Repeated sprint exercise is common across many sports and in experimental research (5,15,17). Performance determinants evaluated in repeated sprint tests include speed or power output and fatigue resistance (18). Quantification of maximum sprint speed/power shows good test-retest reliability (30). Conversely, the best method of quantifying fatigue index (FI) reports a coefficient of variation (CV) of ∼30% (16,17). This large variability may hamper understanding of regulatory processes underpinning performance during, and improvements gained from, repeated sprint exercise.
Pacing tactics may be used before or soon after exercise begins in a feed-forward fashion to prevent significant homeostatic disturbance and premature exercise termination (29). Billaut et al. (7) reported that prior knowledge of the required number of sprints influences power output during a repeated sprint protocol, suggesting that anticipatory self-regulated (SR) pacing may happen during repeated sprint exercise. However, self-regulation was confined to power production, as the recovery periods between sprints were fixed.
Glaister et al. (18) further investigated self-regulation of performance during 12 × 30 m running sprints by allowing subjects to choose their own recovery time based on individual perceptions of recovery. After 2 familiarization trials, subjects were able to self-regulate recovery to maintain a consistent performance (mean CV for recovery time between sessions 3 and 4 of 9.9%). Glaister et al. (18) suggested that these findings justify self-regulation of repeated sprinting as a reliable tool for individuals to quantify their level of fatigue and maintain the quality of repeated sprint sessions. Self-regulation of repeated sprint performance in line with individual physical capabilities would be beneficial in many sport and exercise training scenarios, particularly when individuals train in groups. However, the protocol used by Glaister et al. (18) could not quantify the accuracy of SR repeated sprinting. Therefore, it could not be determined whether subjects overestimated recovery time to allow them to maintain performance. This should be investigated, as the recovery time chosen would influence the physiological demand experienced during the bout (4). Full recovery, defined for the purposes of this study as a return to resting metabolic and intramuscular energy status, is not required for repeated sprint performance to be maintained (15). If SR recovery is overestimated, meaning that subjects give themselves more recovery than is actually necessary to maintain repeated sprint performance, then allowing athletes to self-regulate their performance may not generate the physiological load required to stimulate specific adaptations and performance enhancements or prepare athletes sufficiently for the demands of competition. This can be experimentally tested by establishing individual SR recovery and then reducing this recovery time in a blinded fashion. If such an approach alters physiological and perceptual responses and impairs repeated sprint performance, it would provide an insight into the accuracy of SR repeated sprinting. Currently, no specific research is available that addresses these issues.
The aim of this study was to investigate the reliability of SR performance during repeated sprint exercise and the accuracy of this self-regulation. It was hypothesized that after appropriate familiarization, self-regulation of repeated sprint exercise would allow maintenance of a stable performance level and that reducing SR recovery duration would impair repeated sprint performance.
Experimental Approach to the Problem
Learning effects exist between the first 2 trials of a cycle sprint test (22). Therefore, to ensure sufficient data for familiarization and reliability analysis, subjects completed 4 trials (18), each comprising 10 × 6 seconds cycle sprints on a Monark 894E mechanically braked cycle ergometer against a 7.5% body mass (BM) resistance. Subjects SR the recovery duration between each sprint with the goal of maintaining a stable power output across all sprints. All subjects maintained sprint performance across the 4 trials and therefore took part in a fifth trial. Recovery time was manipulated in a single-blind fashion to investigate the accuracy of SR repeated sprint performance. Each subject's most reliable SR performance from trial 3 or 4 (based on within-trial CV for mean power output [MPO]) was used as the criterion recovery time to manipulate. Each postsprint recovery time was reduced by 10% (reduced recovery [RR] trial). The ergometer was attached to specialist software (Monark Anaerobic Test Software 220.127.116.11; Vansbro, Sweden) that enabled calculation of peak power output (PPO), MPO, and FI for each sprint. Heart rate (HR), physical ratings of perceived exertion (P-RPE), and measures of task effort awareness (TEA) were recorded during each trial to provide an indication of physiological and psychological strain. Within subjects, all trials were conducted at the same time of day, with a minimum of 3 and maximum of 7 days between trials. Subjects completed a food diary for 24 hours before the first trial and were instructed to replicate this before each trial to control for the potential influence of alterations in energy intake on mood state (9) and performance (24). Subjects were asked to consume a light meal at least 2 hours before testing. Pretesting training was not standardized between subjects but was standardized within subjects by requesting that they refrain from strenuous exercise for at least 24 hours before each trial. Adherence to these procedures was verbally confirmed at the beginning of each trial.
Fourteen healthy recreationally active men (24.5 ± 5.0 years, 178 ± 8 cm, and 80.9 ± 13.2 kg) participated, some of whom had experience of repeated cycle sprinting. Subjects took part in a variety of sports (gym training, climbing, football, hockey, volleyball, martial arts) for a mean weekly duration of 6.5 ± 3.9 hours and a mean experience level of 8.1 ± 5.3 years. Subjects were informed of the nature of the investigation, after which they gave written informed consent. The study received approval from the institutional research ethics committee.
Body mass (kg) and standing height (cm) were recorded using a height stadiometer (model 245; Seca, Hamburg, Germany) and digital scale (model 708; Seca), respectively, while wearing shorts. Subjects then completed a standardized warm-up of 4-minute cycling at 60 rpm against a 1 kg resistance and 3 × 3 seconds maximal sprints against a 7.5% BM resistance interspersed with 45-second cycling against no resistance. They then dismounted and sat quietly for 3 minutes before the main component of the trial. In each trial, subjects were informed that they were to complete 10 × 6 seconds cycle sprints to give maximum effort in each sprint and to give themselves sufficient recovery so that in all 10 sprints they were able to replicate the performance achieved in the criterion sprint (instructions adapted from Glaister et al. (18)). Subjects were instructed to remain seated during all sprints. No external performance feedback was provided, but cadence was visible during recovery periods. Vigorous verbal encouragement was provided during every sprint. Subjects were instructed to give a 3-second countdown before starting each sprint and to factor this into their recovery. Recovery time was defined as the period from the end of the previous sprint until the beginning of the next sprint, immediately after the 3-second countdown. All sprints began from 60 rpm with resistance automatically applied to the flywheel on reaching 110 rpm.
A flowchart summarizing the experimental protocol is in Figure 1. Subjects were introduced to the equipment and procedures. They then undertook a single 6-second sprint to familiarize them with the procedure and provide criterion sprint data for comparison with repeated sprint performance. After the warm-up, subjects remounted the ergometer and cycled at 60 rpm against no resistance for 10 seconds, after which they cycled maximally. The load was automatically added to the ergometer on reaching 110 rpm, at which time the 6-second sprint began. On completion, participants cycled easily against a 1 kg resistance for 1 minute, then dismounted the ergometer, and sat quietly for 5 minutes. The test was repeated to identify whether a maximal effort was achieved in the first sprint. If subjects achieved a lower MPO in test 2, the result of test 1 was taken as MPO. If subjects achieved a MPO in test 2 ≥5% than test 1, a third test was undertaken. This was repeated as necessary until MPO no longer increased. A 15-minute seated recovery followed the criterion sprint test.
After the recovery, participants completed the standardized warm-up, then remounted the ergometer, and cycled at 60 rpm for 1 minute. The investigator provided a 3-second countdown, after which the subject completed 10 × 6 seconds cycle sprints against a 7.5% BM resistance with a self-regulated recovery between each sprint. During recovery, participants cycled at 50–60 rpm against no resistance.
Trials 2–4 followed a similar procedure to that of trial 1. However, only the warm-up and the 10 × 6 seconds sprints were completed.
After the first 4 trials, subjects' data were analyzed to determine whether they successfully maintained sprint performance in each trial. Performance maintenance was defined as:
- The absence of an obvious pattern of fatigue, determined by visual inspection of PPO and MPO data for each sprint (18), to confirm no continuous drop-off in performance.
- A within-trial CV for MPO of 5.2% or less (the upper CV of MPO for this type of exercise (10,18)).
All subjects successfully maintained performance in the first 4 trials and progressed to the final trial.
In this trial, SR recovery time was manipulated as described above. The 10% reduction in recovery times is greater than the random variation of recovery time previously reported during self-paced recovery of repeated sprints (18). However, before the session, subjects were informed that their most reliable sprint session was being replicated to investigate repeatability of performance. They were reminded that they should produce their best effort, but this time, the investigator would tell them when to begin each sprint. The investigator informed the subject when there was 10 seconds of a recovery period remaining and provided a 3-second countdown into the next sprint.
In addition to power data, HR was recorded (Polar S610i; Kempele, Finland) at 5-second intervals throughout each trial. Fatigue index was calculated using the formula (18):
where total sprint performance = sum of MPO from all sprints, and ideal performance = number of sprints × greatest MPO. Self-regulated recovery duration between each sprint was recorded with a digital stopwatch to the nearest second (11). Physical ratings of perceived exertion and TEA were recorded 5 seconds after every sprint using procedures described by Swart et al. (27). These scales separately quantify physical and psychological effort during exercise, enabling greater insight into the influence of these factors on exercise performance (27).
Between-trials reliability was assessed by calculating changes in the mean, intraclass correlation coefficient (ICC), CV, and 95% limits of agreement (LoA) using published spreadsheets (20,21). One-way repeated measures analysis of variance (ANOVA) compared mean recovery time between trials 1 and 4, and PPO and MPO between the criterion sprint and all sprints in the SR and RR trials. Physiological, perceptual, and performance measures from each subject's most reliable repeated sprint trial (based on within-trial recovery time) from the first 4 sessions was compared with the RR trial using a two-way (trial × sprint) ANOVA. The Greenhouse-Geisser adjustment was applied if the assumption of sphericity was violated, and post hoc Bonferroni's correction explored significant main effects. Pearson's correlations between sprint number and perceptual responses and between P-RPE and TEA were calculated for each subject. Effect sizes for significant main effects from ANOVA analysis were reported as partial eta-squared (
) and quantified as small (≤0.01), medium (>0.01 and <0.14), and large (≥0.14) (13). Cohen's d effect sizes quantified the magnitude of significant mean differences between trials (small, d ≤ 0.2; medium, d > 0.2 and d < 0.8; and large, d ≥ 0.8; (12)). Statistical significance was set at p ≤ 0.05, and data are mean ± SD unless otherwise stated.
Mean results for each performance variable across the 4 reliability trials are in Table 1. There were no differences across trials for any performance variable (p > 0.05). Reliability statistics are in Table 2. Random variation of recovery time (CV and LoA) was substantially reduced when comparing the final pair of trials with the earlier pairs of trials. Both MPO and PPO demonstrated ICCs >0.95 and CVs ≤3.3% in all comparisons. Conversely, there were high levels of random variation in FI.
As designed, mean recovery time was significantly reduced in the RR trial compared with the SR trial (86.2 ± 31.6 vs. 95.7 ± 35.2 seconds, p ≤ 0.05, d = 0.28), with the RR time lower than that chosen by each participant in trials 3 and 4. There was no significant effect of sprint number for mean recovery time (F[2.69,34.94] = 0.482, p > 0.05,
= 0.07). Power profiles across the SR and RR sprints, and compared with the criterion sprint, are in Figures 2A, B. There was no significant main effect for trial (F[1,13] = 0.134, p > 0.05,
= 0.01) or sprint number (F[3.66,47.54] = 0.820, p > 0.05,
= 0.06) and no interaction effect (F[4.13,53.74] = 0.973, p > 0.05,
= 0.07) for PPO. There was no main effect for trial (F[1,13] = 1.163, p > 0.05,
= 0.08) or sprint number (F[3.01,39.12] = 0.452, p > 0.05,
= 0.03) and no interaction effect (F[2.78,35.98] = 0.840, p > 0.05,
= 0.06) for MPO. There was no significant difference in MPO between the criterion sprint and any sprint in the SR (F[2.76,35.86] = 2.099, p > 0.05,
= 0.14) and RR trials (F[2.96,38.53] = 1.161, p > 0.05,
= 0.08). However, PPO in the criterion sprint was significantly greater than PPO in all sprints of the SR trial (F[3.40, 44.14] = 3.114, p ≤ 0.05,
= 0.19). In the RR trial, PPO in sprints 3–10 was significantly lower than the criterion sprint (F[10,130] = 2.621, p ≤ 0.05,
Perceptual responses to the SR and RR trials are in Figures 3A, B. There was no main effect of trial (P-RPE: F[1,13] = 0.034, p > 0.05,
= 0.0, and TEA: F[1,13] = 0.074, p > 0.05,
= 0.0) and no interaction effect between trials over time (P-RPE: F[3.05,39.70] = 0.920, p > 0.05,
= 0.07, and TEA: F[9,117] = 0.750, p > 0.05,
= 0.06). However, there was a significant time effect for both P-RPE (F[1.43,18.62] = 27.590, p ≤ 0.05,
= 0.68) and TEA (F[1.42,18.44] = 21.950, p ≤ 0.05,
= 0.63). The relationship between sprint number and perceived physical and psychological stress demonstrated wide interindividual variability in the SR (P-RPE: r2 = 0.07–0.98, and TEA: r2 = 0.21–0.88) and RR (P-RPE: r2 = 0.08–0.92, and TEA: r2 = 0.0–0.94) trials. Similarly, the relationship between P-RPE and TEA was r2 = 0.28–1.0 in the SR trial and r2 = 0–0.98 in the RR trial. Heart rate showed no main effect of trial (F[1,10] = 0.949, p > 0.05,
= 0.09) or interaction effect (F[2.78,27.83] = 0.708, p > 0.05,
= 0.07), but there was a main effect of time (F[3.54,35.35] = 41.269, p ≤ 0.05,
= 0.81; Figure 4).
After 2 familiarization sessions, subjects were able to maintain repeated sprint performance with relatively stable SR recovery periods. Reducing SR recovery duration by 10% did not impair maintenance of repeated sprint performance or affect psychophysiological ratings. Therefore, subjects overestimated required recovery time between sprints.
Table 2 shows a notable improvement in the reliability of SR recovery time between trials 1–2, 2–3, and 3–4. The high CV and low ICC for SR recovery time between trials 1–2 compared with trials 2–3 and 3–4 supports the suggestion of Hopkins et al. (22) that learning effects are evident between at least the first 2 trials of cycle sprint tests. The reliability of SR recovery between trials 3 and 4 in this study (CV = 7.5%, ICC = 0.97) is better than that reported by Glaister et al. (18) across the same trials (CV = 9.9%, ICC = 0.83) and is also below the imposed 10% reduction of recovery time in the RR trial. Better reliability may relate to the exercise mode (running vs. cycling) or to the subjects used in this study, some of whom had experience of repeated cycle sprinting. It should also be considered that maintenance of repeated sprint performance depends on sprint duration (1). Therefore, varying sprint duration may influence the ability to self-regulate performance. This should be considered when comparing results between studies and may represent an interesting avenue for further research.
Glaister et al. (18) reported a progressive increase in RPE during repeated sprints, despite a stable performance. This was attributed to subjects giving themselves just enough recovery between sprints. In this study, P-RPE and TEA scores progressively increased throughout both trials, with no significant between-trials differences. The present findings support the observation that although a self-selected recovery will allow performance to be maintained, perceived exertion progressively increases. However, the present findings do not support the suggestion that subjects pace recovery to give just enough time to maintain performance, as when recovery time was reduced by 10% performance was still maintained.
In this study, P-RPE was almost identical at the end of exercise in the SR and RR trials. However, the peak values (∼15) in this study and that of Glaister et al. (18) likely do not reflect the highest tolerable values that subjects could have attained. This is reinforced by the moderate peak TEA values in both trials. Short-duration sprinting is fueled by phosphocreatine (PCr; ∼50%) and glycolysis (∼40%), with a progressive aerobic contribution as sprint number increases (6). The duration of the recovery periods in this study would likely have enabled a continued large contribution of PCr to subsequent sprints, as the half-time of PCr resynthesis in adults is ∼27 seconds (28). Therefore, progressive intramuscular acidosis associated with the glycolytic contribution to the sprints may explain the progressive increase in P-RPE and TEA (18). It has also been shown that the aerobic contribution to repeated sprinting increases, as the number of sprints progresses (8). Increased aerobic contribution would require an increased cardiorespiratory demand, increasing afferent feedback, and potentially elevating RPE and TEA. The potential impact of increased intramuscular acidosis and cardiorespiratory demand may also explain the variable individual relationship between sprint number and perceived physical and psychological stress, as between-subjects differences in aerobic fitness and muscle morphology may have modulated metabolic responses (19,29), and hence, perceptual responses to the sprints. Blood lactate concentration was not measured in this study because of the large variability in blood lactate measures and the greater reliance on PCr as a fuel during repeated sprinting. Therefore, further investigation is required to elucidate these suggestions. Similar P-RPE, TEA, and HR between the SR and RR trials reinforces that when subjects are permitted to select their own recovery, they overestimate the recovery required to maintain performance by at least 10%.
Deception of the number of sprints (with known sprint and recovery duration) to be performed can significantly reduce PPO and work performed from the first sprint, suggesting the presence of a pacing strategy based on factors including the number of sprints required (7). From a practical perspective, pacing during repeated sprint exercise may impair training quality and fitness adaptations. In the SR trial, subjects produced a significantly lower PPO from sprint 1 compared with the criterion sprint. It therefore seems that when subjects were aware that they had to perform multiple sprints, even with a self-selected recovery, they produced submaximal power from the onset of exercise despite being asked to perform maximally. Submaximal power production could be the result of an anticipatory pacing strategy based on knowledge of the number of sprints to be completed (7) or it may be that experience of completing repeated 6-second sprints enabled the subjects to pace differently within each sprint, achieving a lower PPO but maintaining MPO (Figures 2A, B). In this study, it is not possible to determine the relative prevalence of these hypotheses. Billaut et al. (7) did not use a single criterion sprint. Therefore, the true maximal performance of their subjects was unknown, meaning inferences regarding pacing strategies could only be made by comparing between-trials sprint performance during exercise. By comparing repeated sprint performance with that of a single sprint, this study provides the first evidence for suboptimal performance from the onset of a known bout of repeated sprinting in recreationally trained subjects. This finding reinforces the presence of a pacing strategy based either on anticipation of the number of sprints to be completed, prior experience of the repeated sprint protocol, or a combination of the two.
It is well known that the type of pacing strategy used during exercise is influenced by previous related exercise experience (2,25) and the performance level of the subject (23). Possible determinants of the pacing strategies used by different standards of athlete include differences in physiological and psychological parameters (3,23) and the learned aspect of pacing that is developed through experience (14). This study used recreationally trained subjects. Therefore, it cannot be conclusively asserted that using SR recovery in more highly trained and/or experienced athletic populations would generate the same findings as reported in this study or would be a useful strategy for athletes. Future research should explore the influence and efficacy of SR recovery in more elite populations.
In conclusion, after 2 familiarization trials, repeated cycle sprinting performance can be reliably maintained when subjects self-regulate recovery. However, subjects also overestimate by at least 10% the recovery time needed to maintain sprint performance.
Self-regulated recovery seems to be a reliable option for maintaining the quality of repeated sprint exercise and resisting fatigue. This has particular practical relevance when training groups of individuals with differing repeated sprint abilities. Coaches could use SR repeated sprinting as a method of maintaining sprint quality tailored to individual performance rather than using a single fixed recovery period, which may not suit the ability of all individuals. However, this study has demonstrated that individuals overestimate the recovery time needed for maintenance of performance. Many sporting situations require repeated bouts of effort with minimal recovery (26). Therefore, if a goal of training is to prepare for this situation, then allowing individuals to self-regulate recovery may not stimulate the necessary metabolic adaptations for performance improvement. Coaches should be aware of the potential benefits and limitations of SR repeated sprinting and consider the use of SR recovery within the context of specific training aims. The findings of this study should also be treated as population specific, until subsequent study has been conducted in more elite populations to investigate whether high-performing athletes display similar responses to SR repeated sprint exercise.
1. Abt G, Siegler JC, Akubat I, Castagna C. The effects of a constant sprint-to-rest ratio and recovery mode on repeated sprint performance. J Strength Cond Res 25: 1695–1702, 2011.
2. Albertus Y, Tucker R, St Clair Gibson A, Lambert EV, Hampson DB, Noakes TD. Effect of distance feedback on pacing strategy and perceived exertion during cycling. J Med Sci Sports Exerc 37: 461–468, 2005.
3. Baden DA, Warwick-Evans L, Lakomy J. Am I nearly there? The effect of anticipated running distance on perceived exertion and attentional focus. J Sport Exerc Psychol 26: 215–231, 2004.
4. Balsom PD, Seger JY, Sjödin B, Ekblom B. Maximal-intensity intermittent exercise: Effect of recovery duration. Int J Sports Med 13: 528–533, 1992.
5. Billaut F, Basset FA. Effect of different recovery patterns on repeated-sprint ability and neuromuscular responses. J Sports Sci 25: 905–913, 2007.
6. Billaut F, Bishop D. Muscle fatigue
in males and females during multiple-sprint exercise. Sports Med 39: 257–278, 2009.
7. Billaut F, Bishop DJ, Schaerz S, Noakes TD. Influence of knowledge of sprint number on pacing during repeated-sprint exercise. Med Sci Sports Exerc 43: 665–672, 2011.
8. Bogdanis GC, Nevill ME, Boobis LH, Lakomy HK. Contribution of phosphocreatine and aerobic metabolism to energy supply during repeated sprint exercise. J Appl Physiol 80: 876–884, 1996.
9. Butki BD, Baumstark J, Driver S. Effects of a carbohydrate-restricted diet on affective responses to acute exercise among physically active participants. Percept Mot Skills 96: 607–615, 2003.
10. Capriotti PV, Sherman MW, Lamb DR. Reliability of power output during intermittent high-intensity cycling. Med Sci Sports Exerc 31: 913–915, 1999.
11. Castagna C, Impellizzeri FM, Rampinini E, D'Ottavio S, Manzi V. The Yo-Yo intermittent recovery test in basketball players. J Sci Med Sport 11: 202–208, 2008.
12. Cohen J. A power primer. Psychol Bull 112: 155–159, 1992.
13. Ellis PD. The Essential Guide to Effect Sizes: Statistical Power, Meta-analysis, and the Interpretation of Research Results. Cambridge, UK: Cambridge University Press, 2010.
14. Foster C, Hendrickson KJ, Peyer K, Reiner B, deKoning JJ, Lucia A, Battista RA, Hettinga FJ, Porcari JP, Wright G. Pattern of developing the performance template. Br J Sports Med 43: 765–769, 2009.
15. Gaitanos GC, Williams C, Boobis LH, Brooks S. Human muscle metabolism during intermittent maximal exercise. J Appl Physiol 75: 712–719, 1993.
16. Glaister M, Howatson G, Lockey RA, Abraham CS, Goodwin JE, McInnes G. Familiarisation and reliability of multiple sprint running performance indices. J Strength Cond Res 21: 857–859, 2007.
17. Glaister M, Howatson G, Pattison JR, McInnes G. The reliability and validity of fatigue
measures during multiple-sprint work: An issue revisited. J Strength Cond Res 22: 1597–1601, 2008.
18. Glaister M, Witmer C, Clarke DW, Guers JJ, Heller JL, Moir GL. Familiarization, reliability, and evaluation of a multiple sprint running test using self-selected recovery periods. J Strength Cond Res 24: 3296–3301, 2010.
19. Greenhaff PL, Nevill ME, Soderlund K, Bodin K, Boobis LH, Williams C, Hultman E. The metabolic responses of type I and II muscle fibres during maximal treadmill sprinting. J Physiol 478: 149–155, 1994.
20. Hopkins WG. Analysis of validity by linear regression (excel spreadsheet). In: A New View of Statistics [Internet]. New Zealand: Sportscience. Available at: http://sportsci.org/resource/stats/valid.xls
. Accessed July 1, 2013.
21. Hopkins WG. Precision of measurement. In: A New View of Statistics [Internet]. New Zealand: Sportscience. Available at: http://newstats.org/precision.html
. Accessed June 19, 2013.
22. Hopkins WG, Schabort EJ, Hawley JA. Reliability of power in physical performance tests. Sports Med 31: 211–234, 2001.
23. Lima-Silva AE, Bertuzzi RCM, Pires FO, Barros RV, Gagliardi JF, Hammond J, Kiss MA, Bishop DJ. Effect of performance level on pacing strategy during a 10-km running race. Eur J Appl Physiol 108: 1045–1053, 2010.
24. Maughan RJ, Greenhaff PL, Leiper JB, Ball D, Lambert CP, Gleeson M. Diet composition and the performance of high-intensity exercise. J Sports Sci 15: 265–275, 1997.
25. Micklewright D, Papadopoulou E, Swart J, Noakes TD. Previous experience influences pacing during 20 km time trial cycling. Br J Sports Med 44: 952–960, 2010.
26. Orendurff MS, Walker JD, Jovanovic M, Tulchin KL, Levy M, Hoffman DK. Intensity and duration of intermittent exercise and recovery during a soccer match. J Strength Cond Res 24: 2683–2692, 2010.
27. Swart J, Lindsay TR, Lambert MI, Brown JC, Noakes TD. Perceptual cues in the regulation of exercise performance—Physical sensations of exercise and awareness of effort interact as separate cues. Br J Sports Med 46: 42–48, 2012.
28. Taylor DJ, Kemp GJ, Thompson CH, Radda GK. Ageing: Effects on oxidative function of skeletal muscle in vivo. Mol Cell Biochem 174: 321–324, 1997.
29. Tucker R, Marle T, Lambert EV, Noakes TD. The rate of heat storage mediates an anticipatory reduction in exercise intensity during cycling at a fixed rating of perceived exertion. J Physiol 574: 905–915, 2006.
30. Watt KKO, Hopkins WG, Snow RJ. Reliability of performance in repeated sprint cycling tests. J Sci Med Sport 5: 354–361, 2002.