Success in athletic competition often depends on finishing a competition in the shortest possible time. To accomplish this, the energy-producing capacities of the athlete must be exhausted either at, or shortly before, reaching the finish line. The process of regulating energy expenditure in this manner while minimizing the negative consequences of the developing fatigue is referred to as pacing; it is accomplished through variations in momentary power output (3). A common explanation for changes in power output is metabolite accumulation during relatively longer (2-10 min) (13) events, phosphagen depletion for relatively briefer (20-120 s) exercise (6), and glycogen depletion during prolonged exercise (15), all of which are combined into a broad concept known as peripheral fatigue. However, because peripheral fatigue cannot explain all observations during exercise-induced fatigue, an alternative model has recently been proposed. The central governor hypothesis suggests that pacing is under central neural control and that it may involve decreasing power output during time trials to prevent the development of a dangerous internal milieu (16,17). Though Hettinga et al. (12) have shown that the decrease in power output does not occur by downregulating EMG, pacing strategy may still be centrally regulated. This process can be either conscious or subconscious, and the feeling of fatigue will be the result of efferent inhibitory command processes (16,17). The central governor hypothesis has been formulated in terms of minimizing the likelihood of exertion-related failure of skeletal and cardiac muscle during exercise. Foster et al. (8) have described this process as an internal negotiation performed by the athlete regarding the estimation of the magnitude of the task remaining, the momentary aerobic power output, and the estimated remaining anaerobic energetic reserves. This process has been referred to as teleoanticipation by Ulmer (20). This negotiation seems to be more about maximizing performance than about minimizing the risk of catastrophic failure and is modified by a variety factors, both peripheral and central in origin (12). This may be as simple as the athlete developing a stable template of time/distance versus power output and anticipating how much the periphery will fail to respond to the neural drive as the event proceeds (12).
Pacing strategies can be divided into slow-start, even-pace, and all-out strategies. An all-out strategy is characterized by a fast start and a relatively large decrease in power output/deceleration towards the end of the event (3). It has been suggested that an all-out strategy might be optimal for events less than 1.5 min in duration (3,8). A pacing strategy characterized by relatively higher power during the first and last parts, with a more or less constant value through the middle part of an event, seems to be commonly adopted during supramaximal exercise of longer duration (5,8,17). De Koning et al. (3) have presented evidence from modeling studies suggesting that for events about 2 min in duration, such as the 1500-m speed skate, the best pacing strategy is a fast early start with a relatively early (15-20 s) reduction in anaerobic power output to a constant value. Recent experimental tests of this model suggest that this general pattern is adopted during simulated competitions, although the time required to reach constant anaerobic power is on the order of 30-60 s (3,7,8).
Against this background, we had made empirical observations that during very high-level competitions, such as the Olympic Games, some athletes will start with an unrealistically ambitious strategy: a so-called Olympic pacing strategy. The athletes who use this strategy will adopt an early pace designed to compete with the medalists, even though on the basis of pre-Olympic performances, the athlete would not be expected to finish near the medals. In most cases, it is expected that this will result in a deterioration of speed throughout the race and the athlete performing worse than might have been expected from precompetition performances. However, this strategy is apparently successful just often enough to encourage the behavior among athletes. This suggested to us that strong extrinsic motivation might override the experience programmed pacing template. Although it is not feasible to experimentally replicate these extraordinary levels of extrinsic motivation, this study was designed to determine whether the presence of a lesser level of extrinsic motivation, applied immediately before exercise, might influence either the pacing strategy or the ultimate performance during cycling time trials. To our knowledge, no data have been published regarding the stability of pacing strategies in response to extrinsic motivation.
Seven well-trained male cyclists were the subjects for this study. All subjects provided informed consent before participation, and the protocol had been approved by the institutional review board for the protection of human subjects at the University of Wisconsin-La Crosse. Descriptive data of the subjects are provided in Table 1. The cyclists were competitive at a regional level and were habituated to cycling time trials via participation in other studies in our laboratory (7,8). Additionally, to ensure task orientation, a habituation trial was performed by each subject. The testing took place approximately 1 month after the end of the competitive season, during a period of light training (90-120 min at an intensity generally below the ventilatory threshold (VT), 3-4 d·wk−1).
The subjects completed an incremental exercise test to fatigue to allow definition of V˙O2peak and the VT and respiratory compensation threshold (RCT). The test was performed on an electrically braked cycle ergometer (Lode, Groningen, the Netherlands). After a 10-min warm-up at a self-selected intensity, subjects cycled for 3 min at 25 W. Thereafter, the workload increased by 25 W every minute until the subject could no longer continue. Subjects were required to remain seated for the duration of the test. The highest power output maintained for a full minute, plus the interpolated power output for fractions of the terminal minute, was defined as [email protected]˙O2peak. V˙O2 was measured by open-circuit spirometry (Quinton Q-plex, Seattle, WA). V˙O2peak was defined as the highest V˙O2 observed during 30 s. Heart rate (HR) was recorded every 5 s using radio telemetry (Polar Vantage NV, Kempele, Finland). Maximum HR (HRmax) was defined as the highest HR measured during 5 s.
Each subject completed four 1500-m time trials with a 48- to 96-h interval between trials. The time trials were performed on a racing cycle interfaced with a wind load simulator with a heavy flywheel (Findlay Road Machine, Toronto, Canada). Previous studies from our laboratory have suggested that this apparatus provides realistic perceptual and power output responses during time trials (5,7,8). Power output, accumulated distance, elapsed time, and HR were recorded every second using an SRM dynamometer (Koningskamp, Germany). Although the 1500-m distance is not specifically relevant to cycling, the primary interest of our research group has been related to speed skating (4,6), where 1500 m is regarded as the single most representative distance in this sport. Cycling was chosen as an experimental model because of the strong relationship between cycling and speed skating.
Each time trial session began with a standard warm-up. The subjects began cycling for 3 min at 20 km·h−1 (~75 W), after which they increased their velocity by 2 km·h−1 (~20 W) each minute. After each minute, the subjects rated their perceived exertion using the category ratio rating of perceived exertion scale. Once the subjects had rated the intensity equal to 7 (very hard), they reduced their velocity to 20 km·h−1 for 2 min. After this period of active rest, the subjects cycled for 5 min at a velocity requiring a V˙O2 equal to approximately 90% of the VT observed during the screening incremental exercise test. The velocity was based on the power-velocity relationship established for this bike and the power at VT during the incremental test:
where v is velocity in kilometers per hour, and power is expressed in watts.
Gross efficiency (GE) was calculated on the basis of the V˙O2 and power output during the last 2 min of the 5-min submaximal ride (9). After the efficiency ride, the subjects rode at 20 km·h−1 for 5 min, and then they performed an all-out sprint of about 10 s. Peak power was accepted as the highest 5-s average during the all-out sprint. After performing the sprint, the subjects rode an additional 5 min at 20 km·h−1 to allow recovery before beginning the time trial. When the 5 min were completed, the subjects stopped cycling for 1 min. At the end of this minute, the 1500-m time trial was started. For a graphic presentation of the warm-up protocol, see Foster et al. (7).
Before each subsequent time trial, the subjects were provided feedback about their velocity profile and average velocity in consecutive segments of preceding time trials, just as an athlete would be aware of previous performances and split times in previous competitions. The only instruction given before each time trial was to finish in the shortest possible time. To provide general motivation, there was a lab record sheet posted in the laboratory listing the best performances from previous studies in our laboratory, both in terms of total time and average power output (expressed in watts per kilogram). Before the fourth and last time trial, the subjects were unexpectedly offered a monetary reward ($100; external motivation) if they were able to beat their best time by more than 1 s. It is recognized that, ideally, the trial with the monetary reward would have been presented in random order. However, we were concerned about the negative impact if a monetary trial was not offered in subsequent trials. Given the generally high reliability of time trial performance in our lab (5,7,8), we made the decision to only offer the incentive during the terminal trial.
Using the following equations, mechanical work attributable to aerobic and anaerobic energy sources was calculated for each 10-m segment in the race. First, aerobic metabolic power attributable to aerobic sources (PV˙O2) was calculated from the V˙O2 and V˙CO2 data according to Garby and Astrup (9) as
where PV˙O2 is the aerobic metabolic power output in watts, V˙O2 is oxygen consumption during each segment in liters per minute, and RER is respiratory exchange ratio. We assumed that an RER in excess of 1.0 was attributable to buffering. Accordingly, RER values > 1.0 were treated as if they equaled 1.0 relative to calculating metabolic work.
GE was calculated as
where PO is power output in watts during the last 2 min of the submaximal ride, and PV˙O2 is aerobic metabolic power output in watts during the last 2 min of the submaximal ride. Then, mechanical power output attributable to aerobic energy sources during the time trial could be calculated with the following equation:
where Paer is aerobic mechanical power output in watts.
Consequently, mechanical power output attributable to anaerobic energy sources could be calculated by subtraction as:
where Panaer is anaerobic mechanical power output in watts, and Ptot is the total measured power output. For all calculations, we assumed that the GE measured during the submaximal ride at approximately 90% VT was representative of the efficiency during the time trial. This assumption has been used in both our previous studies (3,7,8) and elsewhere (11,19).
Statistical analysis was accomplished using SPSS (version 12), with repeated-measures ANOVA used to compare time, total power output, aerobic and anaerobic power, and V˙O2 for the entire time trial and for different segments of the time trial. Statistical significance was accepted when P < 0.05. Additionally, beyond the planned analysis in relation to the experimental treatment, we ranked the performances by quality (fastest to slowest) per subject and performed a repeated-measures ANOVA on the dependent variables based on the level of performance.
The times (mean ± SD) required for completion of the second, third, and fourth (incentivized) time trial were 133.1 ± 2.1, 134.1 ± 3.4, and 133.6 ± 3.0 s, respectively. The average time, total, aerobic and anaerobic power output, and V˙O2 are presented in Table 2. The serial pattern of responses for average power output, power attributable to aerobic and anaerobic energy sources, and V˙O2 are graphically presented in Figure 1 for the first 100 m and after that for every 200 m. On average, the fastest time was recorded for the second time trial: 133.1 ± 2.1 s. However, there were no significant differences between trials for time, power output, or aerobic and anaerobic contribution. Only two cyclists were able to improve their best time by more than 1 s on the last time trial.
The results when the trials were ranked from fastest to slowest are presented in Table 3 and Figure 2. The mean (± SD) times for the fastest, second-fastest, and third-fastest trials were 131.9 ± 2.4, 133.4 ± 2.4, and 135.4 ± 2.6 s, respectively (P < 0.05). The pattern of power output was characterized by a higher power output in the first 300 m during the fastest trial (compared with both the second-fastest and third-fastest trials) and during the 100- to 300-m segment in the second-fastest trial compared with the slowest trial (P < 0.05). Because there was no difference in the pattern of V˙O2 and power output attributable to aerobic sources, the higher values of power output were attributable to higher anaerobic power output during the relevant segments (P < 0.05).
We hypothesized that the fourth time trial, in which an extra monetary reward was provided, would motivate the cyclists to adopt a more all-out or Olympic pacing pattern. We did not hypothesize that this would necessarily be associated with a faster time, because the informal observations that had suggested this study had indicated that athletes frequently fail to improve when using an overly aggressive pacing pattern. The results revealed no significant differences between the final times on the three different time trials. Only two cyclists were able to beat their time by more than 1 s during the incentivized trial. To support the hypothesized effect of incentive on pacing strategy, a higher anaerobic contribution, particularly during the first 300 m, was expected in the fourth time trial, because variations in total power output have been shown to be mainly caused by variations in anaerobic energy contribution (12). The total and anaerobic power output for the first 100 m and the subsequent 200 m were not significantly different for the different trials. Consequently, the experimental hypothesis is rejected.
The additional monetary reward was intended to increase the cyclists' motivation to accomplish a maximal performance in their last time trial. However, the average best performance was achieved during the second time trial, with no evidence of significance related to either testing sequence or to the incentive. It is possible that the motivation was not increased by the provided reward. An explanation for this might be found in the concept of intrinsic and extrinsic motives. Wann (21) suggests that motivation can be defined as a process of arousal within an organism that helps direct and sustain behavior. Motivation may be intrinsic or extrinsic. Intrinsic motivation lies within an individual and involves enjoyment and involvement in a task. Extrinsic motivation lies outside an individual and involves the rewards of performing a task. Money is an excellent example of an extrinsic motivator. However, in situations in which an individual is performing a task because of intrinsic motivation, presenting the individual with extrinsic rewards to perform the activity may actually lower intrinsic motivation (2). Thus, the additional monetary reward could have lowered the intrinsic motivation of the cyclists and, therefore, not changed the pattern of performance. Alternatively, the monetary reward may have been too small or have been presented too close to the start of the trial to affect the pacing strategy. Simple observation, as well as systematic studies of athletes, suggests that performance is often rehearsed multiple times before the actual competition. If, as suggested by St Clair Gibson and Foster (18), the pacing strategy is dependent on a preestablished template defining the pattern of power output during an event, and athletes typically have a long experience in developing a particular template, it may take either a very large motivation (e.g., Olympic competition) or an appreciable time before competition to modify this template. This is supported conceptually by the transtheoretical model of behavioral change (14).
Both the present study and a previous study from our laboratory have demonstrated a lack of intrasubject differences during time trial exercise. Foster et al. (8) propose that the failure to observe changes in the pattern of power output with repeated performance suggests that the ability to monitor energy resources is learned early in an athlete's experience and is not a highly specialized, learned response. St Clair Gibson et al. (16) have suggested that the motor sequences representative of the performance template are almost certainly programmed into the motor cortex from prior athletic performance, potentially from the time of childhood, and that these would regulate even an initial sprint activity. It is probable that this applies to supramaximal exercise, which would explain why the performance was impossible to change even with the extrinsic motivation provided in this trial.
Although there were no significant differences between the second, third, and last self-paced time trials, there were clear differences when the fastest time trial for each cyclist (regardless of sequence) was compared against their second and third performances, regardless of the sequence. In Figure 2, it can be seen that the fastest time trial is achieved with a more all-out pacing pattern. The average power output during the first 100 m was 567 ± 98 W for the fastest time trial, compared with 491 ± 82 and 493 ± 93 W for the second and third performance times, respectively. For the next 200 m, these values are 565 ± 91, 513 ± 41, and 484 ± 88 W, respectively.
Modeling results by de Koning et al. (3) have shown that the best result on a 1000-m time trial (performance time approximately 60 s) should be obtained when an all-out strategy is used. Conversely, the fastest time on a 4000-m time trial should be achieved with a short initial high-power output, with a transition (after approximately 12 s) to an evenly paced race. In agreement with the present study, previous experimental studies from our group using a spontaneous pacing pattern have shown that 1500-m events are characterized by a decreasing power output during the first 30-40% (30-60 s) of the event rather than 12-15 s as predicted by modeling studies, with a relatively constant power output for the remainder of the time trial (3,7,8). In the energy-flow model of de Koning et al. (3), exhaustion of the anaerobic capacity was assumed. However, in both the present study and other studies from our laboratory (4,7,8), the anaerobic capacity was apparently not fully used before the end of the 1500-m trial, suggesting that the subjects hold some energy reserves even during work that is nominally exhausting. This supports the findings of Foster et al. (5,7,8) that in 1000- to 3000-m time trials, the anaerobic contribution never falls to zero. De Koning (4) has characterized this anaerobic profile by a first-order system in the power-balance model. Willberg and Pratt (22) found a difference in race profile between winners and losers in the 1000-m track cycling event. In contrast to the results of de Koning et al. (3), they suggest that the most likely cause of losing was a too rapid acceleration at the start of the race and, therefore, an inefficient use of the cyclists' anaerobic capacity. However, there was a fairly strong correlation (r = 0.71) between the time for the first lap and the final time, suggesting that (in general) a relatively rapid start is advantageous in events of this duration.
Although the performance time is longer than 2 min, an analysis of pacing strategies of elite competitors in 2000-m rowing (6-8 min) has shown that rowers adopt a fast start strategy regardless of finishing position, gender, or rowing mode (indoor or outdoor) (10). Bishop et al. (1) investigated the effects of an all-out and even-pace strategy on 500-m kayak performance (performance time 115-125 s); their results indicate that performance is significantly better when an all-out strategy is used. These results support modeling results by de Koning et al. in that the strategy of Bishop et al. was a 10- to 15-s maximal start followed by an even pace.
Although it is interesting to compare cycling with other modes of exercise, and it is evident from this comparison that there is no consensus regarding all-out or even-pace strategies, it is possible that cycling encourages a more all-out pacing strategy. It is apparent that the skill of an activity decreases as an athlete fatigues (8). In cycling, skill may be quantitatively less important compared with skating, kayaking, and rowing. Additionally, in cycling, the body weight is supported by the bicycle, and the considerable kinetic energy present during cycling allows for minimal slowdown even with a loss of power output, something not evident during running.
In conclusion, it can be stated that extrinsic motivation provided (in the form of a monetary reward) immediately before a simulated competition did not improve time trial performance. No significant differences were evident for total power output, aerobic and anaerobic power output, and V˙O2 for the total time trial or for intermediate fractions of the time trial. This suggests that performance on a 1500-m time trial is not readily changeable. However, the tendency to achieve faster times when the spontaneous pacing strategy is more all-out in character suggests that the source of improvement in middle-distance events may rely on the ability to tolerate early disturbances in homeostasis. Whether this is attributable to greater preexercise substrate concentrations, better tolerance of accumulated metabolites, faster V˙O2 kinetics, or other factors remains to be determined. In any case, it suggests that well-trained athletes must take some calculated risk to improve performance. However, it may take either a large motivation (such as a major competition) or a longer period to contemplate a new strategy to change a previously developed competitive template.
1. Bishop, D., D. Bonetti, and B. Dawson. The influence of pacing
strategy on V˙O2
and supramaximal kayak performance. Med. Sci. Sports Exerc.
2. Deci, E. L., and R. Ryan. Intrinsic Motivation and Self-Determination in Human Behavior
. New York, NY: Plenum Press, pp. 57-89, 1985.
3. de Koning, J. J., M. F. Bobbert, and C. Foster. Determination of optimal pacing
strategy in track cycling with and energy flow model. J. Sci. Med. Sport
4. de Koning, J. J., C. Foster, J. Lampen, F. Hettinga, and M. F. Bobbert. Experimental evaluation of the power balance model of speed skating. J. Appl. Physiol.
5. Foster, C., A. C Snyder, N. N. Thompson, M. A. Green, M. Foley, and M. Schrager. Effect of pacing
strategy on cycle time trial performance. Med. Sci. Sports Exerc.
6. Foster, C., M. Schrager, A. C. Snyder, and N. N. Thompson. Pacing
strategy and athletic performance. Sports Med.
7. Foster, C., J. J. de Koning, F. Hettinga, et al. Pattern of energy expenditure during simulated competition. Med. Sci. Sports Exerc.
8. Foster, C., J. J. de Koning, F. Hettinga, et al. Effect of competitive distance on energy expenditure during simulated competition. Int. J. Sports Med.
9. Garby, L., and A. Astrup. The relationship between the respiratory quotient and the energy equivalent of oxygen during simultaneous glucose and lipid oxidation and lipogenesis. Acta Physiol. Scand.
10. Garland, S. W. An analysis of the pacing
strategy adopted by elite competitors in 2000 m rowing. Br. J. Sports Med.
11. Gastin, P. B., D. L. Costill, D. L. Lawson, K. Krzeminski, and G. J. McConnell. Accumulated oxygen deficit during supramaximal all-out and constant velocity exercise. Med. Sci. Sports Exerc.
12. Hettinga, F. J., J. J. de Koning, F. T. Broersen, P. van Geffen, and C. Foster. Pacing
strategy and the occurrence of fatigue in 4000-m cycling time trials. Med. Sci. Sports Exerc.
13. Karlsson, J., and B. Saltin. Lactate, ATP and CP in working muscles during exhaustive exercise. J. Appl. Physiol.
14. Prochaska, J. O., and C. C. DiClemente. Stages and process of self-change of smoking: toward an integrative model of change. J. Consult. Clin. Psych.
15. Rauch, H. G. L., A. St Clair Gibson, E. V. Lambert, and T. D. Noakes. A signalling role for muscle glycogen in the regulation of pace during prolonged exercise. Br. J. Sports Med.
16. St Clair Gibson, A., M. I. Lambert, and T. D. Noakes. Neural control of force output during maximal and submaximal exercise. Sports Med.
17. St Clair Gibson, A., E. V. Lambert, L. H. G. Rauch, et al. The role of information processing between brain and peripherial physiologic systems in pacing
and perception of effort. Sports Med.
18. St Clair Gibson, A., and C. Foster. The role of self talk in the awareness of physical state. Sports Med.
19. Seresse, O., J. A. Simoneau, C. Bouchard, and M. Moulay. Aerobic and anaerobic energy contribution during maximal work in 90s determined with various ergocycle workloads. Int. J. Sports Med.
20. Ulmer, H. V. Concept of an extracellular regulation of muscular metabolic rate during heavy exercise in humans by psychophysiological feedback. Experientia
21. Wann, D. L. Sport Psychology
. Upper Saddle River, NY: Prentice Hall, pp. 107-120, 1997.
22. Willberg, R. B., and J. Pratt. A survey of the race profiles of cyclists in the pursuit and kilo track events. Can. J. Sports Sci.