Sleep 13(6):526-532, Raven Press, Ltd., New York 1990 Association of Professional Sleep Societies Daytime Functioning and Nighttime Sleep Before, During, and After a 146-Hour Tennis Match *tjack D. Edinger, tgail R. Marsh, two Vaughn McCall, tc. William Erwin, and t Anne W. Lininger Departments of * Veterans Affairs and tpsychiatry, Duke University Medical Center, Durham, North Carolina, U.S.A. Summary: Two adult males (ages 31 and 35 years) were studied while they participated in a week-long marathon tennis match under conditions of extreme sleep restriction (4-5 h reductions per night). Polysomnographic monitoring was conducted on the two nights prior to the marathon, continuously throughout the match, and on two recovery nights. In addition, measures of daytime sleepiness, mood state, and cognitive performance were obtained during the course of the study. Despite undergoing marked sleep restriction, both players continued to obtain their usual (baseline) amounts of slow wave sleep throughout the marathon. Both players showed a gradually increasing tendency toward daytime dozing across the first few days of the marathon. This tendency decreased on the fifth day but increased again on the sixth day of the match. Also, both players showed a pre- to postmatch decline on some cognitive measures. However, the players differed markedly in their ratings of sleepiness, mood ratings, recovery sleep patterns, and endurance with respect to the demands of the match. Results appear to be consistent with previous laboratory studies in documenting the primacy of the "slow wave sleep drive." Given the marked differences observed between the players, research designed to identify factors that predict response to sleep loss seems to be warranted. Key Words: Daytime functioning-nighttime sleep---tennis match. Numerous laboratory investigations have evaluated the effects of restricted sleep schedules on nighttime sleep and daytime functioning. These studies have shown that sleep reductions of greater than 2 h per night dramatically alter sleep architecture so that slow wave sleep occurs at the expense of other sleep stages during both restriction and recovery phases (1,2). Accompanying these changes are increases in daytime sleepiness (3,4), decrements in both performance and mood (1-3), and a general intolerance for the sleep loss (5). Although these findings are relevant to many "real world" circumstances, studies of restricted sleep schedules outside the laboratory have been Accepted for publication May 1990. Address correspondence and reprint requests to Dr. Jack Edinger, Psychology Service (1 16B), V A Medical Center, 508 Fulton Street, Durham, NC 27705. 526
TENNIS MARATHON 527 limited in number and focus. In the current "field study," two adult males served as subjects while they participated in a week-long marathon tennis match under conditions of extreme sleep restriction (4-5 h reductions per night). These subjects underwent extensive polysomnographic monitoring, and measures of sleepiness, mood state, and cognitive performance were determined during the course of the study. Given the aforementioned laboratory research, it was hypothesized that (a) the extreme sleep restriction would result in an increase in slow wave sleep percent at the expense of other sleep stages; (b) recovery sleep would show an increase in slow wave sleep percent compared to baseline sleep; (c) a gradual increase in daytime sleepiness would occur during sleep restriction; (d) mood would decline over the course of the sleep restriction period; and (e) cognitive functioning would suffer as a result of the cumulative sleep loss. METHODS Subjects The subjects were fit, white males without complaints or histories of sleep problems. Player 1 (age = 31 years) reported that he habitually slept 5 h per night and 1.5 h during the day whereas player 2 (age = 35 years) reported 6.S h of sleep per night and no daytime napping. Minnesota MUltiphasic Personality Inventories and psychiatric interviews completed prior to the marathon suggested neither player had symptoms or a history of psychiatric illness. However, the players were experienced with marathon conditions since they together had previously participated in a marathon during which a Guiness Book world record was established for doubles tennis play. Procedures Baseline measures of mood and daytime sleepiness were obtained for a 2 week period during the month prior to the marathon. Players were provided record forms containing the Stanford Sleepiness Scale (SSS) (6) and a 100-point mood scale and were instructed to rate their daytime sleepiness and mood four times each baseline day (breakfast, noon, dinner, and bedtime). Subsequently, the players participated in a 146 h tennis match that began at 0810 h on September 25, 1988 and ended at 1010 h on October 1, 1988. Play was continuous except that players were granted 90 s breaks between odd numbered games and they earned 5 min of rest for every hour played. Rest time was pooled and used for a 1-h nap (0300-0400 h) on the second calendar day of the match and 2 h naps (0300-0500 h) on the third through the seventh calender days. Objective measures of sleep were obtained from polysomnography (PSG) performed on the two nights preceding the match, continuously throughout the 146 h marathon, and again on two recovery nights following the match. All PSG studies were conducted using Oxford Medilog 9000 (Oxford Medical, Inc., Clearwater, FL, U.S.A.) ambulatory cassette recording devices. The players slept in their own homes on pre- and postmatch nights but they took their 1-2 h naps in a health club adjacent to their tennis court. Player 1 refrained from his usual daytime napping practices on pre- and postmatch nights so the PSG captured all sleep obtained by the players during the study. The PSG consisted of two electroencephalogram (EEG) channels (CrAv Oz-Cz), bilateral electro-oculogram (EOG), submental chin electromyogram (EMG), electrocardiogram (ECG), and two channels of anterior tibialis EMG (the tibialis recording was omitted during the 146 h marathon period). Two experienced polysomnographers (G.R.M. and
528 J. D. EDINGER ET AL. W. V.M.) each blindly scored one-half of the taped data directly on the screen of the ~1:edilog scanner using standard scoring criteria (7) and a previously tested screening scoring method (8). In addition to the PSG measures, ratings of daytime sleepiness and mood were obtained throughout the marathon and into the recovery period. Players were given SSS/mood scale forms and instructed to provide ratings of sleepiness and mood four times each day (0800, 1200,1800, and 2300 h). These forms were collected from the players each day except for the day following the second recovery night. Players were instructed to return their rating forms for their final recovery day to the experimenters by mail. To assess the marathon's effects on cognitive performance, a brief psychometric battery consisting of the Digit Symbol subtest of the Wechsler Adult Intelligence Scale-Revised (9), the Wechsler Memory Scale (10), the Trail Making Test (11), and the Folstein Mini-Mental Status Exam (12) was used. The players were individually administered these tests between 1000 and 1300 h on the day before the match and again between 1030 and 1300 h on the day the match ended (i.e., the seventh calender day). Since the Wechsler Memory Scale consists of alternate forms (form I and form 2) for retest situations, form 1 was used in the premarathon battery whereas form 2 was used in the postmarathon testing. The performance indices derived from these tests were (a) the number of digit symbol items correctly completed with a 90 s time limit, (b) the memory quotient obtained from the Wechsler Memory Scale, (c) the total time required to complete section B of the Trail Making Test, and (d) the number of items answered correctly on the Folstein test. Objective and subjective measures RESULTS The ambulatory PSG devices were well tolerated by our subjects and produced technically acceptable and interpretable sleep/wake records throughout the study. Table I summarizes the changes in sleep stage architecture noted across the various phases of the study. This table shows that the sleep restriction occurring during the marathon resulted in a differential sleep stage deprivation in favor of slow wave sleep. Relative to pre match nights, both players showed modest reductions in stage 1 % and stage 2 %, marked reductions in REM sleep %, and dramatic increases in slow wave sleep % during the match. During recovery, neither player showed a continued enhancement of slow wave sleep %. For player 1, stage 1 %, stage 2 %, REM %, and slow wave sleep % were all at pre match levels on recovery nights. However, since the total sleep time increased during recovery, player I averaged 30 min more of slow wave sleep on recovery nights than he did on prematch nights. In contrast, player 2 showed a decrease in average slow wave sleep time from 54.2 min (prematch) to 35.8 min (recovery). Further, this player's recovery slow wave sleep % was less than half the amount seen during prematch nights whereas his stage 1 % and stage 2 % showed a slight decrease and REM % increased to 1.5 times the level observed prior to the match. Table 2 shows PSG-derived estimates of the players' daytime dozing during the marathon period. This dozing occurred during the players' between-game, 90 s breaks and consisted primarily of stages 1 and 2 sleep. For both players, dozing increased gradually across the first 4 days ofthe marathon, dropped dramatically on the fifth day,
TENNIS MARATHON 529 TABLE 1. Sleep architecture before, during, and after the marathon Parameter BI B2 MI M2 M3 M4 M5 M6 RI R2 Player I Total sleep time (min) 430.0 414.7 41.0 104.0 103.3 108.3 122.3 85.3 919.3 456.0 Slow wave sleep (min) 71.0 36.3 19.7 52.3 60.3 68.3 68.3 44.0 91.3 75.3 Slow wave sleep % 16.5 8.8 48.0 50.3 58.4 63.1 55.9 51.6 9.9 16.5 Stage I % 2.9 2.9 7.3 0.6 1.6 0.6 2.2 1.6 2.1 3.3 Stage 2 % 52.9 65.9 44.7 49.0 39.7 35.4 41.4 46.5 66.0 51.2 REM % 27.6 22.4 0.0 0.0 0.3 0.9 0.5 0.4 22.0 28.9 Player 2 Total sleep time (min) 373.0 323.3 41.3 99.3 97.0 109.7 121.7 88.3 592.3 418.7 Slow wave sleep (min) 58.0 50.3 2.3 47.7 36.0 33.0 64.0 47.3 40.3 31.3 Slow wave sleep % 15.5 15.6 5.6 48.0 37.1 30.1 52.6 53.6 6.8 7.5 Stage I % 2.4 1.2 6.5 2.0 3.1 1.8 1.8 3.0 1.5 1.0 Stage 2 % 61.1 73.3 87.9 48.7 59.8 61.1 31.8 34.7 70.8 60.4 REM % 20.9 9.9 0.0 1.3 0.0 7.0 13.7 8.7 20.9 31.1 BI and B2 = two prematch nights; MI-M6 = scheduled sleep periods (naps) during the marathon; RI and R2 = two recovery nights. All percentages are rounded to the nearest tenth of a percent. Because of rounding, sleep stage percentages may not add up to 100%. increased again on the sixth day, and dropped on the seventh. However, the players were not monitored during the afternoon hours on the day they completed the match, so the dozing recorded for day 7 likely underestimates the actual amount that occurred. Also shown in Table 2 are mean sleepiness and mood ratings for each player during the two baseline weeks (BWI and BW2), the seven calender days of the marathon (MI-M7), and the day after the first recovery night (RI). Neither player returned his sleepiness and mood ratings for the day after the second recovery night, so these data are absent from the tables. Both players showed a relative reduction in subjective sleepiness on the fifth marathon day compared to days 3, 4, 6, and 7. Similarly, both players showed a gradual decline in mood on days 2, 3, and 4 followed by an increase in mood on the fifth day. However, player 1 reported less sleepiness in general during the marathon and on the recovery day than he did during baseline, whereas the reverse was true for player 2. Player 1 also showed mood ratings at baseline levels or higher during and after the marathon but player 2 showed an opposite trend. Results of psychometric testing were mixed. For player 1, performance on the Digit Symbol test dropped from a score of 65 (prematch) to a score of 61 (postmatch) and his TABLE 2. Dozing time, SSS scor'es, and mood ratings Parameter BWI BW2 MI M2 M3 M4 M5 M6 M7 RI Player I Dozing time (min) 13.3 24.3 40.2 64.8 12.5 50.4 24.2 SSS scores 2.8 2.9 2.0 2.5 2.2 2.0 1.2 1.8 2.2 1.5 Mood rating 56.1 62.3 56.2 62.5 56.2 50.0 62.5 56.2 75.0 62.5 Player 2 Dozing time (min) 0.0 0.0 12.8 33.6 0.5 41.3 19.0 SSS scores 1.2 1.5 1.8 1.5 1.8 2.5 1.5 2.5 3.2 2.0 Mood ratings 92.9 83.9 87.5 93.8 87.5 81.2 93.8 62.5 68.8 75.0 BWI and BW2 = two baseline weeks; MI-M7 = seven calender days on which tennis play occurred; RI = day after first recovery night of sleep. SSS scores were ratings made on a seven-point scale (1 = "feeling active, and vital; alert; wide awake" and 7 = "almost in reverie; sleep onset soon, lost in struggle to remain awake"). Mood ratings were made on a 0 to 100 scale, with 0 = "an extremely negative mood" and 100 = "an extremely positive mood."
530 J. D. EDINGER ET AL. Wechsler memory quotient dropped from 106 (prematch) to 97 (postmatch). Digit Symbol scores for player 2 dropped from 51 (pre match) to 45 (postmatch) and his memory quotient dropped from 96 (prematch) to 83 (postmatch). In contrast, the Folstein test scores of both players remained unchanged. Player 1 made only one error on both the pre- and postmatch Folstein exams and player 2 made only two errors on this test during each testing session. Finally, player 1 showed a worsening of his performance on the Trail Making Test (prematch score = 75 s; postmatch score = 82 s), whereas player 2 showed an improvement on this test (prematch score = 89 s; postmatch score = 81 s). Behavioral observations and physical findings Over the course of the 7-day event, both players reported increasing fatigue, heaviness in their legs, and soreness in their feet. Player 1 remained fully oriented and developed only one obvious physical injury, a blister on his right (preferred) hand. This injury developed early on the first marathon day and was well healed by a postmatch physical exam. In contrast, player 2 developed swelling of his right (preferred) forearm near the end of the second marathon day and this injury required repeated bandaging and ice applications throughout the remainder of the match. By the sixth day, player 2 intermittently appeared to be disoriented and required frequent prompting and instruction to continue his tennis play. Also on the sixth. day, he developed swelling and soreness in his left achilles tendon. Postmatch physical findings included pain and edema in both the right forearm (in the distribution of the wrist extensors) and the left achilles tendon extending up to the soleus muscles. During a postmatch interview, player 2 admitted to having become disoriented and forgetful in regard to the sequence of events he was required to perform during the tennis games. He also reported experiencing mild derealization in the sense that the tennis court seemed progressively less familiar to him. DISCUSSION This field study tested a number of predictions derived from laboratory studies of sleep restriction. Consistent with our first prediction, both players displayed a relative sleep stage deprivation in favor of slow wave sleep while maintaining restricted sleep schedules. This finding replicates the results of laboratory research (1,2) and supports the theory that the drive for slow wave sleep is a dominant sleep drive. Although neither player showed the predicted increase in slow wave sleep percent during recovery, player I did show an increase (relative to prematch nights) in slow wave sleep time on recovery nights. This finding suggests that player 1 incurred a "slow wave sleep debt" during the match that was "paid back" during recovery. In the case of player 2, physical injuries suffered during the match likely contributed to sleep disruption and a reduction in slow wave sleep on recovery nights. Since neither player showed an increase in slow wave percent during recovery, the current results seemingly contradict previous studies (1,2) in which subjects showed a relative increase in slow wave sleep time and percent following sleep restrictions. However, in these previous studies, time in bed was held constant on baseline and recovery nights. In this study, time in bed increased dramatically during recovery (relative to prematch nights) and thus slow wave sleep time and percent did not covary. Like the results obtained for nighttime sleep, the findings with regard to daytime
TENNIS MARATHON 531 measures supported some but not all predictions. The PSG-derived dozing time generally showed the expected cumulative daytime sleepiness across the marathon. SSS scores showed only mild levels of sleepiness for both players during the marathon and player I actually reported less daytime sleepiness during the match than he did during baseline. However, the SSS is likely a far less sensitive measure of sleepiness than is the PSG-monitored dozing activity. Thus, these findings are generally consistent with previous studies (3,4) and suggest that sleep restriction contributes to daytime sleepiness. Contrary to prediction, neither player showed a consistent decline in mood across the match. Mood states remained near baseline levels through the first five marathon days. The completion of the match was associated with a mood enhancement in player 1 and a marked mood decrease in player 2. Thus, mood ratings primarily highlighted player differences in response to the match. Finally, our prediction with regard to the effects of the marathon on cognitive performance was, in part, supported. Although two of the four tests used failed to indicate a consistent decline in performance, both players showed a pre- to postmatch decline in memory (Wechsler Memory Scale) and perceptual-motor coding (Digit Symbol test). Given the design of this study, it is difficult to determine whether these declines resulted from loss of sleep, fatigue, decreased motivation, or other factors. Nevertheless, these cognitive changes seem to be noteworthy and may warrant attention in "real world" situations where individuals are required to perform for extended periods on restricted sleep. In addition, the differences observed between our two players warrant consideration. Measures of recovery sleep, daytime sleepiness, and mood suggested that player 2 was more taxed by the match than was player 1. It is likely that the physical injuries suffered by player 2 contributed to a reduced tolerance for the extended period of sleep restriction. However, why such significant injuries developed in only one of the players is difficult to determine from available data. Whether the contrasting resiliencies of the two players were merely coincidental or the result of dispositional differences remains a question for future investigations to address. Acknowledgment: The authors would like to extend their appreciation to Mary Jo Martin, Sandra Clenney, and William Brickel for their technical support and advice during the course of this investigation. REFERENCES I. Tilley AJ, Wilkinson RT. The effects of a restricted sleep regimen on the composition of sleep and on performance. Psychophysiology 1984;21:406-12. 2. Tilley AJ. Recovery sleep at different times of the night following loss of the last 4 hours of sleep. Sleep 1985;8: 129-36. 3. Herscovitch J, Broughton R. Sensitivity of the Stanford Sleepiness Scale to the effects of cumulative partial sleep deprivation and recovery over sleeping. Sleep 1981 ;4:83-92. 4. Carskadon MA, Dement WC. Cumulative effects of sleep restriction on daytime sleepiness. Psychophysiology 1981;18:107-13. 5. Johnson LC, Naitoh P, Moses JM, Lubin A. Variations in sleep schedules. Waking Sleeping 1977;1: 133-7. 6. Hoddes E, Zarcone VP, Smythe H, Phillips R, Dement WC. Quantification of sleepiness: a new approach. Psychophysiology 1973;1O:431...{,. 7. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system of sleep stages of human subjects. Los Angeles: UCLA Brain Information Service/Brain Research Institute, 1968. Sleep, Vol. ]3, No.6, 1990
532 J. D. EDINGER ET AL. 8. Hoelscher TJ, McCall WV, Powell J, Marsh GR, Erwin CWo Sleep scoring with the Oxford Medilog 9000: comparison to conventional paper scoring. Sleep 1989;12:133-9. 9. Wechsler DA. Manualfor the Wechsler Adult Intelligence Scale-Revised. New York: Psychological Corporation, 1981. 10. Wechsler DA. A standardized memory scale for clinical use. J Psychol 1945;19:87-95. 11. Army individual test. Manual of directions and scoring. Washington, D.C.: War Department, Adjudant General's Office, 1944. 12. Folstein MF, Folstein SE, McHugh PRo "Mini Mental State." J Psychiatr Res 1975;12:189-98.