The Impact of Circadian Phenotype and Time since Awakening on Diurnal Performance in Athletes

Transcription

1 Report The Impact of Circadian Phenotype and Time since Awakening on Diurnal Performance in Athletes Graphical Abstract Authors Elise Facer-Childs, Roland Brandstaetter Correspondence In Brief Facer-Childs and Brandstaetter report significant differences of daily physical performance between circadian phenotypes in athletes. This study establishes that circadian phenotype and time since entrained awakening, i.e., entrainment status of the circadian system reflecting internal biological time, are major determinants of athletic performance. Highlights d Athlete performance shows significant diurnal variation d d d Personal best performance times differ significantly between circadian phenotypes Internal biological time is the most reliable predictor of peak performance time Diurnal performance variations can be as pronounced as 26% in the course of a day Facer-Childs & Brandstaetter, 2015, Current Biology 25, February 16, 2015 ª2015 Elsevier Ltd All rights reserved

2 Current Biology 25, , February 16, 2015 ª2015 Elsevier Ltd All rights reserved The Impact of Circadian Phenotype and Time since Awakening on Diurnal Performance in Athletes Report Elise Facer-Childs 1 and Roland Brandstaetter 1, * 1 School of Biosciences, University of Birmingham, Birmingham B15 2TT, UK Summary Circadian rhythms, among other factors, have been shown to regulate key physiological processes involved in athletic performance [1 7]. Personal best performance of athletes in the evening was confirmed across different sports [8 12]. Contrary to this view, we identified peak performance times in athletes to be different between human larks and owls (also called morningness/eveningness types [13] or chronotypes [14] and referred to as circadian phenotypes in this paper), i.e., individuals with well-documented genetic [15 20] and physiological [21 24] differences that result in disparities between their biological clocks and how they entrain to exogenous cues, such as the environmental light/dark cycle and social factors. We found time since entrained awakening to be the major predictor of peak performance times, rather than time of day, as well as significant individual performance variations as large as 26% in the course of a day. Our novel approach combining the use of an athlete-specific chronometric test, longitudinal circadian analysis, and physical performance tests to characterize relevant sleep/wake and performance parameters in athletes allows a comprehensive analysis of the link between the circadian system and diurnal performance variation. We establish that the evaluation of an athlete s personal best performance requires consideration of circadian phenotype, performance evaluation at different times of day, and analysis of performance as a function of time since entrained awakening. Results We recruited 121 competition level athletes (70 females and 51 males; average age 22.5 years) to complete the RB-UB chronometric test, a novel chronometric questionnaire specifically designed to study sleep/wake-related parameters and training, competition, and performance variables in athletes. After comprehensive analysis and scoring of selected parameters, all individuals were categorized as either early circadian phenotype (ECT), intermediate circadian phenotype (ICT), or late circadian phenotype (LCT); 28% of the individuals were ECTs (n = 34), 48% were ICTs (n = 58), and 24% were LCTs (n = 29). This circadian phenotyping methodology proved consistent with relevant circadian parameters, such as wakeup times, sleep-onset times, and sleep durations, validating behavioral circadian differences between the individual phenotypes; wake-up times, both on weekdays and weekends, were significantly different between the circadian phenotypes (Kruskal-Wallis, p < ). Significant differences were also *Correspondence: r.brandstaetter@bham.ac.uk seen in sleep-onset times (Kruskal-Wallis, p < ) and sleep durations (Kruskal-Wallis, p < ) (Figures 1 and S1). From these 121 athletes, 20 with comparable age and fitness levels and with circadian phenotypes matching the whole population i.e., 25% versus 28% ECTs (n = 5), 50% versus 48% ICTs (n = 10), and 25% versus 24% LCTs (n = 5) were selected to conduct BLEEP fitness tests at six different times of day. All 20 were field hockey players with an average age of 20.4 years competing at regional club level, with seven out of these 20 individuals additionally competing at international level. The BLEEP test is a progressive aerobic cardiovascular endurance test widely used by sports coaches to estimate athletes maximum oxygen uptake, i.e., cardiovascular fitness, one of the most important components of physical fitness [25, 26]. Analysis of personal best BLEEP test performance of all 20 subjects revealed average lowest performance at 07:00 a.m., intermediate performance values at 10:00 a.m., 1:00 p.m., and 10:00 p.m., and highest performance values at 4:00 p.m. and 7:00 p.m., with a considerable performance difference of 11.2% between the minimum and maximum average performance (Figure 2A). Analysis considering circadian phenotype, however, revealed significant differences in peak performance, with the highest performance for ECTs at hr, ICTs at hr, and LCTs at hr (Figures 2B 2D). Diurnal changes in performance were 7.62% % in ECTs as compared to 10.03% % in ICTs and a striking 26.2% % in LCTs (Figure 2). Analysis of the data as a function of time since entrained awakening, i.e., performance evaluated against time in hours after entrained wake-up time, diminished the time difference between peak performance times in ECTs and ICTs, with the highest average performance for ECTs at hr and ICTs at hr, i.e., being only 0.96 hr apart and not significantly different any longer (Kruskal-Wallis, p > 0.05; Figures 3B, 3C, and 4). Average LCT peak performance time, however, was hr after entrained wake-up and was significantly delayed as compared to ECT and ICT peak performance times (Kruskal-Wallis, p < 0.01; Figures 3D and 4). Thus, our study of cardiovascular endurance, a major component of physical fitness, establishes that circadian phenotype and time since entrained awakening, i.e., the entrainment status of the circadian system reflecting internal biological time, are major determinants of diurnal athletic performance. Discussion Our results shed new light on our understanding of personal best performance in athletes by showing (1) significant differences in peak performance times between circadian phenotypes, (2) time since entrained awakening to be the major and most reliable predictor of peak performance, and (3) significant individual performance variations up to 26% in the course of a day. It is recognized that different sports require different skills, such as cognitive abilities, muscle strength, accuracy, or combinations of these [27]. Further factors affecting sports performance are diet [28, 29], motivation and competition [30], and

3 519 Figure 1. Analysis of Relevant Sleep/Wake Parameters Validating Circadian Phenotyping (A) Wake-up time on weekdays (WU WD). (B) Wake-up time on weekends (WU WE). (C) Sleep onset on weekdays (SO WD). (D) Sleep onset on weekends (SO WE). (E) Sleep duration on weekdays (SD WD). (F) Sleep duration on weekends (SD WE). White boxes represent early circadian phenotypes (ECTs), light-gray boxes are intermediate circadian phenotypes (ICTs), and late circadian phenotypes (LCTs) are shown as dark-gray boxes. Data are shown as Tukey boxplots; the line in the box indicates the median, the mean value is represented by the + symbol, and whiskers represent 1.5 times the interquartile or highest/lowest point distance. Statistical analysis was carried out using the Kruskal-Wallis test combined with Dunn s multiple-comparison post test. ns, not significant; *p < 0.05, **p < 0.01, ***p < See also Figure S1. muscle fatigue [31]. Our study shows that circadian phenotype, i.e., the entrainment status of the circadian system, is also a major determinant of athletic performance. Previous performance studies failed to distinguish between these different types [8 10, 12] or classified all participants, sometimes with the exception of only one or two individuals, as intermediate types [4, 11]. Our study, on the other hand, included performance tests of ECTs, ICTs, and LCTs of the same gender, comparable age, and comparable fitness levels, and we report significant differences in peak performance times between the circadian phenotypes, with ECTs performing their personal best around mid-day, ICTs performing best mid-afternoon, and LCTs showing peak performance in the evening. Strikingly, performance variation in the course of the day differed considerably between circadian phenotypes. While ECTs and ICTs showed comparable performance Figure 2. Diurnal Performance Variation as a Function of Time of Day (A) Performance values of all subjects (n = 20) expressed as percentage of individual personal best performance. (B) Performance values of ECTs (n = 5) expressed as percentage of individual personal best performance. (C) Performance values of ICTs (n = 10) expressed as percentage of individual personal best performance. (D) Performance values of LCTs (n = 5) expressed as percentage of individual personal best performance. The x axes show the time of day in hours. Curve fits are second-order polynomial non-linear regressions. Symbols represent the median 6 interquartile range. See also Figure S2 and Table S1. differences in the range of 7% to 10%, LCT performance varied substantially, by 26% on average. These enormous performance differences may have a big impact on talent finding, performance evaluation, and success in competition, and may explain why previous studies had identified international elite athletes mostly to be early types [32, 33]. In the sports world, a competitive advantage can be as little as 1%; at the 2008 Beijing Olympics, for example, a 1% increase in the 9.93 s time gained by fourth place in the men s 100 m sprint would have resulted in the silver medal. Similarly, for the women s road race, 400 m swim, and 400 m sprint, a 1% improvement would have won a gold medal for the fourthplace competitor. A major impact of our study comes from the analysis of performance as a function of time since entrained awakening. While time of day analysis revealed that ECTs, ICTs, and LCTs performed best at different times of day, these results changed significantly as a function of time since entrained awakening. Time of day is an exogenous factor and is only partly related to the circadian physiology of an individual, and our data show that measurements of diurnal performance as a function of time of day have only limited value. Irrespective of the time of day, ECTs wake up earlier and go to sleep earlier than LCTs; thus, their individual periods of wakefulness, i.e., their biological days, differ significantly from each other. Our 10:00 a.m. performance test, for example, took place about

4 520 Figure 4. Peak Performance Times as Functions of Time of Day and Time since Entrained Awakening (A) Peak performance times in real time, i.e., time of day in hours. (B) Peak performance times expressed as time since entrained awakening in hours. White bars represent ECTs, light-gray bars are ICTs, and LCTs are shown as dark-gray bars. Data are shown as Tukey boxplots; the line in the box indicates the median, the mean value is represented by the + symbol, and whiskers represent 1.5 times the interquartile or highest/lowest point distance. Statistical analysis was carried out using the Kruskal-Wallis test combined with Dunn s multiple-comparison post test. ns, not significant; *p < 0.05, **p < 0.01, ***p < Figure 3. Diurnal Performance Variation as a Function of Time since Awakening BLEEP test performance values of all subjects (n = 20; A), ECTs (n = 5; B), ICTs (n = 10; C), and LCTs (n = 5; D) expressed as the percentage of individual personal best performance. The x axes show the time since awakening in hours. Symbols represent individual performance test results. Lines are second-order polynomial non-linear regressions. See also Figure S2 and Table S1. 3 hr after the average wake-up time of the ECT participants and 2 hr after the entrained wake-up time of our ICT participants, but only 15 min after the entrained average wake-up time of the LCTs. Evaluating our data as a function of time since awakening revealed outstanding results; ECT and ICT performance curves were nearly identical, demonstrating that the time difference of peak performance between ECTs and ICTs was more or less entirely caused by the distinct phasing of their sleep/wake cycles. The diurnal performance curve of LCTs, however, showed a distinct shape and slope as compared to ECTs and ICTs, suggesting differences in the underlying physiology controlling performance. Possible explanations for this discrepancy come from endocrine studies showing that ECTs have higher cortisol levels in the morning and a distinct highamplitude diurnal profile of cortisol, while LCTs have lowered cortisol in the morning and a flattened diurnal profile that cannot be explained by different wake-up times and sleep durations and thus suggest that intrinsic physiological features are responsible for these differences [34 37]. Cortisol production is controlled by circadian mechanisms [38, 39], and LCTs have been shown to have significantly delayed melatonin rhythms as compared to ECTs [24]. This suggests that the phase-shifted, i.e., delayed, circadian rhythm in LCTs may cause a partial suppression and delay of cortisol, which in turn negatively affects physical performance, as cortisol is essential for muscle function [40, 41]. We can exclude impact of age as well as external influences on the results of this study as all participants were of comparable age, completed daily sleep/wake diaries (Figure S1), and comprehensive pre- and post-test forms with each performance test, monitoring a large number of variables, including sleep/wake times and food and caffeine intake. Additionally, to exclude any impact due to changes in sleep duration the night before the actual performance test, we re-analyzed all data by omitting all performance tests that took place earlier than entrained wake-up, i.e., that required a change in wake-up time and thus shortened sleep; this analysis confirmed the differences in peak performance times between the circadian phenotypes (Table S1). For ECTs and ICTs, peak performance times were similar about 5.5 hr and 6 hr after entrained wake-up, respectively, while LCTs reached their peak performance about 11 hr after their biological start of the day (Figure 3). Thus, the differences in peak performance times are the consequence of both internal physiological mechanisms and differential entrainment of the circadian system to environmental cues. To further validate these striking results and ensure that these results were not specific to one particular performance test and/or the particular group of athletes selected, we conducted additional performance tests with an independently selected group of squash players showing highly comparable performance results (Figure S3). Our results are the first known performance data in athletes that have observed different peak performance between circadian phenotypes in both real time and time since awakening. Desynchronization of internal body clocks can result from sleep disruptions, jet lag, shift work, and various other circadian disorders [42], including the mismatch of internal biological time and exogenous environmental time, a phenomenon that a high proportion of individuals in the current population, particularly LCTs, experience on a daily basis; all of these circadian disruptions are known to have detrimental effects on performance, health, and well-being [43 47]. With our increasing knowledge of the impact of circadian disruptions and novel tools to study circadian phenotypes, such as the ones introduced in this study, effective interventions can be designed to minimize circadian disruption, stabilize circadian rhythmicity, and enhance well-being and performance [48, 49]. Studies like this will provide athletes and coaches in the sports world with new insights that will allow them to improve

5 521 performance, as well as also create awareness in the corporate world to adapt schedules to achieve maximum performance of the workforce and increase safety [50, 51]. Our results leave no doubt that the correct determination of an athlete s personal best performance requires consideration of circadian phenotype, performance evaluation at different times of day, and analysis of performance as a function of time since entrained awakening. For an athlete to optimize performance, entrained wake-up time appears to be the most important and reliable predictor of optimal performance. It does not necessarily matter at what time of day personal best performance has to be achieved; what matters for an athlete is how many hours after entrained wake-up the competition or performance evaluation takes place. We herewith introduce novel tools for performance evaluation and enhancement, including a chronometric test specifically designed for athletes and longitudinal sleep/wake diaries that allow a detailed analysis of circadian disruptions, contributing factors, and internal biological time. Supplemental Information Supplemental Information includes Supplemental Experimental Procedures, two figures, and one table and can be found with this article online at Author Contributions R.B. developed the chronometric test and longitudinal sleep/wake diary and designed the study. E.F.-C. collected the data. E.F.-C. and R.B. contributed equally to the data analysis and wrote the paper. Acknowledgments Our sincere thanks are due to all participating athletes for their endurance during weeks of repeated performance testing and their dedication for completing sleep/wake diaries and pre-/post-test forms with great accuracy. Our thanks are due to Christian Portz for assistance in conducting performance tests with the University of Birmingham squash squad players. Special thanks are due to Professors Steve Busby and Jeff Bale at the University of Birmingham for invaluable comments on earlier versions of the manuscript. Received: March 23, 2014 Revised: September 25, 2014 Accepted: December 11, 2014 Published: January 29, 2015 References 1. Ericsson, K.A., and Lehmann, A.C. (1996). Expert and exceptional performance: evidence of maximal adaptation to task constraints. Annu. Rev. Psychol. 47, Wedman, J.F., and Graham, S.W. (1998). Introducing the concept of performance support using the performance pyramid. J. Contin. Higher Educ. 46, Gauthier, A., Davenne, D., Martin, A., Cometti, G., and Van Hoecke, J. (1996). Diurnal rhythm of the muscular performance of elbow flexors during isometric contractions. Chronobiol. Int. 13, Callard, D., Davenne, D., Gauthier, A., Lagarde, D., and Van Hoecke, J. (2000). Circadian rhythms in human muscular efficiency: continuous physical exercise versus continuous rest. A crossover study. Chronobiol. Int. 17, Souissi, N., Gauthier, A., Sesboüé, B., Larue, J., and Davenne, D. (2002). Effects of regular training at the same time of day on diurnal fluctuations in muscular performance. J. Sports Sci. 20, Atkinson, G., Jones, H., and Ainslie, P.N. (2010). Circadian variation in the circulatory responses to exercise: relevance to the morning peaks in strokes and cardiac events. Eur. J. Appl. Physiol. 108, Chtourou, H., and Souissi, N. (2012). The effect of training at a specific time of day: a review. J. Strength Cond. Res. 26, Conroy, R.T., and O Brien, M. (1974). Proceedings: diurnal variation in athletic performance. J. Physiol. 236, 51P. 9. Rodahl, A., O Brien, M., and Firth, R.G. (1976). Diurnal variation in performance of competitive swimmers. J. Sports Med. Phys. Fitness 16, Arnett, M.G. (2002). Effects of prolonged and reduced warm-ups on diurnal variation in body temperature and swim performance. J. Strength Cond. Res. 16, Atkinson, G., Peacock, O., St Clair Gibson, A., and Tucker, R. (2007). Distribution of power output during cycling: impact and mechanisms. Sports Med. 37, Kline, C.E., Durstine, J.L., Davis, J.M., Moore, T.A., Devlin, T.M., Zielinski, M.R., and Youngstedt, S.D. (2007). Circadian variation in swim performance. J. Appl. Physiol. 102, Horne, J.A., and Ostberg, O. (1976). A self-assessment questionnaire to determine morningness-eveningness in human circadian rhythms. Int. J. Chronobiol. 4, Roenneberg, T., Wirz-Justice, A., and Merrow, M. (2003). Life between clocks: daily temporal patterns of human chronotypes. J. Biol. Rhythms 18, Katzenberg, D., Young, T., Finn, L., Lin, L., King, D.P., Takahashi, J.S., and Mignot, E. (1998). A CLOCK polymorphism associated with human diurnal preference. Sleep 21, Carpen, J.D., von Schantz, M., Smits, M., Skene, D.J., and Archer, S.N. (2006). A silent polymorphism in the PER1 gene associates with extreme diurnal preference in humans. J. Hum. Genet. 51, Vandewalle, G., Archer, S.N., Wuillaume, C., Balteau, E., Degueldre, C., Luxen, A., Maquet, P., and Dijk, D.J. (2009). Functional magnetic resonance imaging-assessed brain responses during an executive task depend on interaction of sleep homeostasis, circadian phase, and PER3 genotype. J. Neurosci. 29, Dijk, D.J., and Archer, S.N. (2010). PERIOD3, circadian phenotypes, and sleep homeostasis. Sleep Med. Rev. 14, Pedrazzoli, M., Secolin, R., Esteves, L.O., Pereira, D.S., Koike, Bdel.V., Louzada, F.M., Lopes-Cendes, I., and Tufik, S. (2010). Interactions of polymorphisms in different clock genes associated with circadian phenotypes in humans. Genet. Mol. Biol. 33, Kunorozva, L., Stephenson, K.J., Rae, D.E., and Roden, L.C. (2012). Chronotype and PERIOD3 variable number tandem repeat polymorphism in individual sports athletes. Chronobiol. Int. 29, Bailey, S.L., and Heitkemper, M.M. (2001). Circadian rhythmicity of cortisol and body temperature: morningness-eveningness effects. Chronobiol. Int. 18, Brown, F.M., Neft, E.E., and LaJambe, C.M. (2008). Collegiate rowing crew performance varies by morningness-eveningness. J. Strength Cond. Res. 22, Lack, L., Bailey, M., Lovato, N., and Wright, H. (2009). Chronotype differences in circadian rhythms of temperature, melatonin, and sleepiness as measured in a modified constant routine protocol. Nat. Sci. Sleep 1, Nováková, M., Sládek, M., and Sumová, A. (2013). Human chronotype is determined in bodily cells under real-life conditions. Chronobiol. Int. 30, Léger, L.A., Mercier, D., Gadoury, C., and Lambert, J. (1988). The multistage 20 metre shuttle run test for aerobic fitness. J. Sports Sci. 6, Léger, L., and Gadoury, C. (1989). Validity of the 20 m shuttle run test with 1 min stages to predict VO2max in adults. Can. J. Sport Sci. 14, Lericollais, R., Gauthier, A., Bessot, N., Sesboüé, B., and Davenne, D. (2009). Time-of-day effects on fatigue during a sustained anaerobic test in well-trained cyclists. Chronobiol. Int. 26, Bougard, C., Bessot, N., Moussay, S., Sesboue, B., and Gauthier, A. (2009). Effects of waking time and breakfast intake prior to evaluation of physical performance in the early morning. Chronobiol. Int. 26, Javierre, C., Calvo, M., Díez, A., Garrido, E., Segura, R., and Ventura, J.L. (1996). Influence of sleep and meal schedules on performance peaks in competitive sprinters. Int. J. Sports Med. 17, Duffy, J.F., and Wright, K.P., Jr. (2005). Entrainment of the human circadian system by light. J. Biol. Rhythms 20, Edwards, J.D., Wadley, V.G., Vance, D.E., Wood, K., Roenker, D.L., and Ball, K.K. (2005). The impact of speed of processing training on cognitive and everyday performance. Aging Ment. Health 9,

6 Lastella, M., Roach, G.D., Hurem, D.C., and Sargent, C. (2010). Does chronotype affect elite athletes capacity to cope with the training demands of elite triathlon? In Living in a 24/7 World: The Impact of Circadian Disruption on Sleep, Work and Health, C. Sargent, D. Darwent, and G.D. Roach, eds. (Australasian Chronobiology Society), pp Silva, A., Queiroz, S.S., Winckler, C., Vital, R., Sousa, R.A., Fagundes, V., Tufik, S., and de Mello, M.T. (2012). Sleep quality evaluation, chronotype, sleepiness and anxiety of Paralympic Brazilian athletes: Beijing 2008 Paralympic Games. Br. J. Sports Med. 46, Bailey, S.L., and Heitkemper, M.M. (1991). Morningness-eveningness and early-morning salivary cortisol levels. Biol. Psychol. 32, Backhaus, J., Junghanns, K., and Hohagen, F. (2004). Sleep disturbances are correlated with decreased morning awakening salivary cortisol. Psychoneuroendocrinology 29, Kudielka, B.M., Federenko, I.S., Hellhammer, D.H., and Wüst, S. (2006). Morningness and eveningness: the free cortisol rise after awakening in early birds and night owls. Biol. Psychol. 72, Oginska, H., Fafrowicz, M., Golonka, K., Marek, T., Mojsa-Kaja, J., and Tucholska, K. (2010). Chronotype, sleep loss, and diurnal pattern of salivary cortisol in a simulated daylong driving. Chronobiol. Int. 27, Weitzman, E.D., Fukushima, D., Nogeire, C., Roffwarg, H., Gallagher, T.F., and Hellman, L. (1971). Twenty-four hour pattern of the episodic secretion of cortisol in normal subjects. J. Clin. Endocrinol. Metab. 33, Krieger, D.T., Allen, W., Rizzo, F., and Krieger, H.P. (1971). Characterization of the normal temporal pattern of plasma corticosteroid levels. J. Clin. Endocrinol. Metab. 32, Crewther, B.T., Cook, C., Cardinale, M., Weatherby, R.P., and Lowe, T. (2011). Two emerging concepts for elite athletes: the short-term effects of testosterone and cortisol on the neuromuscular system and the doseresponse training role of these endogenous hormones. Sports Med. 41, Teo, W., Newton, M.J., and McGuigan, M.R. (2011). Circadian rhythms in exercise performance: implications for hormonal and muscular adaptation. J. Sports Sci. Med. 10, Roth, T. (2012). Appropriate therapeutic selection for patients with shift work disorder. Sleep Med. 13, Horne, J., and Moseley, R. (2011). Sudden early-morning awakening impairs immediate tactical planning in a changing emergency scenario. J. Sleep Res. 20, Horne, J. (2012). Working throughout the night: beyond sleepiness impairments to critical decision making. Neurosci. Biobehav. Rev. 36, Roenneberg, T., Kantermann, T., Juda, M., Vetter, C., and Allebrandt, K.V. (2013). Light and the human circadian clock. Handbook Exp. Pharmacol. 217, Samuels, C. (2009). Sleep, recovery, and performance: the new frontier in high-performance athletics. Phys. Med. Rehabil. Clin. N. Am. 20, , ix. 47. Short, M.A., Gradisar, M., Lack, L.C., and Wright, H.R. (2013). The impact of sleep on adolescent depressed mood, alertness and academic performance. J. Adolesc. 36, Smith, R.S., Efron, B., Mah, C.D., and Malhotra, A. (2013). The impact of circadian misalignment on athletic performance in professional football players. Sleep 36, Rosenberg, J., Maximov, I.I., Reske, M., Grinberg, F., and Shah, N.J. (2014). Early to bed, early to rise : diffusion tensor imaging identifies chronotype-specificity. Neuroimage 84, Atkinson, G., Batterham, A., and Drust, B. (2008). Is it time for sports performance researchers to adopt a clinical-type research framework? Int. J. Sports Med. 29, Schaefer, E.W., Williams, M.V., and Zee, P.C. (2012). Sleep and circadian misalignment for the hospitalist: a review. J. Hosp. Med. 7,

7 Current Biology Supplemental Information The Impact of Circadian Phenotype and Time since Awakening on Diurnal Performance in Athletes Elise Facer-Childs and Roland Brandstaetter

8 Supplemental Data Supplemental Figure 1 A B C Time Of Day (Hours) Circadian parameter ECT (A) ICT (B) LCT (C) Earliest wake up Latest wake up Average wake up 07.42±0.32 (***) 09.62±0.86 (***) 11.11±1.05 (***) Earliest sleep onset Latest sleep onset Average sleep onset 23.51±0.97 (***) 01.00±1.13 (***) 02.73±1.15 (***) Supplemental Figure 1, Related to Figure 1: Daily activity patterns from performance test participants recorded with sleep/wake diaries. A) Early circadian phenotype (ECT), B) Intermediate circadian phenotype (ICT), C) Late circadian phenotype (LCT). Horizontal black bars indicate time from wake up to sleep onset. Grey area indicates average activity period of ECT. These activity records and the adjacent table show the difference in phasing of the activity periods between the circadian phenotypes and the increasing variability of wake up times and sleep onset times from ECT to LCT. Statistical analysis revealed wake up times and sleep onset times to be significantly different between all three circadian phenotypes (Kruskal-Wallis, p<0.001). Bartlett s test for equal variances revealed significant differences between variances (p<0.0001) of wake up times between participants demonstrating the significantly increased variability, i.e. instability, of sleep/wake rhythmicity in LCTs as compared to ECTs. Each subject taking part in the performance tests completed daily sleep/wake dairies for a minimum of 3 weeks, covering the whole period of performance testing.

9 Supplemental Figure 2 Time since awakening (hrs) Time since awakening (hrs) Supplemental Figure 2, Related to Figures 2,3. Diurnal performance variation in squash players. Performance variation of early/intermediate (n = 16) and late (n = 6) circadian phenotypes conducting court sprints as a function of time of day (A) and time since entrained awakening (B). Performance variation of early/intermediate (n = 16) and late (n = 6) circadian phenotypes conducting accuracy/efficiency performance tests as a function of time of day (C) and as a function of time since entrained awakening (D). Symbols represent mean values ± SEM. Grey dots are ECTs/ICTs, black dots are LCTs. Lines (solid for LCTs; stippled for ECTs/ICTs) represent second order polynomial non-linear regression curve fits. Analysis of the data as a function of time since awakening, i.e. performance evaluated against time in hours after entrained wake-up time showed significantly different peak performance times of both performance tests in ECTs/ICTs at 5.28±1.03 hr and LCTs at 11.65±0.15 hr (Kruskal-Wallis, p<0.01). Time of day average peak performance for ECTs/ICTs was at 14.32±0.81 hr as compared to 22.74±0.26 hr in LCTs (Kruskal-Wallis, p<0.01).

10 Supplemental Table 1 Peak performance ECT ICT LCT A) Time since awakening - all tests included 05.60±1.44 (** to LCT) 06.54±0.74 (** to LCT) 11.18±0.93 (** to ECT/LCT) B) Time since awakening test after entrained wake-up only 06.20±0.94 (** to LCT) 05.52±0.93 (** to LCT) 11.82±0.83 (** to ECT/LCT) Statistical difference between A and B ns p=0.74 ns P=0.39 ns P=0.63 Supplemental Table 1, Related to Figures 2,3: Peak performance as a function of time since awakening. This table shows peak performance times from the performance tests in hockey players calculated from second-order polynomial nonlinear regressions for all tests (A) and after exclusion of all tests that took place before entrained wake up and, thus, only include performance test that took place after entrained wake up correcting for the possible effects of shortened sleep and disrupted wake-up time on the day of the performance tests (B). Peak performance times remained comparable, i.e. were not significantly altered, while significant differences between circadian phenotypes, as reported in Figure 3, were retained with ECTs and ICTs not showing significant differences but both, ECTs and ICTs, being significantly different from LCTs (**=p<0.01).

11 Supplemental Experimental Procedures Chronometric testing and circadian phenotyping The RBUB chronometric test collects information on wake up times, sleep onset times, sleep onset delays, sleep duration, alarm use, light exposure, food intake, exogenous schedules (work, training, competition, school and/or university timetables), sleep quality, daytime naps, periods of mental and physical high and low activity, energy drink consumption, alcohol consumption, caffeine consumption, and smoking. Completion of the chronometric test took athletes 10 minutes on average. For each individual, scores were allocated to wake up times and sleep onset times during weekdays, weekends, and free days, the time lag between weekday and weekend wake up times, self-reported times of high (mentally and physically active) and low (tiredness, fatigue) activity periods and meal times. Masking factors such as working hours, University timetables, and training schedules were considered when allocating scores. Scores represented time in hours and were used to categorize into early (ECT), intermediate (ICT), and late (LCT) circadian phenotypes. All data were collected according to the Human Ethics regulations of the University of Birmingham and treated anonymously according to the UK Data Protection Act The chronometric test is the property of the corresponding author, Dr Roland Brandstaetter, University of Birmingham; it is protected by copyright and other intellectual property laws and can be used on the condition that it is not copied, reprinted or disclosed to a third party either in whole or in part without the prior written consent of the University of Birmingham. For further information or to request utilization of the questionnaire, contact the corresponding author at r.brandstaetter@bham.ac.uk. Physical performance tests BLEEP tests were carried out with 20 athletes, all of which were hockey players of comparable fitness and competition levels ranging from regional club level to international level. The sample consisted of 5 ECTs, 10 ICTs, and 5 LCTs to mirror the circadian phenotype distribution of the whole population studied. All subjects completed at least five out of six tests, scheduled at six different times of day spread out over a period of three weeks (0700, 1000, 1300, 1600, 1900, and 2200 hr). We used a multi stage fitness test called the UK standard Bleep Test; the initial running velocity was 8.0 km/hr, which increased to 9.0 km/hr at the next stage and by 0.5 km/hr thereafter. The participants completed continuous 20m shuttle runs until exhaustion and the levels achieved were recorded. The level attained provides an estimation of maximal oxygen uptake. All performance results are expressed as percentage of individual maximum performance. The test facility was an indoor gym providing consistent lighting and temperature conditions throughout all performance tests. 20m lines were measured accurately and the same lines used for each of the tests. Warm up routines were kept consistent for each test and pre/post-test forms were completed by all participants for each test to account for external factors that

12 might have influenced performance and for recording the athletes self-assessment of their physical conditions and perception of the tests. Further 22 athletes (average age 19.6 years) from the University of Birmingham squash team participated in additional performance tests. Of these 22 participants, 16 were early and intermediate circadian phenotypes and 6 were late circadian phenotypes. 2 of the 22 participants were of international standard, 5 of national standard, 8 of regional/county standard and 7 of club standard. Performance tests were carried out at 0700, 1100, 1500, 1900, and 2300 hr on consecutive s over a period of 5 weeks, with the order of times of day selected randomly. The performance tests consisted of two components; an accuracy/efficiency test requiring squash-specific skills, and a timed fitness component, i.e. repeated squash court sprints. The accuracy/efficiency component required participants to hit as many shots as possible into a 6-inch target area adjacent to the side wall of the squash court within 2 minutes and performance was evaluated as total hits into the target zone as percent of personal best. The fitness component involved four consecutive sets of court sprints. The time difference between the first and the last sets of court sprints was calculated and performance plotted as court sprint time in percent of fastest court sprint time. All participants completed specifically designed sleep/wake diaries and pre/post test forms. Sleep/wake diaries were used to allow a longitudinal detailed analysis of circadian parameters on an individual and daily basis. Measures included alarm clock use, wake up times, sleep onset times, sleep onset latency, sleep duration, sleep quality, meal times, snack times, caffeine intake, training/competition times, self-reported high activity, self-reported tiredness and fatigue, and daytime naps. Pre/post test forms recorded parameters such as wake up times over the last two days before the performance test, sleep onsets of the last two nights before the test, sleep duration, food intake, caffeine intake and exercise schedules over the last 24 hours before the actual performance test, time of last meal and details of carbohydrate intake. For the calculation of time since entrained awakening, wake-up times of at least three consecutive days ahead of each performance test were taken from the sleep/wake diaries and the pre/post-test forms of all individuals and the average wake-up time considered as entrained wake-up time. All performance test results were then calculated as a function of time since entrained awakening, i.e. number of hours deviating from entrained wake-up time. Parametric and non-parametric curve fits were initially applied to the performance data and 2 nd order polynomial non-linear regression chosen for further analysis according to tests of goodness of fit, normality of residuals, and replicates test of lack of fit with GraphPad Prism software. Individual peak performance values were derived from the XY coordinates of the 2 nd order non-linear regression curves.