training or training onset and was also associated with significantly higher tension and depression scores compared with peak training. Confusion at peak training volume was significantly lower than scores measured at either onset of training or taper.
In this group of competitive athletes, time spent in SWS at the higher training volumes of onset and peak was significantly more than at taper, supporting the restorative theory of sleep(1,35,43,46) that proposes that it is the SWS component that is physiologically restorative. Consequently, the greater the physical workload, the greater the need for SWS, and hence the establishment of a SWS:exercise relationship. The decrease in training volume in our swimmers during taper was accompanied by a decrease in SWS toward levels similar to nonexercising women (15-22% TST)(9,27,51). With similar workloads (training distances) at the onset and peak training periods, the amount of time spent in SWS was similarly high. Some studies have shown that SWS is elevated in response to increased training load (5,46), and it has been suggested that there may be a ceiling or “maximum level of SWS in a single night” (43) beyond which increases in workload may not be matched by further increases in SWS. While we were not able to examine the effects of larger workloads on SWS, we did find that decreases in training volumes were associated with reductions in time spent in SWS.
Studies on the effect of the type of exercise on sleep have shown that SWS is more likely to be affected by aerobic training than by anaerobic training such as power lifting (49). The change in training type from predominantly medium intensity aerobic exercise (onset) to high intensity aerobic exercise plus maximal intensity anaerobic exercise training (peak) to low volume anaerobic exercise (taper) training sessions may have influenced our SWS results. In our subjects the lowest amount of SWS occurred at taper when the training volumes were lowest and coincided with high proportions of anaerobic exercise, which may have exacerbated the changes seen in SWS levels.
It has been proposed that it is not the exercise per se that causes changes in SWS, but habitual exercise or fitness(28,46). Increases in SWS after exercise have been shown in fit but not in unfit subjects (10). Furthermore, not only should subjects be physically fit, but daily exercise should be of relatively long duration and conducted at high rates of energy expenditure for SWS increases to be detected (48). Our highly competitive, elite female swimmers were physically fit at all three phases of their training program, yet SWS was significantly higher in the onset and peak periods than after precompetition taper. While the taper-training may not meet the criterion of adequate duration(48), considering the intensity of the exercise the energy expenditure would have been extremely high. Therefore, the decrease in SWS in our athletes at taper, after a full 6 months of exercise training, to SWS levels similar to those of sedentary controls contradicts the SWS:fitness theory and lends support to other relationships such as SWS:exercise or SWS:temperature.
Another factor that may have influenced SWS is body temperature since high, sustained rates of body heating, whether passive or caused by exercise, has been shown to increase SWS(18,19,20,44). However, head-cooling during exercise appears to prevent this rise in SWS(17). This thermal response to exercise in water is influenced by individual insulation (25,42). Although there was a significant reduction in body fat percent in our subjects from onset to peak training, the time spent in SWS was not affected. While we intended to measure overnight rectal temperatures, it was impossible to obtain consent from our subjects or their guardians. Therefore, in assessing the SWS:temperature relationship, we had to rely on the literature, which shows that while heat gain during swimming is related to both work intensity and water temperature (32), even high intensities of water exercise produces minimal changes in body temperature when compared with land exercise (24). It is therefore improbable that the increases in SWS during high training loads can be ascribed to changes in either insulation or body temperature.
Increases in lean body mass may be associated with decreased SWS(45). Although body fat decreased in our subjects in the first 3 months of training, it was also associated with significant reductions in body mass, suggesting that lean body mass was in fact unchanged. In this group of athletes, there was no clear association between changes in body mass and body fat percent with the amount of SWS, thus supporting the suggestion(45) that the relationship between SWS and body composition is not simple.
In the same vein, dietary variations, in particular negative energy situations such as severe restrictions, have been shown to affect sleep(8,21), while the high carbohydrate/low fat diet currently recommended for athletes has been associated with decreased SWS(40). Our subjects followed a high carbohydrate/low fat dietary regime consisting of proportionately more carbohydrate, less fat, and less protein but mean energy intake similar to other elite female swimmers of comparable age (13). Although the energy content was higher during peak training when energy demands were higher, dietary composition was not significantly changed owing to differences in training, and diet is therefore unlikely to have influenced sleep.
Sleep disturbance is often quoted as a side effect of excessive training although the precise nature of the disturbances varies(12,23). Since sleep disruptions may compromise sleep quality and thus may affect mood, the number of movements and shifts to awake-as indications of sleep disturbance-were included in the sleep analyses. Shifts to awake and time spent awake after sleep onset were not different at any of the study phases; however, subjects tended to have higher movement scores at the higher training volumes of onset and peak. This increase in movement may have been caused by mild muscle fatigue and soreness following the high training loads (7), and although it suggests some sleep disruption, the disruption was not excessive enough to affect either subjective quality of sleep or mood parameters.
The most positive mood state occurred at the heaviest training load, while the poorest mood occurred at the lowest load immediately prior to national competition. These results are in contrast with those observed by others(12,29,31,34), where short-term manipulations in training load affected mood in a dose-dependent manner. In previous studies, increased training volume was associated with increased mood disturbance while mood improved in response to decreased training. Morgan et al. (29) examined mood in swimmers at the start, middle, and end of their season and found that the greatest mood disturbances occurred midseason in association with the highest workloads (10 km·d-1 training). A possible explanation for the more positive mood in our swimmers at the midseason peak following a high training load is that our subjects were not overtrained. The swimmers in the Morgan study (29) showed a reduction in performance, and those in the O'Connor study(33) had elevated cortisol levels, both factors associated with overtraining. The improvement in performance time and low levels of tension and anger (31) at peak training, as well as the consistency of early morning pulse rate, all support the contention that the swimmers in our study were not overtrained.
The high fatigue scores observed in our swimmers at onset and peak training are concomitant with the high training volumes (50); yet the high vigor levels at peak are unexpected. Our finding that the high intensity anaerobic exercise of the taper period had a significantly higher depression level than the other two periods (which included aerobic sessions) fits in well with the theory that aerobic exercise decreases depression(29).
Possible explanations for the poor mood in association with reduced training volume are “exercise addiction” (4) and the stress imposed by the impending competition. However, since exercise addiction was not investigated in this study, the origin of the poor mood at taper requires evaluation.
In conclusion, the decrease in SWS in response to the reduced physical demands associated with training taper supports the hypothesis that SWS is physically restorative. Ideally, it would be important to include nonexercising baseline results for our subjects since this would allow us to see whether increases in training volume at the start of the swimming season were associated with increases in SWS proportions. However, as competitive athletes, these subjects maintain a significant level of indoor training and gym-work throughout the year.
Of particular importance was the poor mood associated with the reduction in training and/or the impending competition, indicating a disadvantageous attitude in the swimmers immediately prior to national competition. Since performance did improve across the season, it is possible that the poor mood in these swimmers was benign. On the other hand, poor mood is a factor that may have prevented even greater improvement at Nationals and as such is an area that needs addressing. Physiological taper is essential for improvement in performance, so that the poor mood prior to competition cannot be handled by maintaining elevated training loads; it must, instead, be handled by mental preparation and competent redressing of the negative attitudes in these competitive female swimmers.
1. Adam, K. and I. Oswald. Sleep is for tissue restoration.J. R. Coll. Physicians Lond.
2. Agnew, H. W., W. B. Webb, and R. L. Williams. The first night effect: an EEG study of sleep. Psychophysiology
3. Almirall, H., A. Aguirre, R. V. Rial, A. Daurat, J. Foret, and O. Benoit. Temperature drop and sleep: testing the contribution of SWS in keeping cool. Neuroreport
4. Anthony, J. Psychologic aspects of exercise. Clin. Sports Med.
5. Baekeland, F. and R. Lasky. Exercise and sleep patterns in college athletes. Percept. Mot. Skills
6. Berger, R. J. and N. H. Phillips. Comparative aspects of energy metabolism, body temperature and sleep. Acta Physiol. Scand.
133(Suppl. 574):21-27, 1988.
7. Costill, D. L., M. G. Flynn, J. P. Kirwan, et al. Effects of repeated days of intensified training on muscle glycogen and swimming performance. Med. Sci. Sports Exerc.
8. Crisp, A. H. and E. Stonehill. Aspects of the relationship between sleep and nutrition: a study of 375 psychiatric outpatients. Br. J. Psychiatry
9. Driver, H. S., D-J. Dijk, E. Werth, K. Biedermann, and A. A. Borbély. Sleep and the sleep electroencephalogram across the menstrual cycle in young healthy women. J. Clin. Endocrinol. Metab.
10. Driver, H. S., A. F. Meintjies, G. G. Rogers, and C. M. Shapiro. Submaximal exercise effects on sleep patterns in young females before and after an aerobic training programme. Acta Physiol. Scand.
133(Suppl. 574):8-13, 1988.
11. Durnin, J. V. G. A. and J. Wormersley. Body fat assessed from total body density and its estimation from skinfold thickness. Measurement on 481 men and women aged 16 to 72 years. Br. J. Nutr.
12. Fry, R. W., J. R. Grove, A. R. Morton, P. M. Zeroni, S. Gaudieri, and D. Keast. Psychological and immunological correlates of acute overtraining. Br. J. Sports Med.
13. Grandjean, A. C. and J. S. Ruud. Energy intake of athletes. In: Oxford Textbook of Sports Medicine
. M. Harries, C. Williams, W. D. Stanish, and L. J. Micheli (Eds.). Oxford: Oxford University Press, 1994, pp. 53-65.
14. Griffin, S. J. and I. Trinder. Physical fitness, exercise and human sleep. Psychophysiology
15. Hooper, S. L., L. T. MacKinnon, A. Howard, R. D. Gordon, and A. W. Bachmann. Markers for monitoring overtraining and recovery.Med. Sci. Sports Exerc.
16. Horne, J. Human slow wave sleep: a review and appraisal of recent findings, with implifications for sleep functions, and psychiatric illness. Experientia
17. Horne, J. A. and V. J. Moore. Sleep EEG effects of exercise with and without additional body cooling. Electroencephalogr. Clin. Neurophysiol.
18. Horne, J. A. and A. J. Reid. Nighttime EEG sleep, EEG changes following body heating in a warm bath. Electroencephalogr. Clin. Neurophysiol.
19. Horne, J. A. and L. H. E. Staff. Exercise and sleep: body heating effects. Sleep
20. Jordan, J., I. Montgomery, and J. Trinder. The effect of afternoon body heating on body temperature and slow wave sleep.Psychophysiology
21. Karklin, A., H. S. Driver, and R. Buffenstein. Restricted energy intake affects nocturnal body temperature and sleep patterns. Am. J. Clin. Nutr.
22. Kirwan, J. P., D. L. Costill, M. G. Flynn, et al. Physiological responses to successive days of intense training in competitive swimmers. Med. Sci. Sports Exerc.
23. Lehmann, M., C. Foster, and J. Keul. Overtraining in endurance athletes: a brief review. Med. Sci. Sports Exerc.
24. McArdle, W. D., J. R. Magel, G. R. Lesmes, and G. S. Pechar. Metabolic and cardiovascular adjustment to work in air at 18, 25, and 33°C. J. Appl. Physiol.
25. McArdle, W. D., J. R. Magel, R. J. Spina, T. J. Gergley, and M. M. Toner. Thermal adjustment to cold-water exposure in exercising men and women. J. Appl. Physiol.
26. McNair, D. M., M. Lorr, and L. F. Droppleman (Eds.).Profile of Mood States Manual.
San Diego, CA: Edits Educational and Industrial Testing Service, 1992.
27. Meintjies, A. F., H. S. Driver, and C. M. Shapiro. Improved physical fitness failed to alter the EEG patterns of sleep in young women. Eur. J. Appl. Physiol.
28. Montgomery I., J. Trinder, G. Fraser, and S. J. Paxton. Aerobic fitness and exercise: effect on the sleep of younger and older adults.Aust. J. Psychol.
29. Morgan, W. P., D. R. Brown, J. S. Raglin, P. J. O'Connor, and K. A. Ellickson. Psychological monitoring of overtraining and staleness. Brit. J. Sports Med.
30. Morgan, W. P., D. L. Costill, M. G. Flynn, J. S. Raglin, and P. J. O'Connor. Mood disturbance following increased training in swimmers. Med. Sci. Sports Exerc.
31. Murphy, S. M., S. J. Fleck, G. Dudley, and R. Callister, Psychological and performance concomitants of increased volume training in elite athletes. Appl. Sport Psychol.
32. Nadel, E. R., I. Holmer, U. Bergh, P-O. Astrand, and J. A. J. Stolwijk. Energy changes of swimming man. J. Appl. Physiol.
33. O'Connor, P. J., W. P. Morgan, and J. S. Raglin. Psychobiologic effects of 3 d of increased training in female and male swimmers. Med. Sci. Sports Exerc.
34. O'Cornnor, P. J., W. P. Morgan, J. S. Raglin, C. M. Barksdale, and N. H. Kalin. Mood state and salivary cortisol levels following overtraining in female swimmers. Psychoneuroendocrinology
35. Oswald I. Sleep, the great restorer. New Sci.
36. Oswald I. The normal record of sleep. In: A Textbook of Clinical Neurophysiology
. A. M. Halliday, S. R. Butler, and R. Paul (Eds.). London: John Wiley and Sons, 1987, pp. 173-185.
37. Paxton, S. J., I. Montgomery, J. Trinder, J. Newman, and A. Bowling. Sleep after exercise of variable intensity in fit and unfit subjects. Aust. J. Psychol.
38. Paxton, S. J., J. Trinder, and I. Montgomery. Does aerobic fitness affect sleep? Psychophysiology
39. Paxton, S. J., J. Trinder, I. Montgomery, I. Oswald, K. Adam, and C. Shapiro. Body composition and human sleep. Aust. J. Psychol.
40. Phillips, F., A. H. Crisp, B. McGuinness, et al. Isocaloric diet changes and electroencephalographic sleep. Lancet
41. Rechtscaffen, A. and A. Kales. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects
. Washington, DC: US Government Printing Office, pp. 1-60, 1968.
42. Sagawa, S., K. Shiraki, M. K. Yousef, and N. Konda. Water temperature and intensity of exercise in maintenance of thermal equilibrium. J. Appl. Physiol.
43. Shapiro, C. M. Energy expenditure and restorative sleep. Biol. Psychol.
44. Shapiro, C. M., M. Allan, H. Driver, and D. Mitchell. Thermal load alters sleep. Biol. Psychol.
45. Shapiro, C. M., H. Driver, K. Cheshire, A. Carver, and J. Hamman. Sleep and body composition. In: Advances in in Vivo Body Composition Studies
, S. Yasamura, J. E. Harrison, K. G. McNeill, A. D. Woodhead, and F. A. Dimanian (Eds.). New York: Plenum Press, 1990, pp. 231-235.
46. Shapiro, C. M., P. M. Warren, J. Trinder, et al. Fitness facilitates sleep. Eur. J. Appl. Physiol.
47. Tipton, M. J. and F. Golden. Immersion in cold water: effects on performance and safety. In: Oxford Textbook of Sports Medicine
, M. Harries, C. Williams, W. D. Stanish and L. J. Micheli (Eds.). Oxford: Oxford University Press, 1994, pp. 205-217.
48. Trinder, J., I. Montgomery, and S. J. Paxton. The effect of exercise on sleep: the negative view. Acta Physiol. Scand.
133(Suppl. 574):14-20, 1988.
49. Trinder, J., S. J. Paxton, I. Montgomery, and G. Fraser. Endurance as opposed to power training: their effect on sleep.Psychophysiology
50. Verde, T., S. Thomas, and R. J. Shephard. Potential markers of heavy training in highly trained distance runners. Br. J. Sports Med.
51. Williams, R. L., I. Karacan, and C. J. Hursch.Electro-encephalography (EEG) of Human Sleep
. New York: John Wiley and Sons, 1974, p. 73.
EXERCISE; SWIMMING; MOOD; WOMEN; ATHLETES