Influence of Feedback and Prior Experience on Pacing during a 4-km Cycle Time Trial


Medicine & Science in Sports & Exercise: February 2009 - Volume 41 - Issue 2 - pp 451-458
doi: 10.1249/MSS.0b013e3181854957
Applied Sciences

Purpose: To determine the importance of distance knowledge, distance feedback, and prior experience on the setting of a pacing strategy.

Methods: Eighteen well-trained male cyclists were randomly assigned to a control (CON) group or an experimental (EXP) group and performed four consecutive 4-km time trials (TT), separated by a 17-min recovery. The CON group received prior knowledge of distance to be cycled and received distance feedback throughout each TT; the EXP group received neither but knew that each TT was of the same distance.

Results: The EXP group was significantly slower than the CON group to complete TT1 (367.4 ± 21 vs 409.4 ± 45.5 s, P < 0.001). Differences between groups in completion time reduced over successive TT (CON TT4 = 373.9 ± 20 s vs EXP TT4 = 373.8 ± 14.4 s), shown by a significant linear contrast (F1,16 = 12.39, P < 0.0005). Mean speed and power output also showed significantly reduced differences between groups over successive TT (P < 0.0005). However, peak power output showed no such convergence between groups over TT. End blood lactate was significantly different between groups in TT1, but differences between groups converged with successive TT.

Conclusion: The progressively improving completion times in the EXP group show that distance feedback is not essential in developing an appropriate pacing strategy. Prior experience of an unknown distance appears to allow the creation of an internal, relative distance that is used to establish a pacing strategy.

School of Sport and Health Sciences, University of Exeter, Devon, UNITED KINGDOM

Address for correspondence: Craig Anthony Williams, Ph.D., School of Sport and Health Sciences, University of Exeter, St. Luke's Campus, Heavitree Road, Exeter, EX1 2LU, United Kingdom; E-mail:

Submitted for publication April 2008.

Accepted for publication June 2008.

Article Outline

Where the time to completion is the measure of success in an event lasting longer than 60 s, a pacing strategy exerts a large influence over success or failure (10). A pacing strategy has been defined as the variation of speed over the race by regulating the rate of energy expenditure (14). Appropriate regulation of energy expenditure should reduce the rate of fatigue development and thus enhance performance.

Two main theories exist that describe the origins of fatigue. The most popular theory is termed peripheral fatigue and can be defined as fatigue produced by changes at or distal to the neuromuscular junction (11). This theory revolves around causative factors such as metabolic inhibition of the contractile process and excitation-contraction postulated failure via accumulation of intramuscular metabolites, although the particular metabolite(s) that has the most important role is disputed (4,6,19,34). The other postulated origin of fatigue is from a supraspinal or spinal level, otherwise termed central fatigue (11). Central fatigue can be defined as a reduction in neural drive to the muscles, resulting in a decline in force production or tension development that is independent of changes in skeletal muscle contractility (9).

One explanation for how peripheral contractility is controlled centrally has been proposed by Ulmer (32) in the theory of teleoanticipation. Before an exercise bout, the body "anticipates" the optimal arrangement of exertion so as to avoid early exhaustion before reaching a finish point. Thus, the importance of the knowledge of exercise end point is stressed. Ulmer theorized that there must be a "central programmer" that focuses on the finishing point of the task and works backward from that point when regulating optimal metabolic rate and motor output. More recently, a new model of central fatigue has been proposed, termed the central governor model (CGM), which is based on teleoanticipation. It is hypothesized that physical activity is controlled by a central governor in the brain, which interprets fatigue as a sensation, based on afferent feedback from the periphery. The governor then adjusts skeletal muscle motor unit recruitment, which is manifested in a continuously refining pacing strategy, with the primary aim of avoiding physiological "catastrophe" (24). If the aim of a pacing strategy is to regulate energy expenditure, then this should be under the control of the theorized central governor and thus be an important variable to monitor. CGM supporters stress the importance of recognizing fatigue as a sensation, and so RPE is a tool that has been frequently used in research to investigate the role of perceived exertion on the setting of a pacing strategy (1,33).

The majority of evidence that is used to support teleoanticipation, and the CGM comes from exercise in environments such as heat (21,25,31) and at altitude (7,23,24,26). Yet in these examples, ecological validity has been limited due to power output (PO) being either fully or partly dictated by the protocol. Where exercise has been self-paced, similar results have been found in several studies (2,3,22,27,36). However, the data from these studies are difficult to reconcile due to methodological differences.

The measurement of EMG changes in exercised muscle is widely used in research investigating central fatigue. This is based on the assumption that these changes reflect variations in central neural drive (30). According to peripheral theories, integrated EMG (iEMG) activity should increase progressively to compensate for impaired contractile function caused by metabolic changes. Central theories, however, state that both force and iEMG activity will fall in parallel, indicating that the recruitment of fewer motor neurones has caused the work output of the muscles to fall (3,14,30).

Although the importance of distance knowledge, feedback, and prior experience is heavily stressed in the theories of the CGM and teleoanticipation (24,32), there is limited research that focuses on these factors and their role in the setting of a pacing strategy. To provide stronger scientific grounding for these theories of central fatigue, empirical research should concentrate on explaining these principal components of the model. Athletes in competitive situations are not constrained by fixed cadences, PO, or intensities based on V˙O2, and this should therefore be taken into account in studies investigating this phenomenon.

The aim of the current study was, therefore, to establish whether an athlete can adopt a successful pacing strategy that is based on prior experience alone. A group of well-trained cyclists were randomly assigned to two groups (experimental [EXP] and control [CON]) and were given different levels of distance feedback over a series of time trials (TT). It was hypothesized that distance-blinded participants in the EXP group with prior experience of the TT would be able to develop a pacing strategy that was as successful as the CON group who received prior distance knowledge and distance feedback.

Back to Top | Article Outline


Back to Top | Article Outline


Eighteen well-trained (V˙O2peak = 61 ± 5 mL·kg−1·min−1, max PO = 397 ± 51), competitive male cyclists, aged 28 ± 10 yr, were recruited to participate in this study. These participants were selected because they participated in regular, structured training (11.5 ± 3.5 h·wk−1) and competed frequently (approximately one competition per week). The study was conducted with approval of the Institutional Ethics Committee, and participants read and signed a form of written informed consent. Before testing, participants were randomly assigned to two groups: CON and EXP. The subjects were told that the study was about the reliability of repeated TT of the same distance on a new cycle ergometer. However, on completion of the study, all participants were debriefed about the deception and the completed protocol.

Back to Top | Article Outline

CompuTrainer cycle ergometry system.

Before testing, all participants were given a familiarization period in the laboratory. All tests were conducted on a CompuTrainer cycle ergometry system (RacerMate®; CompuTrainer, Seattle, WA), which allowed each cyclist to ride his own bicycle in the laboratory. Previous research has shown these ergometers to provide a reliable measure of PO when compared with standard laboratory ergometers (8). Further calibration and reliability testing in house demonstrated that the CompuTrainer provided an accurate and reliable measure of PO (at PO above 150 W, measured PO on the CompuTrainer was always within 6% of actual PO).

Back to Top | Article Outline

V˙O2peak test.

On the second visit to the laboratory, cyclists completed a maximal incremental exercise (MIE) test to exhaustion during which oxygen consumption (Cortex Metalyser 11R; Cortex GmbH, Lepzig, Germany), PO, RPE, and HR (Polar Heart Rate Chest Strap T31, New York) were recorded for the duration of the test. After a 10-min warm-up at a self-selected intensity, the test commenced at a PO of 150 W. Thereafter, the PO increased by 20 W·min−1 until the subject could no longer maintain the required PO (34). RPE (Borg's 6-20 scale) was recorded every minute (5). The researcher gave verbal encouragement throughout all maximal tests. Peak PO (POmax) was defined as the highest PO, averaged over 10 s, attained during the V˙O2peak test.

Back to Top | Article Outline

Time trials.

Within 7 d of the MIE, each participant returned to the laboratory at the same time of day (±2 h) to perform four consecutive 4-km TT. Successive TT were eparated by approximately 17 min, which included a 10-min recovery period where the cyclist could do as they liked (this involved either spinning or a brief period off the bike to stretch), a 5-min cycling warm-up at a self-selected intensity, and a 2-min countdown during which participants cycled at a speed of 25-35 km·h−1. The participants in the CON group were told that they would complete 4 × 4-km TT. The CON group was given distance feedback throughout each TT in the form of elapsed distance shown on a computer screen. The EXP group was told that they would complete four TT of the same distance but they were not told the distance to be completed. The EXP group did not receive any distance feedback during the TT. Participants were told that each TT would involve a bout of maximal aerobic exercise. For both groups, all time cues were removed from sight. Both groups were told to perform each TT in the fastest time possible but in the knowledge that four TT were to be performed. Neither group was informed of their completion times until after the final TT. A fan was positioned in front of the subjects throughout testing, and participants were allowed to drink water ad libitum during the 10-min rest period between each TT. The CompuTrainer ergometry system continuously recorded PO, speed, and time. V˙O2 and HR were recorded continuously as previously described. At the start of each 2-min countdown, immediately after each TT and 5 min after the end of each TT, a finger tip blood sample was taken using an automated lancet (Hemocue, Angelholm, Sweden) and then immediately analyzed for B[La] with a YSI 2300 Stat Plus Analyzer (Yellow Springs Instruments, Ohio). At the start of each TT, the end of each TT, and for each kilometer elapsed, participants in the CON group were asked their RPE. In the EXP group, it was paramount that they received no external cues of time or elapsed distance; therefore, RPE was asked at the beginning and at the end of each TT, and participants were asked to call out their RPE for every kilometer they thought had been completed.

Back to Top | Article Outline

Isometric testing of skeletal muscle function.

Immediately before the MIE test and the TT test, each participant's peak isometric force was measured on an isokinetic dynamometer (Biodex System 3 Pro; Biomedical Systems Inc., Shirley, NY). The peak force produced by an isometric contraction of the quadriceps muscles of the right leg was tested at an angle of 60°, with full knee extension being the 0° reference. Participants performed four 50%, two 70%, one 90%, and one 100% maximum voluntary contraction familiarization of 5 s each as a warm-up (3). After this, each participant performed four 5-s MVC with a 5-s rest between trials.

Back to Top | Article Outline

EMG activity.

Muscle recruitment was assessed during the isometric test, as well as during the TT by measuring EMG activity of the vastus lateralis muscle of the right leg. Silver/Silver chloride electrodes (Ambu Blue Sensor T, Ballerup, Denmark) with a rectangular sensor area of 27 mm2 (9 × 3 mm) set in an area of 254 mm2 of highly conductive gel were attached to the subjects lower limb before the start of all testing. The skin overlying the vastus lateralis was first shaven, abraded with sandpaper, and then cleaned using an alcohol swab. The bipolar electrode was placed according to the figures provided in the EMG software (MegaWin; Mega Electronics Ltd., Kuopio, Finland) on the area of greatest muscle bulk, lateral to the rectus femoris on the distal half of the thigh. Electrodes were linked to the unit used to record the EMG signal (Muscle Tester ME3000PB, Mega Electronics Ltd.) that was connected to a host computer. To secure electrodes in place, a sports bandage was wrapped around the participant's leg. EMG activity was recorded during the second MVC of the isokinetic test. EMG activity was recorded during the whole TT with a sample rate of 2000 Hz (15). The EMG activity coinciding with the second MVC was used to normalize the EMG values recorded during the TT. EMG was collected in raw form and stored using the MegaWin software to preserve the integrity of the signal and to allow a variety of subsequent manipulations using software (12). EMG data for the 10-s period surrounding each kilometer completed were cut from the raw data to be analyzed. To remove external interference noise and movement artifacts, the raw EMG signals from these 10-s periods were filtered with a second order Butterworth band-pass filter (10-400 Hz). The filtered EMG data were full-wave rectified and smoothed with a low-pass, second order Butterworth filter with a cut-off frequency of 10 Hz (14). Mean EMG (iEMG) was calculated over a 10-s period surrounding each elapsed kilometer and was normalized relative to the iEMG recorded during the MVC.

Back to Top | Article Outline

Statistical analysis.

After the MIE, an independent t-test was used to determine whether there were significant baseline differences between groups, of which none were observed. To examine differences between groups for completion time, PO, speed, and blood lactate ([B]La), a two-way ANOVA with repeated measures was used. Linear and quadratic interaction terms were used to examine any changes in variables between groups over TT. Where convergence in variables was observed over TT, differences in TT1 were examined using an independent t-test. Pearson correlations were used to analyze the relationship between iEMG and percentage of mean PO sustained. Data points for iEMG and percentage of mean PO were taken for every kilometer of each TT in each group so that 16 data points were available for both variables for each group. Differences in these correlations were then assessed using data matrix on the Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL) and regression analysis. Significance was accepted at P < 0.05.

Back to Top | Article Outline


Back to Top | Article Outline

Time to complete TT.

Independent-samples t-tests revealed that the time to complete TT1 was significantly different between groups (t16 = 2.51, P = 0.03), with the CON group completing TT1 in a faster time (367.4 ± 21 s) than the EXP group (409.4 ± 45.5 s). A two-way ANOVA with repeated measures revealed a significant main effect for time trial on completion time (P < 0.05) and time x group interaction effect on completion time (P < 0.00). A significant linear contrast in completion time between groups was found (F1,16 = 12.39, P = 0.00). No significant quadratic contrast between groups was revealed (F1,16 = 3.14, P > 0.05), and no significant group effect was evident (F1,16 = 2.21, P > 0.05). Therefore, the magnitude of the difference in completion time between groups reduced linearly across TT1-TT4 (see Fig. 1).

Back to Top | Article Outline

Mean speed and PO.

Independent-samples t-tests revealed both mean speed (t16 = −2.5, P = 0.02) and mean PO (t16 = −2.5, P = 0.02) in TT1 to be significantly different between the two groups. A two-way ANOVA with repeated measures found a significant main effect for time trial on mean speed (P < 0.00) and time trial x group interaction effect on mean speed (P < 0.05). No significant main effect for time trial on PO was found (P < 0.05), but a significant main effect for time trial x group on PO was observed (P < 0.05). A significant linear contrast in mean speed (F1,16 = 12.94, P < 0.0005) and mean PO (F1,16 = 13, P < 0.0005) was found, but no quadratic contrast between groups was found for either variable (P > 0.00). No significant group effect was observed in either variable (P > 0.05). Thus, differences in mean speed and mean PO between groups reduced linearly with each subsequent TT. Figure 2 shows groups mean PO over all TT.

Back to Top | Article Outline


A two-way ANOVA with repeated measures revealed a significant main effect for time trial on POmax (P < 0.05) but no significant main effect for time trial x group on POmax (P < 0.05). No linear or quadratic contrast in POmax was observed between groups (P > 0.05). A significant main effect for group was found (F1,16 = 9.15, P < 0.0005). Independent-samples t-tests showed a significant difference between groups in TT1, TT3, and TT4, with the CON group producing higher POmax (TT1 t16 = −3.13, P = 0.01; TT3 t16 = −2.72, P = 0.01; TT4 t16 = −2.73, P = 0.01). POmax for both groups across TT are shown in Table 1.

Back to Top | Article Outline

End blood lactate (E[B]La).

An independent samples t-test revealed a significant difference for E[B]La between groups in TT1 (t16 = −3.54, P < 0.0005). A two-way ANOVA with repeated measures found a significant main effect for time trial on E[B]La (P < 0.05) and a significant main effect for time trial x group on E[B]La (P < 0.05). A significant linear (F1,16 = 9.6, P < 0.0005) and quadratic (F1,16 = 9.35, P < 0.0005) contrast in E[B]La between groups was found; E[B]La for both groups across TT is shown in Table 1. E[B]La therefore displayed a convergence between groups from TT1 to TT4. No significant group effect was observed (P > 0.05).

Back to Top | Article Outline

Integrated EMG (iEMG).

In all the CON trials, iEMG dynamically tracked changes in PO. Figure 3 shows how iEMG and PO rise and fall in parallel during each TT. A significant positive correlation between iEMG and percentage of mean PO in the CON group was found (R = 0.52, P = 0.02). Furthermore, lower levels of iEMG were observed in the TT, where time to completion was longer and mean speed and mean PO were lower.

As shown in Figure 4, iEMG during the EXP TT did not track PO as it did during the CON TT. Indeed, a significant negative correlation between iEMG and percentage of mean PO in the EXP group was found (R = −0.499, P = 0.03). Regression analysis showed the correlations in iEMG and percentage of mean PO between groups to be significantly different (P < 0.0005).

Back to Top | Article Outline


It was hypothesized that distance-blinded participants in the EXP group with prior experience of the TT would be able to develop a pacing strategy that was as successful as the CON group. Our principal finding was that the EXP group developed a similar pacing strategy to the CON group, as the final EXP TT was only 2 s slower than the average completion time of the four CON group. This is remarkable considering that the EXP group was distance blinded and received no feedback, whereas the CON group knew distance and had feedback. Foster et al. (10) has stated that performance time differences between gold and silver medalists for distances of 1-4 km are in the region of 1%. In the present case, the difference between the average time to completion for the CON and the time to completion for the EXP group in TT4 was only 0.5%, indicating that the EXP group, despite their apparent disadvantage, could still produce times that were competitive compared with the CON group. Therefore, we accepted the hypothesis that an athlete can adopt a successful pacing strategy based just on prior experience.

The EXP group appeared to use TT1 as a "safe practice," completing it at an intensity that was low enough to complete a much longer distance without fatiguing. According to teleoanticipation (32), an athlete works backward from the end of an exercise bout and subconsciously creates a pacing strategy based on this knowledge of end point. In the case of the EXP TT1, the end point of exercise is not known, and therefore a competitive pacing strategy cannot be set. Indeed, the EXP mean PO for TT1 was only 60% of their peak power and never rose above 79% of their peak power. Once this first TT had been completed, however, the athlete would have had experience of the exercise and therefore some knowledge of the distance. Once familiarized with the task, the EXP group constructed a crude pacing strategy by working backward from the knowledge of end point. With each TT completed, more experience is gained, and therefore a more robust internal, relative, knowledge of distance is acquired. This in turn allows for a more accurate knowledge of end point and finally a more successful setting of pacing strategy.

Albertus et al. (1) found that cyclists could produce similar times for 20 km TT, even when during some of these TT they received incorrect distance feedback. This suggests that when an athlete knows the distance they have to complete, accurate distance feedback is not a prerequisite for optimum performance. Our data support this notion. By TT4, the EXP group had an idea of relative distance through their previous TT, and therefore by TT4, the only major difference between the groups was that the CON group had distance feedback. The similarity between the two groups in completion time at this stage indicates that distance feedback is surplus information that is not necessarily required to produce successful performances.

If exercise end point is not known, then according to teleoanticipation and CGM an initial pacing strategy cannot be set, which leaves only afferent feedback to regulate PO. Indeed, in EXP TT1, the group's mean B[La] did not rise above 3.76 mmol·L−1, and RPE did not go higher than 15. This value for B[La] would likely be below participants' maximal lactate steady state (17,20). Furthermore, RPE did not show the same linear increase across EXP TT1 as it did with all other trials. Therefore, the fact that participants maintained an intensity below these higher levels of exertion potentially provides support for the use of afferents as a power regulating system. Although B[La] has been suggested to be one of the afferents that is used by the CG to regulate PO (24), additional variables that were not measured in this research such as Pi, K+, H+, core temperature, or remaining anaerobic work capacity should be acknowledged as being other possible afferent signalers.

The observation that the differences between the groups E[B]La reduced with successive TT could be indicative of a more sensitive central safety mechanism in the EXP group, which limited exertion level due to the limited information that was received in the early TT. As more information is received through further TT, pacing strategy is resultantly more robust, and thus the sensitivity of the safety mechanism and reliance on afferent feedback is reduced, hence the increased level of [B]La that is "allowed."

After TT1, the EXP group displayed a linear increase in RPE during each trial and end RPE increased across TT. Furthermore, end RPE showed increases in accordance with PO in the EXP group over TT (Table 1). Therefore, in the EXP group, it seems likely that a pacing strategy may have been based around conscious feelings of exertion to manage PO. Indeed, the shortest time to completion was in TT4, where end RPE and E[B]La were highest. In the CON group, the fastest time to completion was achieved in TT1, where end RPE was at its lowest, yet E[B]La was at its highest. It may be the case, therefore, that the CON group initially set an overall pacing strategy for the whole 4 × 4 km, whereas the EXP group set strategy based on their level of knowledge and experience for each TT. The tracking of RPE against PO suggests that a pacing strategy is related in some fashion to RPE.

The participants in the EXP group were asked to shout out their RPE for every kilometer they thought they had traveled-a task that they could not do accurately, even by TT4. This seems at odds with the importance stressed of knowing the distance of a TT to produce a competitive performance. Therefore, it appears as though knowledge of "absolute" distance is not important, and thus the athlete must create a "relative" distance to be completed subconsciously. This relative or virtual distance is enough to generate a basic pacing strategy; however, increased experience of the bout will improve the accuracy of this virtual distance. Accuracy appears to improve up to a point where so much experience is gained that the knowledge of the virtual distance is so strong it enables the athlete to complete the bout in a time that is as competitive as an athlete who knows actual distance and receives feedback. St. Clair Gibson et al. (29) have previously postulated the existence of an internal clock that uses scalar rather than absolute time scales. This study lends support to this suggestion and also provides further insight into the relative importance and interplay between known distance and prior experience.

A surprising finding of the study was the marked difference in iEMG activity observed between the groups. Indeed, a significant positive correlation between iEMG and percentage of mean PO was found in the CON group, suggesting that iEMG dynamically tracked PO across TT (Fig. 3). This is a finding that is in accordance with numerous previous studies supporting theories of central control (3,18,31). However, a significant negative correlation between iEMG and percentage of mean PO was observed in the EXP group, and no tracking between these variables occurred (see Fig. 4). To ascertain the reasons for these observations, further research is needed, either examining more muscle groups or analyzing iEMG more frequently for longer periods during self-paced exercise.

In summary, this study has shown that with no knowledge of distance and no external feedback, cyclists can still complete a 4-km TT in a competitive time (in relation to a similarly placed athlete who knows these variables) when sufficient prior experience has been gained. Although cyclists do not appear to be able to accurately judge the elapsed distance they have traveled, this does not seem to matter because a relative or a scalar, rather than an absolute, distance is created within the brain from which a pacing strategy is apparently generated. Pacing seemed to be based, a least partially, on RPE and more predominantly so when the athlete received less external feedback. When distance knowledge and feedback are available, iEMG appears to dynamically track PO, suggestive of central regulation of exercise intensity. However, when external feedback is not available and a relative distance is calculated based on prior experience, this relationship appears to be more complex.

Back to Top | Article Outline


1. 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. Med Sci Sports Exerc. 2005;37(3):461-8.
2. Ansley L, Robson PJ, St Clair Gibson A, Noakes TD. Anticipatory pacing strategies during supramaximal exercise lasting longer than 30 s. Med Sci Sports Exerc. 2004;36(2):309-14.
3. Ansley L, Schabort E, St Clair Gibson A, Lambert MI, Noakes TD. Regulation of pacing strategies during successive 4-km time trials. Med Sci Sports Exerc. 2004;36(10):1819-25.
4. Baker AJ, Kostov KG, Miller RG, Weiner MW. Slow force recovery after long-duration exercise: metabolic and activation factors in muscle fatigue. J Appl Physiol. 1993;74(5):2294-300.
5. Borg G. Perceived exertion as an indicator of somatic stress. Scand J Rehabil Med. 1982;2(2):92-8.
6. Cady EB, Jones DA, Lynn J, Newham DJ. Changes in force and intracellular metabolites during fatigue of human skeletal muscle. J Physiol. 1989;418:311-25.
7. Calbert JA, Boushel R, Radegran G. Why is V˙O2max after altitude acclimatization still reduced despite normalization of arterial V˙O2 content? Am J Physiol. 2003;284(2):R304-16.
8. Davison RCR, Corbett J, Ansley L. Influence of temperature and protocol on the calibration of the CompuTrainer electromagnetically braked cycling ergometer. In: British Association of Sport & Exercise Sciences Annual Conference: Wolverhampton (UK); 2006. p. 12.
9. Enoka RM, Stuart DG. Neurobiology of muscle fatigue. J Appl Physiol. 1992;72(5):1631-48.
10. Foster C, Schrager M, Snyder AC, Thompson NN. Pacing strategy and athletic performance. Sports Med. 1994;17(2):77-85.
11. Gandevia SC. Spinal and supraspinal factors in human muscle fatigue. Physiol Rev. 2001;81(4):1725-89.
12. Gleeson NP. Assessment of neuromuscular performance using electromyography. In: Eston, RG, Reilly, T, editors. Kinanthropometry and Exercise Physiology Laboratory Manual: Tests, Procedures and Data, Volume 2, Exercise Physiology. London: Routledge; 2001. pp. 37-63.
13. Gregor RJ, Conconi F. Road Cycling. Oxford: Blackwell Science; 2000. p. 10.
14. Hettinga FJ, De Koning JJ, Broersen FT, Van Geffen P, Foster C. Pacing strategy and the occurrence of fatigue in 4000-m cycling time trials. Med Sci Sports Exerc. 2006;38(8):1484-91.
15. Hunter AM, St. Clair Gibson A, Lambert MI, Nobbs L, Noakes TD. Effects of supramaximal exercise on the electromyographic signal. Br J Sports Med. 2003;37(4):296-9.
16. van Ingen Schenau GJ, Dorssers WM, Welter TG, Beelen A, de Groot G, Jacobs R. The control of mono-articular muscles in multijoint leg extensions in man. J Physiol. 1995;484(Pt 1):247-54.
17. Jones AM, Doust JH. The validity of the lactate minimum test for determination of the maximal lactate steady state. Med Sci Sports Exerc. 1998;30(8):1304-13.
18. Kay D, Marino FE, Cannon J, St. Clair Gibson A, Lambert MI, Noakes TD. Evidence for neuromuscular fatigue during high-intensity cycling in warm, humid conditions. Eur J Appl Physiol. 2001;84(1-2):115-21.
19. Kent-Braun JA. Central and peripheral contributions to muscle fatigue in humans during sustained maximal effort. Eur J Appl Physiol Occup Physiol. 1999;80(1):57-63.
20. Kilding AE, Jones AM. Validity of a single-visit protocol to estimate the maximum lactate steady state. Med Sci Sports Exerc. 2005;37(10):1734-40.
21. Nielsen B, Hales JR, Strange S, Christensen NJ, Warberg J, Saltin B. Human circulatory and thermoregulatory adaptations with heat acclimation and exercise in a hot, dry environment. J Physiol. 1993;460:467-85.
22. Nikolopoulos V, Arkinstall MJ, Hawley JA. Pacing strategy in simulated cycle time-trials is based on perceived rather than actual distance. J Sci Med Sport. 2001;4(2):212-9.
23. Noakes TD, Peltonen JE, Rusko HK. Evidence that a central governor regulates exercise performance during acute hypoxia and hyperoxia. J Exp Biol. 2001;204(Pt 18):3225-34.
24. Noakes TD, St Clair Gibson A, Lambert EV. From catastrophe to complexity: a novel model of integrative central neural regulation of effort and fatigue during exercise in humans: summary and conclusions. Br J Sports Med. 2005;39(2):120-4.
25. Nybo L, Nielsen B. Hyperthermia and central fatigue during prolonged exercise in humans. J Appl Physiol. 2001;91(3):1055-60.
26. Peltonen JE, Tikkanen HO, Rusko HK. Cardiorespiratory responses to exercise in acute hypoxia, hyperoxia and normoxia. Eur J Appl Physiol. 2001;85(1-2):82-8.
27. Rejeski WJ, Ribisl PM. Expected task duration and perceived effort: an attributional analysis. J Sport Psychol. 1980;39:249-54.
28. St. Clair Gibson A, Lambert ML, Noakes TD. Neural control of force output during maximal and submaximal exercise. Sports Med. 2001;31(9):637-50.
29. St. Clair Gibson A, Lambert EV, Rauch LH, et al. The role of information processing between the brain and peripheral physiological systems in pacing and perception of effort. Sports Med. 2006;36(8):705-22.
30. St. Clair Gibson A, Noakes TD. Evidence for complex system integration and dynamic neural regulation of skeletal muscle recruitment during exercise in humans. Br J Sports Med. 2004;38:797-806.
31. Tucker R, Rauch L, Harley YX, Noakes TD. Impaired exercise performance in the heat is associated with an anticipatory reduction in skeletal muscle recruitment. Pflugers Arch. 2004;448(4):422-30.
32. Ulmer HV. Concept of an extracellular regulation of muscular metabolic rate during heavy exercise in humans by psychophysiological feedback. Experientia. 1996;52(5):416-20.
33. Ulmer HV, Schneider A, Neumahr S, Freitag W. Precision of time and velocity estimation as a base of teleoanticipation. Pflügers Arch. 1996;431(Suppl R 48):1-5.
34. Wilson JR, McCully KK, Mancini DM, Boden B, Chance B. Relationship of muscular fatigue to pH and diprotonated Pi in humans: a 31P-NMR study. J Appl Physiol. 1988;64(6):2333-9.
35. Zhang YY, Johnson MC, Chow N, Wasserman K. Effect of exercise testing protocol on parameters of aerobic function. Med Sci Sports Exerc. 1991;23(5):625-30.
36. Zohar D, Spitz G. Expected performance and perceived exertion in prolonged physical tasks. Percept. Motor Skills. 1981;52:975-84.


©2009The American College of Sports Medicine