Regulation of Pacing Strategies during Successive 4-km Time Trials : Medicine & Science in Sports & Exercise

Journal Logo

Applied Sciences: Physical Fitness and Performance

Regulation of Pacing Strategies during Successive 4-km Time Trials


Author Information
Medicine & Science in Sports & Exercise 36(10):p 1819-1825, October 2004. | DOI: 10.1249/01.MSS.0000142409.70181.9D
  • Free


The capacity to produce and maintain power output decreases rapidly during bouts of high-intensity exercise (1,35). Therefore, in order to sustain a high power output in events lasting longer than several seconds, athletes adopt a pacing strategy to delay fatigue and optimize performance (15), where pacing strategy is defined as the conscious or subconscious regulation of work output according to a predetermined plan in order to maximize performance without causing irreparable harm to physiological systems. The typical pacing strategy comprises several stages: a short initial period of high power output followed by a sharp reduction to a power output, which is usually maintained until the final period of the exercise bout when power output may again increase (8).

It is usually believed that these changes in power output are determined by the development of a critically low muscle pH that impairs skeletal muscle contractile function (20,21), so-called “peripheral fatigue.” However, this peripheral model of fatigue does not satisfactorily explain all the known features of exercise fatigue (17,23,24), most notably the phenomenon of the “lactate paradox” of high altitude (12,31).

Foster et al. (6) have proposed an alternate theory in which the pacing strategy is under central neural control. Indeed, this theory that the central nervous system regulates athletic performance was first proposed over a century ago (34). But the complexity of quantifying any central neural regulation of exercise performance has influenced researchers to focus on the peripheral, skeletal muscular determinants of fatigue and exercise performance with, until recently, less attention on any central component (7,17,24,29).

Interval training, in which athletes work repeatedly at near maximum effort with short periods of recovery between exercise bouts, represents one of the most intense forms of physical activity (30). Because the activity is repetitive, athletes must adopt a pacing strategy within the first interval to ensure that they can produce repetitive intervals with similar total work output. To determine any contribution of the central nervous system to the regulation of this pacing strategy, we measured the work output and associated changes in physiological, integrated electromyographic (iEMG), and metabolic parameters in consecutive 4-km cycling intervals performed by highly trained cyclists in the laboratory. We theorized that central regulation would be characterized by iEMG activity that dynamically tracks the changes in power output, particularly towards the end of each time trial. In contrast, peripheral causes of fatigue would produce changes in power output that mirror changes in blood lactate concentrations, and in which power output would steadily decline toward the latter stages of the exercise bout despite a progressive increase in iEMG activity. In addition, a high proportion of motor units activated during the maximal voluntary contraction would be recruited during each interval.



Seven highly trained, competitive male cyclists were selected to participate in this study (Table 1). These subjects were chosen because they were well-trained, accustomed to high-intensity exercise, and familiar with laboratory testing because they frequently participated in previous trials in this laboratory. At the time of investigation, they were cycling 400–800 km·wk−1. The study was conducted with the approval of the Ethics in Human Research Committee of the Faculty of Health Sciences at the University of Cape Town; before the trial, all subjects were informed of the nature of the investigation, after which they gave written informed consent.

Subject characteristics and results from the maximal incremental exercise test (N = 7).

Kingcycle ergometry system.

Each subject completed two trials in the laboratory. On the first occasion, maximal power output and peak oxygen consumption (V̇O2peak) were measured. On the second occasion seven days later, subjects performed three consecutive, 4-km time trials (TT) with a 17-min rest between each. All tests were conducted on a Kingcycle ergometry system (Kingcycle Ltd., High Wycombe, UK) that allowed each cyclist to ride his own bicycle in the laboratory. Previous studies have shown that performance tests conducted on these ergometers have good reliability when compared to standard laboratory ergometers (18,25,26).

V̇O2peak test.

On their first visit to the laboratory, cyclists completed the maximal incremental exercise test (MIE) to exhaustion during which oxygen consumption (V̇O2), power output, and heart rate (HR) were recorded for the duration of the test. After a 10- to 15-min warm-up at a self-selected intensity, the test commenced at a workload of 200 W. Thereafter, the workload increased by 20 W·min−1 until the subject could no longer maintain the required power output (36). The same investigators gave verbal encouragement throughout all maximal tests. Peak power output (PPO) was defined as the highest power output, averaged over 10 s, attained during the V̇O2peak test. Subjects were requested to remain in a seated position for the duration of the MIE and the time trials.

For the measurement of V̇O2 during the MIE and the 4-km time trials, subjects wore a mask covering the nose and mouth; the expired air passed through an on-line computer system attached to an automated gas analyzer (Oxycon model Alpha, Mijnhardt, The Netherlands). Before each test, the gas analyzer was calibrated with a Hans Rudolph 5530, 3-L syringe and a span gas mixture (15.24% O2/5.11% CO2). Analyzer outputs were processed by a computer that calculated V̇O2 and carbon dioxide production using conventional equations (16). V̇O2peak was defined as the highest V̇O2 measured over 10 s during the MIE.

HR was recorded every 10 s using a Polar Accurex Plus heart rate monitor (Polar Electro, Kempele, Finland). The data were downloaded via an interface into HRM files after completion of the trial. Maximum HR was taken as the highest HR recorded at any stage during the test.

Time trials.

Within 7 d of the MIE, each subject returned to the laboratory at the same time of day to perform three consecutive 4-km time trials. Successive time trials were separated by approximately 17 min, which comprised 10-min inactive recovery, 5-min warm-up cycling, at a self-selected intensity, and a 2-min countdown. Subjects were requested to perform the same type of training for the duration of the trial and to refrain from heavy physical exercise for 24 h before the TT testing.

After a 10-min warm-up at a self-selected intensity, the first TT started after a 2-min countdown during which the cyclists maintained a cycling speed of 35–40 km·h−1. This speed had to remain constant until the start of the time trial. After each time trial, the subjects engaged in inactive rest for 10 min after which they performed a 5-min warm-up before commencing the 2-min countdown. Subjects were told to perform each time trial in the fastest time possible but in the knowledge that three trials were to be performed. The elapsed distance was the only feedback given to the subjects during the time trials. They were not informed of completion times until after the final time trial. Power output, HR, and V̇O2 measurements were recorded continuously during the trials. A fan was positioned in front of the subjects during their time trials, and during the 10-min rest interval, subjects were allowed to drink water ad libitum.

Before the MIE and time trials a 20-gauge Jelco cannula (Critikon, Halfway House, South Africa) was inserted into an antecubital forearm vein for blood sampling. Exactly 3 min (19) after both the V̇O2peak test and each TT, blood was drawn and placed into tubes containing potassium oxalate and sodium fluoride for the measurement of plasma lactate concentration. The blood samples were kept on ice until centrifuged at 3000 × g for 10 min at 4°C and the plasma stored at −20°C until later analysis. Plasma lactate concentrations were determined by spectrophotometric (Beckman Spectrophotometer–M35) enzymatic assays using the conventional assay (Lactate PAP, bioMérieux Kit, Marcey l’Etoile, France).

Isometric testing of skeletal muscle function.

Immediately before the MIE test and time trials, each subject’s peak isokinetic force was measured on a Kin-Com isokinetic dynamometer (Chattanooga Group Inc., U.S.). The peak force produced by the quadriceps muscles was tested isometrically at an angle of 60°, with full knee extension being the 0° reference. Subjects performed four 50%, two 70%, one 90%, and one 100% familiarization trials of 5 s each as a warm-up. After this warm-up, each subject performed four 5-s maximal voluntary contractions (MVC) with a 5-s rest between trials. The EMG activity coinciding with the peak torque of the second effort was used to normalize the EMG values recorded during the time trials (14). The subjects were verbally encouraged during the tests to exert a maximal effort.

Electromyographic activity.

Muscle recruitment was assessed during the isometric test, as well as during the time trials by measuring EMG activity of the rectus femoris muscle. Electrodes (Thought Technology TriodeTM MIEPO1-00, Montreal, Canada) with a bandwidth of 20–500 Hz and sensitivity of < 0.08 μV were attached to the subject’s lower limb before the start of all testing. The skin overlying the rectus femoris muscle was first shaven, after which the outer layer of epidermal cells was abraded with sandpaper and the dirt removed from the skin using an alcohol swab. The triode electrode was placed in the middle of the subject’s rectus femoris muscle, secured with self-adherent wrap (Coban 3M, 1582, St. Paul, MN) and linked via a fiber-optic cable to a Flexcomp/DSP EMG apparatus (Thought Technology, Montreal, Canada) and host computer (13). EMG activity was recorded for 5 s during the isokinetic test. During the time trials, EMG activity was recorded from 60 to 80 s to coincide with the maximal power output and in 10 s intervals between 200 s to exhaustion during the TT.

A toggle switch was activated at the beginning of each test to mark the start point of the test procedure, with each activity being sampled at a 1984-Hz capture rate. A 50-Hz line filter was applied to the raw EMG data to prevent any external interference from electrical sources. The EMG signals from the electrode were band-pass filtered (20–500 Hz) and amplified using standard differential amplifiers (Thought Technology; common mode rejection ratio > 130 dB at 1 kHz, input impedance = 106 MΩ, adjustable gain up to 1600). The raw EMG signals were subsequently full wave rectified, movement artifact removed using a high-pass second order Butterworth filter with a cut-off frequency of 15 Hz and then smoothed with a low-pass second-order Butterworth filter with a cut-off frequency of 5 Hz. This was performed using MATLABTM gait analysis software. The integrated EMG data (iEMG) were used for subsequent analyses.

EMG data analysis.

All EMG data were normalized by dividing the value at each time point during the cycle trial by the EMG value obtained during the MVC performed before the start of the performance test. Further normalization ofthe EMG recorded over the final 90 s of exercise was performed against the mean EMG recorded over the period 60 to 80 s. These data are expressed as a percentage of the mean EMG during the peak power output.

Statistical analysis.

Data for HR, V̇O2, and power output were averaged over 30 s from 0 to 300 s. EMG data were compared every 10 s from 200 s to fatigue. An analysis of variance was used to assess differences between and within the trials. Once main effects were identified individual differences between the means were located using Tukey’s HSD post hoc procedure. Significance was accepted at P < 0.05. All data are expressed as mean ± standard deviation (SD).


Maximum values of V̇O2, HR, and plasma lactate concentrations.

The V̇O2peak attained during maximal incremental exercise (MIE) was similar to the highest V̇O2 measured during each of the 4-km time trials (MIE 71.4 ± 2.3; TT1 70.2 ± 4.1; TT2 70.7 ± 5.3; TT3 69.8 ± 4.2 mL·kg−1·min−1) (Table 2). Maximum HR reached during MIE (193 ± 7 beats·min−1) was significantly higher than during TT1 (186 ± 6 beats·min−1) and TT2 (187 ± 6 beats·min−1) but was not different from TT3 (189 ± 7 beats·min−1). Postexercise plasma lactate concentrations were similar for the MIE and all the time trials (MIE 14.5 ± 0.8, TT1 12.7 ± 2.0; TT2 12.8 ± 2.5; TT3 14.0 ± 0.4 mmol·L−1) (Table 2 and Fig. 1d), and the concentrations were not different from those observed after the incremental maximal exercise test (14.5 ± 0.8 mmol·L−1).

Mean values of V̇O2peak, maximal heart rate, and maximal lactate concentrations attained during the maximal incremental exercise and three consecutive 4-km time trial intervals.
FIGURE 1— Peak power output, mean power output, performance time and peak lactate concentrations for the three consecutive 4-km time trial intervals; γ first interval is significantly different from the second and third intervals (:
P < 0.05); * second interval is significantly different from the first interval ( P < 0.05).

Peak and average power outputs.

The peak power outputs, reached at 60 s (Fig. 2a), were significantly higher in the first interval (TT1 556 ± 51 W; TT2 504 ± 44 W; TT3 506 ± 41 W; P < 0.05) (Fig. 1a). However, the average power outputs for each interval were not different (TT1 447 ± 30 W; TT2 425 ± 33 W; TT3 434 ± 33 W) (Fig. 1b). Nor was the time taken to complete the first interval (284 ± 8 s) and the average velocity (51 ± 1 km·h−1) significantly faster than for the third interval third interval (287 ± 9 s and 50 ± 2 km·h−1) although it was faster than the second interval (290 ± 9 s and 50 ± 2 km·h−1) (Fig. 1c). Completion times were not different between the second and the third interval (Fig. 1c). In all trials, power output declined significantly from 60 s (Fig. 2a) after which it reached a plateau. Beyond 240 s, power output rose until completion of the time trial (Fig. 2a and 3a–c).

FIGURE 2— Power output, V̇O2, and HR profiles for the duration of the consecutive 4-km time-trial intervals.
FIGURE 3— Relationship between power output and iEMG over the final 90 s of each of the three consecutive 4-km time trials relative to the maximum values at 60 s.

V̇O2 and HR.

The rate of oxygen consumption was similar for all intervals after the first minute (Fig. 2b) and did not increase significantly over the duration of the interval. Likewise, HR did not vary between intervals, but it did increase significantly from the 30th to the 60th second (P < 0.05), after which it did not increase significantly (Fig. 2c).


The iEMG recorded during the period of maximal power output during the time trials was less than 25% of iEMG activity during the MVC (TT1 24.2 ± 3.5; TT1 21.1 ± 4.2%; TT3 23.7 ± 3.8%). There was no difference in iEMG between intervals. However, in all trials iEMG dynamically tracked changes in power output (Fig. 3a–c) with the highest values being measured at 60 s. Furthermore, iEMG did not increase sequentially across successive intervals and was not higher at 300 s than at 60 s in any interval.


The first important finding of this study was that the first time trial and the last time trial were completed in a similar time (Fig. 1c), even though the subjects received no external feedback regarding their cycling speeds, power output, elapsed time, or HR. Distance covered was the only external feedback provided. However, the peak power output was significantly higher in the first time trial than the last time trial (Fig. 2a), which indicates that the subjects adopted a modified pacing strategy in the second and third time trials, which allies allocation of effort with the remaining duration of the work bout and comprises both feed-forward calculations; feedback regulation from afferent changes associated with peripheral metabolic structures; and the external environment and also assimilates knowledge acquired from prior exercise bouts. The pacing strategy during the three consecutive 4-km time trials was characterized by an output that peaked at 60 s and then declined steeply before leveling out, between 120 and 240 s, at a similar magnitude to the average power output over the first 30 s. Thereafter, power output progressively increased and, in the case of the second and third time trials, approached the peak magnitude recorded at 60 s (Fig. 3). This pattern of fluctuation in power output indicates that the subjects did not reach a state of absolute fatigue during any of the time trials because they were able to increase power output towards the end of the trials. Whereas, if they had been approaching absolute fatigue, they would not have been capable of increasing their work rate, and, in fact, a continuous, progressive decline in power output throughout the time trial would be expected.

There is indirect evidence of a learning effect between the time trials in the form of a modified power output in the initial stages (up to 90 s) of the second and third time trials compared with the first time trial. Although it could be argued that this is, in fact, a manifestation of fatigue, and therefore a “fatigue effect,” as a result of incomplete recovery of the phosphocreatine (PCr) stores within the muscle. There are two reasons why this argument is not valid in these circumstances. First, Harris et al. (11) have shown that PCr stores are completely resynthesized within 2–3 min of stopping exercise, and more recently Bogdanis et al. (3) were able to demonstrate an 80% recovery of intramuscular PCr within 4 min of recovery; therefore, the 10 min of recovery provided to the subjects offered sufficient time for recovery of PCr stores. Second, the completion times for the first and third time trial were similar, which is unlikely to have occurred if there had been progressive depletion of intramuscular creatine over the duration of the trial. Furthermore, despite this significant modification in peak power output, the mean power output was similar for all three time trials (Fig. 2b). It is possible that there are mechanisms that affect peak power more than mean power, but it is more likely that the changes are due to pacing strategy and therefore this suggests that subjects may have consciously or, more likely, subconsciously altered their pacing strategies in the second and third intervals, based upon their experience in the first trial in order to either minimize fatigue or optimize performance, or perhaps a combination of both. Importantly, there was no evidence for the development of accumulated, progressive fatigue because by its very nature progressive, absolute fatigue of a peripheral mechanism must manifest as a decreasing power output power output over the duration of the trial, whereas power output actually increased over the last 60 s of each time trial (Fig. 2a).

The second important finding of the study was that the power output increased during the final 60 s of each time trial, a phenomenon that has also been noted in a longer duration time trials lasting 60 min which included six 1-min sprint intervals (17). These authors also found an increase in iEMG activity and power output to near initial values during the final sprint, after progressive declines in the intervening sprints. If the power output during the time trials had been regulated by peripheral fatigue mechanisms, an increased central neural recruitment of additional motor units would be expected to occur (5,9,32) in an attempt to compensate for the reduced power output from the fatiguing motor units. This would be shown as a progressive increase in iEMG activity and an irreversibly falling power output. However, in this study the iEMG changes tracked the power output changes in a nonmonotonic fashion (i.e., up and down), indicating dynamic regulatory activity and the increase in power output toward the end of the time trials was dynamically tracked by the iEMG. One might suggest that a lack of difference in difficult to interpret, but based on findings by Kay et al. (17), and with the nonmonotonic changes in power output in this study, we believe that iEMG is tracking power output. More significantly, iEMG were not progressively higher in successive time trials as would be expected if peripheral fatigue caused a progressive impairment of the power output of individual motor units, requiring the recruitment of progressively more motor units to compensate for the reduced power output of the fatiguing units.

Less than a quarter of the available motor units in the studied muscles were recruited during the time trials, a finding that is consistent with previous observations (17,28,29) but is incompatible with the peripheral fatigue model. The peripheral model predicts that there must be near total motor unit recruitment at exhaustion so that fatigue results from peripheral metabolite-induced modulation of the contractile activity of all the recruited motor units. However, the iEMG findings need to be evaluated with caution, because of the complex nature of the surface signal and the number of factors which influence it. Similarly, the level of recruitment as compared with the MVC may be affected by the dynamic nature of the cycling trial. Nevertheless, previous work has shown this method of normalization and iEMG analysis to be acceptable during dynamic activity (4,14).

The initial rapid increase in power output during the first time trial, which is characterized by a disproportionately large percentage of the total work performed in the initial stages of exercise, is indicative of the lag phase inherent in the feedback model (2). However, in the final time trial, the initial spike in power output is significantly dampened even though, importantly, time to completion of the time trial was similar to the first interval. This observation is more consistent with the peripheral monitoring theory proposed by Foster et al. (6) in which a component of the pacing strategy is to complete exercise with specific terminal plasma (and muscle) lactate concentrations. They suggest that athletes learn to “sense” the intramuscular pH and adapt their workload on the basis of this feedback so that a critically low muscle pH is never approached. The proposed mechanistic functioning of this feedback system involves the detection of pH changes by nociceptors or chemoreceptors that transmit an inhibitory signal to the central nervous system, which is perceived as a “sensation” (10). As a result, the motor cortex decreases efferent neural drive directed to those motor units in the exercising muscles from which the afferent signals were received. The degree of reduction in efferent activity is proportional to the magnitude of afferent feedback activity. This interpretation is further supported by the unexpected finding that the plasma lactate concentrations were similar at the end of each time trial (Fig. 2d). Although muscle pH and lactate may act as the monitored chemicals, neither would appear to be limiting factors according to those studies that have dissociated blood lactate concentrations and pH from changes in performance (27) by altering blood lactate and pH during stochastic 40-km time trials without equivalent changes in performance. Similarly, Medbø and Sejersted (19) concluded that blood lactate concentrations do not limit physical performance during exhaustive exercise. In fact, lactic acid has been observed to have a protective effect on force generation in skeletal muscle (22).

An alternative explanation, which also fits the scenario of a reduced initial power output but similar completion times, proposes a feed-forward mechanism whereby a subconscious “controller” determines the overall pacing strategy during exercise by matching the rate of energy expenditure and the current energy reserves with the predicted energy cost of the exercise (33), but within the physiological capacity of the individual. The pacing strategy would therefore result from changes in the number of motor units that are alternatively recruited and de-recruited during different stages of exercise. Ulmer (33) was the first to propose that the metabolic rate during exercise is maintained by the action of a programmer that takes cognizance of the duration or distance, or both, of the planned effort and calculates the pacing strategy on the basis of previous experience. This preemptive mechanism would not only pattern a single time trial but would be responsible for the pacing strategy adopted in the consecutive trials that comprise the entire exercise bout.

In summary, this study shows that the pacing strategies adopted by athletes during three successive 4-km cycling time trials that each elicited a maximal V̇O2 response were remarkably similar, despite the absence of external feedback to the subjects other than distance covered. Importantly, power outputs rose progressively from 240 to 300 s of exercise in all time trials, and the iEMG changes tracked the increases and decreases in power output in a dynamic manner. On the basis that the trend in the relationship between iEMG and power output shows a general linearity, although there is variability within this relationship and the finding that the performances were remarkable similar, that iEMG activities were not progressively greater in successive intervals and that only about 25% of available motor units were recruited at the point of peak power output, suggests the presence of a centrally determined, regulation of the adopted pacing strategies. Such regulation would result from a centrally determined recruitment and de-recruitment of more or less motor units during the exercise bout. The extent to which the pacing strategy is influenced by feedback from metabolic events in the active muscles or is dependent on preemptive feed-forward could not be determined by this study.

Funding for this study was provided by the Beatrix Waddell Scholarship Fund, the Lowenstein Scholarship Trust, and the Harry Crossley Staff Research Fund, all of the University of Cape Town; the National Research Foundation and the Medical Research Council of South Africa; and Discovery Health Pty Ltd.


1. Bar-Or, O. The Wingate anaerobic test: an update on methodology, reliability and validity. Sports Med. 4:381–394, 1987.
2. Bhalla, U. S., and R. Iyengar. Robustness of the bistable behavior of a biological signaling feedback loop. Chaos 11:221–226, 2001.
3. Bogdanis, G. C., M. E. Nevill, L. H. Boobis, and H. K. A. Lakomy. Contribution of phosphocreatine and aerobic metabolism to energy supply during repeated sprint exercise. J. Appl. Physiol. 80:876–884, 1996.
4. Burden, A. M., M. Trew, and V. Baltzopoulos. Normalisation of gait EMGs: a re-examination. J. Electromyogr. Kinesiol. 13:519–532, 2003.
5. Crenshaw, A. G., S. Karlsson, B. Gerdle, and J. Friden. Differential responses in intramuscular pressure and EMG fatigue indicators during low- vs. high-level isometric contractions to fatigue. Acta Physiol. Scand. 160:353–361, 1997.
6. Foster, C., M. Schrager, A. C. Snyder, and N. N. Thompson. Pacing strategy and athletic performance. Sports Med. 17:77–85, 1994.
7. Gandevia, S. C. Spinal and supraspinal factors in human muscle fatigue. Physiol. Rev. 81:1725–1789, 2001.
8. Hagerman, F. C. Applied physiology of rowing. Sports Med. 1:303–326, 1984.
9. Hakkinen, K., and P. V. Komi. Electromyographic and mechanical characteristics of human skeletal muscle during fatigue under voluntary and reflex conditions. Electroencephalogr. Clin. Neurophysiol. 55:436–444, 1983.
10. Hampson, D. B., A. St Clair Gibson, M. I. Lambert, and T. D. Noakes. The influence of sensory cues on the perception of exertion during exercise and central regulation of exercise performance. Sports Med. 31:935–952, 2001.
11. Harris, R. C., R. H. Edwards, E. Hultman, L. O. Nordesjo, B. Nylind, and K. Sahlin. The time course of phosphorylcreatine resynthesis during recovery of the quadriceps muscle in man. Pflugers Arch. 367:137–142, 1976.
12. Hochachka, P. W., C. L. Beatty, Y. Burelle, M. E. Trump, D. C. McKenzie, and G. O. Matheson. The lactate paradox in human high-altitude physiological performance. News Physiol. Sci. 17:122–126, 2002.
13. Hunter, A. M., A. St Clair Gibson, M. I. Lambert, L. Nobbs, and T. D. Noakes. Effects of supramaximal exercise on the electromyographic signal. Br. J. Sports Med. 37:296–299, 2003.
14. Hunter, A. M., A. St Clair Gibson, M. I. Lambert, and T. D. Noakes. Electromyographic normalization method for cycle fatigue protocols. Med. Sci. Sports Exerc. 34:857–861, 2002.
15. Ingen Schenau, G. J., J. J. de Koning, and G. de Groot. Optimisation of sprinting performance in running, cycling and speed skating. Sports Med. 17:259–275, 1994.
16. Jones, N. L. Clinical Exercise Testing. Philadelphia: W. B. Saunders, 1987, pp. 292–300.
17. Kay, D., F. E. Marino, J. Cannon, A. St Clair Gibson, M. I. Lambert, and T. D. Noakes. Evidence for neuromuscular fatigue during high-intensity cycling in warm, humid conditions. Eur. J. Appl. Physiol. 84:115–121, 2001.
18. Lindsay, F. H., J. A. Hawley, K. H. Myburgh, H. H. Schomer, T. D. Noakes, and S. C. Dennis. Improved athletic performance in highly trained cyclists after interval training. Med. Sci. Sports Exerc. 28:1427–1434, 1996.
19. Medbø, J. I., and O. M. Sejersted. Acid-base and electrolyte balance after exhausting exercise in endurance-trained and sprint-trained subjects. Acta Physiol. Scand. 125:97–109, 1985.
20. Metzger, J. M., and R. H. Fitts. Role of intracellular pH in muscle fatigue. J. Appl. Physiol. 62:1392–1397, 1987.
21. Nakamaru, Y., and A. Schwartz. The influence of hydrogen ion concentration on calcium binding and release by skeletal muscle sarcoplasmic reticulum. J. Gen. Physiol. 59:22–32, 1972.
22. Nielsen, O. B., F. de Paoli, and K. Overgaard. Protective effects of lactic acid on force production in rat skeletal muscle. J. Physiol. 536:161–166, 2001.
23. Noakes, T. D. 1996 J. B. Wolffe Memorial Lecture. Challenging beliefs: ex Africa semper aliquid novi. Med. Sci. Sports Exerc. 29:571–590, 1997.
24. Noakes, T. D. Physiological models to understand exercise fatigue and the adaptations that predict or enhance athletic performance. Scand. J. Med. Sci. Sports. 10:123–145, 2000.
25. Palmer, G. S., S. C. Dennis, T. D. Noakes, and J. A. Hawley. Assessment of the reproducibility of performance testing on an air- braked cycle ergometer. Int. J. Sports Med. 17:293–298, 1996.
26. Schabort, E. J., J. A. Hawley, W. G. Hopkins, I. Mujika, and T. D. Noakes. A new reliable laboratory test of endurance performance for road cyclists. Med. Sci. Sports Exerc. 30:1744–1750, 1998.
27. Schabort, E. J., S. C. Killian, A. St Clair Gibson, J. A. Hawley, and T. D. Noakes. Prediction of triathlon race time from laboratory testing in national triathletes. Med. Sci. Sports Exerc. 32:844–849, 2000.
28. Sjogaard, G. Force-velocity curve for bicycle work. In: Biomechanics VI, E. Asmussen and K. Jorgensen (Eds.). Baltimore: University Park Press, 1978, pp. 33–39.
29. St Clair Gibson, A., E. J. Schabort, and T. D. Noakes. Reduced neuromuscular activity and force generation during prolonged cycling. Am. J. Physiol. 281:R187–R196, 2001.
30. Stepto, N. K., D. T. Martin, K. E. Fallon, and J. A. Hawley. Metabolic demands of intense aerobic interval training in competitive cyclists. Med. Sci. Sports Exerc. 33:303–310, 2001.
31. Sutton, J. R., J. T. Reeves, B. M. Groves, et al. Oxygen transport and cardiovascular function at extreme altitude: lessons from Operation Everest II. Int. J. Sports Med. 1(Suppl. 13):S13–S18, 1992.
32. Taylor, A. D., R. Bronks, P. Smith, and B. Humphries. Myoelectric evidence of peripheral muscle fatigue during exercise in severe hypoxia: some references to m. vastus lateralis myosin heavy chain composition. Eur. J. Appl. Physiol. 75:151–159, 1997.
33. Ulmer, H. V. Concept of an extracellular regulation of muscular metabolic rate during heavy exercise in humans by psychophysiological feedback. Experentia 52:416–420, 1996.
34. Waller, A. D. The sense of effort: an objective study. Brain 14:179–249, 1896.
35. Withers, R. T., W. M. Sherman, D. G. Clark, et al. Muscle metabolism during 30, 60 and 90 s of maximal cycling on an air- braked ergometer [published errata appear in Eur. J. Appl. Physiol. 1992;64(4):387 and 1992;64(5):485]. Eur. J. Appl. Physiol. 63:354–362, 1991.
36. Zhang, Y. Y., M. C. Johnson, N. Chow, and K. Wasserman. Effect of exercise testing protocol on parameters of aerobic function. Med. Sci. Sports Exerc. 23:625–630, 1991.


©2004The American College of Sports Medicine