Pacing strategy refers to the variation of speed over the race by regulating the rate of energy expenditure (6,7,19). This involves controlling multiple variables, all of which interact with each other. First, biomechanical factors are of importance to regulate energy expenditure efficiently. External power losses must be minimized. For a 4000-m time trial, it is clear that an evenly paced strategy is optimal to minimize external power losses (15,19). Because athletes are very fatigued at the end of the time trials, however, it can be hypothesized that attempts to maintain power output were achieved only by progressively increasing the subjects' effort. Thus, fatigue is another important factor in pacing strategy and may depend on the presence of a preestablished estimate of the power output that can be sustained over the duration of the event.
Fatigue has been described as a decrease in force production and power output, or as the inability to maintain force despite the presence of an increased perception of action (12). It has traditionally been believed that decreases in force production and power output are attributable to an impaired muscle contractile function caused by physiological changes within the muscle. This phenomenon is generally described as "peripheral fatigue" and appears in the presence of an unchanged or increasing neural drive (29). If power output is regulated by a peripheral fatigue mechanism, an increased central neural recruitment of additional motor units would be expected to occur to compensate for the reduced power output from the fatiguing motor units (11).
Bangsbo et al. (2) showed that metabolic changes in muscles were not tightly correlated with decreasing power output. They suggested that particularly central factors might cause fatigue during prolonged exercise. An alternative hypothesis has been proposed (23) in which a central neural governor acts to reduce muscle recruitment during fatiguing exercise. In light of this hypothesis, it has been suggested that in time trials, central fatigue occurs as a safety mechanism for preventing potentially dangerous metabolic disturbances and preserving the integrity of the muscle fibers (24). In this case, the reduction in power output is suggested to be the result of a reduction in central neural drive. In support of this hypothesis, studies comparing integrated electromyography (iEMG) and power output in time trial exercise with self-paced strategy demonstrated that iEMG decreased in parallel with power output during bouts of high-intensity time trial exercise during a 100-km time trial (28).
Further support for the central governor hypothesis was provided by Ansley et al. (1), who found that iEMG and power output rose at the end of a 4000-m cycling time trial exercise. They concluded that power output paralleled iEMG, and that fatigue is not regulated by peripheral mechanisms but is centrally controlled. This conclusion was drawn based on iEMG measurements during self-paced exercise in the rectus femoris muscle only. To determine whether power output parallels iEMG and whether central fatigue plays a role in time trial exercise, a range of pacing strategies must be performed, measuring iEMG in various muscles. If fatigue during time trial exercise was centrally controlled, as suggested in the governor hypothesis of Noakes (23), it should be visible in a pacing strategy where a relatively high power output is delivered in the beginning of the trial, with a subsequent slowdown in the latter part.
This study was conducted to create a better understanding of the role of central and peripheral fatigue on performance during middle-distance exercise. Therefore, influence of pacing strategy on the occurrence of fatigue was investigated by comparing changes in power output and iEMG activity during 4000-m cycling time trials. Furthermore, the contribution of the anaerobic and aerobic energy system to the total mechanical power output was calculated for the different pacing strategies.
Eight men, highly motivated to produce maximal time trial efforts, were recruited for this study. They were all well trained and familiar with cycling exercise. Before the experiment, all subjects were informed of the nature of the investigation, after which they gave written informed consent. The protocol has been approved by the human ethics committee. The characteristics of the subjects are given in Table 1.
The experiment consisted of five tests with at least 48 h between trials. All subjects completed a maximal incremental exercise test and four cycling time trials of 4000 m, all conducted on a custom-made, electronically braked laboratory cycle ergometer simulating real competition. The incremental test was cycled at a pedal frequency of 90 rpm. After a warm-up of 2 min at a power output (PO) of 100 W, the test started at a PO of 150 W. PO was increased by 30 W every 3 min until the subject could no longer maintain the required power output or the pedal frequency dropped below 80 revolutions per minute (rpm). All trials were performed under standard conditions with a constant temperature of 15°C, 50% relative humidity (RH). Oxygen consumption (V̇O2), respiratory exchange ratio (RER), PO, and heart rate (HR) were recorded during all tests. Respiratory gas exchange was measured breath-by-breath using open-circuit spirometry (Oxycon Alpha, Mijnhardt, The Netherlands). Before each test, the gas analyzer was calibrated using a Jaeger 3-L syringe, room air, and a standard gas mixture (5.04% CO2). HR was recorded every 15 s using radio telemetry (Polar Electro, Kempele, Finland).
The first time trial was designed to determine at which PO subjects had to cycle their different strategies. After warming up for 7 min and 30 s (100 W, 90 rpm), including three sprints of 15 s at minutes 5 (200 W), 6 (400 W), and 7 (300 W), subjects were told to complete the distance as fast as possible, as in competition. Subjects remained seated during the whole trial, and oxygen consumption was measured. After the time trial, subjects cooled down for 5 min at 100 W (90 rpm). The profile of the entire protocol is shown in Figure 1.
The three remaining trials, performed in random order, were aimed at an evenly paced time trial (EVEN), a submaximal time trial (SUB), and a supramaximal time trial (SUPRA). All trials started with a warm-up at 100 W (90 rpm) of 7 min and 30 s (Fig. 1). During the first 5 min of the warm-up, efficiency was estimated (10). After 5, 6, and 7 min, short sprints of 15 s at a higher PO (200, 400, and 300 W) were performed. After warming up, the subjects had 1 min of rest before the time trial. As in the first trial, subjects remained seated during the entire trial. PO was dictated during the first 2000 m of all three time trials, based on the mean power output (MPO) measured during the first trial. Elapsed distance and virtual velocity were the only feedback given to the subjects. During the first 2000 m of the SUB trial, the subjects were instructed to maintain a virtual velocity corresponding to a PO representing 95% of the MPO measured during the first trial. During EVEN, the subjects rode the first 2000 m at a velocity corresponding to the MPO of the first trial. During the first 2000 m of the SUPRA, 105% of the MPO measured during the first trial was dictated. In all time trials, after the first 2000 m, the subjects were instructed to complete the time trial in as little time as possible. After performing the time trial, subjects cooled down for 5 min at 100 W (90 rpm).
According to the protocol, three different strategies performed by all subjects were the aim, which are broadly representative of spontaneous variations in pacing strategy adopted by athletes (9,20). Because in all strategies subjects were instructed to get the bike going as fast as possible, all strategies had a high peak power output during the first 200 m of the trial, which is not representative of the rest of the race. Therefore, the first 200 m were not incorporated in these analyses.
During all time trials, electromyographic muscle activity (EMG), virtual velocity, PO, V̇O2, RER, and HR were measured. EMG was measured over the rectus femoris (RF), vastus lateralis (VL), and biceps femoris (BF) muscles. These muscles are highly active during cycling (17). Anaerobic power output (Pan) and aerobic power output (Paer) were calculated every 200 m as described in de Koning et al. (20), using V̇O2, RER, and efficiency estimated during the warm-up (10). At the conclusion of 2000 and 4000 m, the rate of perceived exertion (RPE) was noted on a scale of 1-10, based on the Borg scale (4). Blood lactate concentration (BLC) was measured before the time trial, after 2000 m, and at the end of the time trial using dry chemistry (Lactate Pro, Arkray, Kyoto, Japan).
To normalize EMG data, the maximal EMG level of each muscle was measured during maximal voluntary contractions (MVC). Before warming up for the time trials, each subject performed six isometric MVC (three extensions and three flexions of the knee joint of the left lower limb) of 5-s duration with 2 min of rest in between. During MVC, subjects sat in a specially designed chair in which the knee angle was set at 90°, and torque was measured by a built-in load cell. During extension, maximal EMG level of the RF and VL muscles were measured, and during flexion, maximal EMG level of the BF muscle was measured. During both the MVC test and all time trials, muscle recruitment was assessed by measuring EMG activity of the RF, VL, and BF muscles of the left upper leg. EMG activity was recorded during the whole time trial with a sample rate of 2000 Hz (14). The EMG activity coinciding with peak torque of the best-effort MVC was used to normalize the EMG values recorded during the time trials.
To remove external interference noise and movement artifacts, the raw EMG signals were filtered with a second-order Butterworth band-pass filter (10-400 Hz). The already filtered EMG data were full-wave rectified and smoothed with a low-pass, second-order Butterworth filter with a cutoff frequency of 10 Hz. Mean EMG (iEMG) was calculated over every successive 200 m.
Differences in mean values for efficiency (measured before the time trial) and final time between strategies were tested using ANOVA repeated measures.
For PO, Paer, Pan, RPM, HR, BLC, RPE, and iEMG, the mean values for the 200- to 2000-m interval were compared with mean values for the 2000- to 4000-m interval per strategy using 3 × 2 (strategy × interval) ANOVA repeated measures. If main effects were found, a pairwise comparison with Bonferroni adjustment was performed to find which variables differed significantly between strategies. In case of an interaction effect of strategy and interval, paired-sample t-tests were performed (P < 0.05) to compare mean values of variables over the first interval of the time trial with mean values of the corresponding variables over the last 2000-m interval.
PO, Paer, and Pan profiles are shown per 200-m segment in Figure 2a (SUB), 2b (EVEN), and 2c (SUPRA). A 3 × 2 ANOVA with repeated measures revealed a main effect for strategy for PO. A pairwise comparison (with Bonferroni adjustment) showed a significant difference between SUB and SUPRA. Furthermore, an interaction effect was found. Paired-sample t-tests showed that PO increased significantly in SUB. In EVEN, no significant difference was seen, and in SUPRA, PO decreased significantly (Table 2). For Paer, a main effect of strategy was found, but a pairwise comparison did not show any significant differences between strategies. For interval, a main effect was found for Paer. Also, an interaction effect was found. For all strategies, paired-sample t-tests showed that Paer was significantly higher in the second interval of the trial compared with the first interval (Table 2). For anaerobic power output, no main effects were found for strategy and interval. An interaction effect was found. A significant increase in Pan comparing the second interval of the trial with the first interval was found in the SUB trial. In the SUPRA time trial, Pan decreased significantly (Table 2).
iEMG and PO profiles are shown per 200-m segment for each strategy in Figure 3. For iEMG, a 3 × 2 ANOVA with repeated measures performed per muscle revealed main effects for strategy and interval for the VL and RF muscles, but not for the RF muscle. For all three muscles, an interaction effect was found. Paired-sample t-tests revealed a significant increase between iEMG in the second half compared with the first half of the race for the VL and BF (Table 2). For the RF, no significant differences were found between the time trial intervals. For the VL and BF muscles, the iEMG of the second 2000 m of the time trial was higher than the first part for all strategies. For the RF muscle, however, a significant increase in iEMG level was not found. Thus, iEMG increased significantly (VL and BF) or remained constant (RF) in SUPRA, even when PO decreased significantly (Fig. 3). It can also be seen that iEMG never rises above 50% MVC in any of the pacing conditions.
Heart rate, pedal frequency, blood lactate concentration, and rate of perceived exertion.
For RPM, a main effect of strategy was found as well as an interaction effect. Paired-sample t-tests comparing the second interval of the trial with the first revealed a significant increase in the SUB strategy and a significant decrease in the SUPRA strategy. Mean values for the first and second interval per strategy are shown in Table 3. For HR, strategies did not differ significantly. HR increased significantly in the second half compared with the first half for all strategies. For BLC, a main effect for strategy was found, but no significant differences between strategies were found in the pairwise comparison. Furthermore, BLC was significantly higher in the second interval compared with the first interval for all strategies (Table 3). For RPE, a main effect for strategy was found. A pairwise comparison with Bonferroni adjustment showed a significant difference between SUB and SUPRA. Also, a main effect for interval was found as well as an interaction effect. Paired-sample t-tests revealed significant increases in RPE in SUB and SUPRA (Table 3).
Mean values for efficiency and final time.
Mean values for efficiency measured during the warm-up before the trial did not differ significantly between SUB (0.17 ± 0.01%), EVEN (0.18 ± 0.02%), and SUPRA (0.17 ± 0.01%). Also, final times were not significantly different per strategy (351.1 ± 6.2, 348.9 ± 7.5, and 345.8 ± 5.8 s, respectively).
The main focus of the present study was to investigate the influence of different pacing strategies on fatigue by comparing changes in PO and iEMG activity during 4000-m cycling time trials to create a better understanding of the role of central and peripheral fatigue on performance during middle-distance time trial exercise. Previous studies indicated that PO profile mirrored iEMG pattern in self-paced exercise. In these studies, iEMG was measured solely in the RF muscle (1,14,18,27). We measured iEMG in three leg muscles (RF, VL, BF) and found that iEMG patterns of the muscles differed. iEMG of the RF did not change or decreased slightly in all strategies, whereas iEMG of BF and VL increased significantly in all strategies, even when PO decreased. An increase in iEMG is contrary to the predictions of the central governor hypothesis but consistent with a peripheral locus of fatigue.
Ebenbichler et al. (5) has also found differences in EMG patterns of different muscles, in particular between mono- and biarticular muscles. Their functional difference might explain the occurrence of differences in fatigue patterns found between mono- and biarticular muscles. Biarticular muscles (RF, BF) seem to be responsible for controlling the distribution of net moments around the joints crossed (16), and are thus particularly involved in the regulation of movement direction and its external force (e.g., net torque on an environment). As a consequence, differences in seating position and technique will influence iEMG pattern. Conversely, monoarticular muscles (VL) seem to work mainly as work generators, and thus differences in seating position and technique are expected to have little influence on the iEMG pattern. Therefore, the monoarticular VL muscle may potentially be the most appropriate muscle to monitor for evidence of central downregulation. The choice of which muscle to monitor may be a significant issue in future studies of the causes of fatigue.
In the previously mentioned studies (1,28) on iEMG and PO during time trials, PO seemed to mirror iEMG. Additionally, these studies measured iEMG activity only in the RF muscle and only employed self-selected pacing strategies. The PO profile was either even or slightly increasing toward the end of the trial, as was iEMG pattern of the RF. We studied a range of strategies, with a particular interest in a strategy during which PO could not be maintained despite increasing effort.
In SUB, where PO increased toward the end of the trial, we found both iEMG and PO to increase in all muscles. This is in accordance with earlier studies (1,28). However, iEMG of the VL and BF muscles showed a significant increase toward the end of the trial in all conditions, irrespective of PO profile. RF activity was independent of PO. This leads us to the main finding of our study: a decrease in PO was accompanied by constant or increasing iEMG activity. This is contrary to one of the central tenets of the central governor hypothesis, which suggests that decreases in PO occur secondary to downregulation of iEMG, potentially to prevent the development of unreasonable disturbances of homeostasis within the muscle. Although subjects experienced subjective fatigue in all trials, as shown by the high RPE and BLC concentration at the end of the trial, no evidence was seen of a central downregulation of muscle recruitment with decreasing power output. On the basis of these findings, centrally mediated downregulation of neural drive does not necessarily accompany fatigue in middle-distance time trial exercise. Peripheral fatigue, as shown by the decreasing PO in the presence of an unchanged or increasing central neural drive and the high occurring BLC, may be a better explanation for the cause of fatigue.
In their recently published papers, Noakes et al. (25) and Lambert et al. (21) addressed the importance of peripheral fatigue in pacing, again in a different form. They suggested that changes in peripheral physiological systems act as afferent signalers and that peripheral metabolites provide information to the central controller by way of afferent neural pathways, and are therefore an integral part of the regulatory process (25).
Another finding of our study was that, in all strategies, maximal iEMG level was not higher than approximately 50% MVC for BF, 30% MVC for VL, and 15% MVC for RF. St. Clair Gibson et al. (27) found that 20% of MVC was attained in RF muscle. The fact that fewer fibers are recruited during cycling time trial exercise than during MVC was also interpreted in favor of the governor hypothesis, as an indication that exercise performance had to be regulated by the central nervous system to ensure that catastrophic loss of homeostasis does not occur during normal exercise (25). It has to be pointed out that in calculating iEMG at MVC, the highest mean value from seconds 2 until 4 of a 5-s isometric knee-extension (RF, VL) or knee-flexion (BF) contraction was used. This value was used to normalize iEMG during a 200-m segment. Because cycling is a cyclic movement, bursts of high EMG activity repeatedly occurred, alternating with zero-activity intervals. In attaining iEMG during cycling during 200-m segments, the zero activity is also included in the calculations, whereas the value for iEMG at MVC is calculated during the burst of activity only. Accordingly, iEMG during cycling will therefore not reach the maximal MVC values. Furthermore, iEMG at MVC is attained during isometric exercise. Maximal iEMG during cycling in a particular muscle might differ from maximal iEMG during isometric exercise. Hunter et al. (13) showed that iEMG for MVC was significantly greater than iEMG during a dynamic single maximal revolution of a cycle pedal, which reaches a value of about 75% of MVC. This might be explained by the fact that muscle coordination and timing are very important in dynamic tasks (16). For a dynamic movement such as cycling, coordination is therefore important. It might not be favorable to recruit all possible fibers maximally and for the same duration as during the MVC. Therefore, conclusions about the causes of fatigue based on comparing maximal iEMG during dynamic exercise with iEMG attained during isometric MVC have to be interpreted with caution.
We were also interested in the contribution of aerobic and anaerobic energy systems to total PO because pacing strategy refers to the variation of speed over the race by regulating the rate of energy expenditure (19) and thereby the pattern of PO. The different pacing strategies performed in this study were defined by PO pattern. In the SUB trial, PO increased significantly from the first half of the trial to the second half; in the EVEN trial, PO remained constant; and in the SUPRA trial, PO decreased significantly. PO was subdivided into aerobic and anaerobic power. For Paer, a main effect was found when comparing the first and second interval of the trial, which showed that aerobic contribution increased toward the end of the race, independent of strategy. This seems to indicate that the aerobic energy contribution in supramaximal exercise is not a defining variable of pacing strategy. No matter what strategy is chosen, even if PO decreases significantly, Paer increases significantly. Also, a main effect for HR was found, which showed that HR increased significantly in all strategies. Because athletes are operating close to V̇O2max velocity in time trial exercise, they do not seem able to vary aerobic energy contribution much. Pan, on the other hand, seems to be the more important variable in pacing. Pan increases significantly from the first to last 2000 m in SUB, as does PO. In SUPRA, Pan decreases significantly, as does PO. In EVEN, Pan and PO do not differ significantly between intervals. Although not significant, the somewhat higher value in the first interval for Pan in the EVEN strategy can be expected, because to establish a constant power output profile, the increasing Paer has to be compensated. Additionally, in the first part of the race (the first 500 m), anaerobic energy contribution is relatively high because anaerobic energy is immediately available, whereas aerobic energy production needs some time to increase. Noticing the comparable profiles of Pan and PO, it seems that the different pacing strategies, as already suggested by de Koning et al. (19), are mainly regulated by variation in anaerobic energy expenditure. In accordance with this finding, a terminal acceleration in self-paced time trials is often found, accompanied by an increase in anaerobic energy production (8,9). This increase in PO at the end of exercise has also been found by Ansley et al. (1) in self-paced time trials, accompanied by an increase in iEMG. Because the amount of anaerobic energy that can be generated during a 4000-m time trial is limited (22), it is important to distribute this amount optimally over the race.
In calculating aerobic and anaerobic power, it is assumed that efficiency measured at submaximal intensities does not change compared with at higher intensities. For the profile of energy distribution, we do not expect this to have a big impact on the outcome. Further, no significant differences were found for efficiency measured before each time trial per strategy, so differences between strategies are not caused by efficiency measurements. It should be kept in mind that efficiency might have a slightly differing value at higher intensities, which might result in an underestimation of aerobic power and an overestimation of anaerobic power.
Also, no significant differences were found between final times per strategy. Although not significant, it can be seen that the SUPRA strategy has the fastest final time and SUB the slowest. Additionally, for PO, a main effect of strategy was found. A pairwise comparison showed that PO in SUB was lower than in SUPRA. This could mean that in an aggressive strategy (SUPRA), subjects are able to win so much in the first interval of the race that it is impossible to make up for this in the second interval with a conservative (SUB) strategy. For PO, strategy thus affects energy distribution, but can also have an influence on the total amount of energy that can be generated during the race. For BLC, a pairwise comparison clearly revealed that the greatest (although not significant) difference can be found between SUB and SUPRA. This, combined with an absence of a decrease in PO in SUB, indicates that in SUB, peripheral mechanisms interfere less with performance than in SUPRA, where they even evoke a decrease in PO. An effect of strategy was also found for Paer, EMG (VL, BF), RPM, and RPE (mean values over the whole trial). Thus, strategy influences these variables, which should be taken into account in pacing. It is therefore very important to gain more insight into fatigue.
Our findings suggest that reaching unsustainable metabolic disturbances during time trial performance might be prevented by peripheral regulation. This does not mean that central factors are not playing a role in pacing. Pacing is about regulating speed by varying energy expenditure and, thereby, PO pattern (19). Successful athletes are skilled in this monitoring process and prevent a significant slowdown late in the event by regulation of their early pace (e.g., PO) (20). Pacing is about managing energy expenditure with its resulting peripheral fatigue so that no factor will be limiting before the end of the trial. Otherwise, performance will be suboptimal. Although homeostasis is not endangered during a time trial, as can also be seen in the variety of possible pacing strategies, energy must be distributed during the race to attain optimal performance. This is regulated centrally by regulation of muscle activation, as is also mentioned by Rauch et al. (26). Tucker et al. (30) showed that during self-paced time trials in the heat, reduced PO and EMG activity compared with normal conditions occurred before any abnormal increase in rectal temperature was shown. The subjects, apparently anticipating these circumstances, avoided reaching limiting factors and distributed energy resources optimally. Rauch et al. (26) found that carbohydrate-loaded subjects had a higher mean power output during exercise than subjects who were not carbohydrate loaded. It was suggested that the brain was able to anticipate the rate of muscle glycogen concentration according to an internal metabolic calculation based on an individual's own critical level end-point muscle glycogen ("glycostat"). Carbohydrate-loaded subjects had more energy to distribute, and thus could afford an increase in skeletal muscle recruitment compared with those who were not.
As mentioned, we found that anaerobic energy resources seem to be of importance in regulating pacing strategy. This corresponds to Foster et al. (8), whose data showed that athletes monitor some aspect of anaerobic energy expenditure during high-intensity exercise so that near zero values are not reached until the finish line is approached. When instructed to finish as fast as possible, athletes apparently negotiate regarding an estimation of the task remaining, momentary PO, and remaining anaerobic reserves (9). This monitoring process to optimize distribution of energetic resources is consistent with the governor hypothesis, but not reaching near-zero values of anaerobic energy expenditure can also be seen as support for a metabolite accumulation or phosphagen-depletion model of fatigue (8). Possible mechanisms in regulating pacing strategy should be sought in the anaerobic energy production system.
Lastly, the fact that pacing is possible during a time trial makes riding a time trial different from exercise till exhaustion, in which pacing is not an option because a fixed intensity has to be attained. In time trial exercise, limiting factors can be avoided or postponed by proper pacing. Central fatigue does not necessarily occur. However, an athlete might fail to select a proper pacing strategy, for example, in unexpected circumstances. In such a case, peripheral fatigue is not well managed, and downregulation of EMG as a protective mechanism may reasonably be expected.
In time trials with different first-half strategies imposed, no reduction in iEMG occurs with decreasing power output. Centrally mediated downregulation of mechanical power output does not necessarily accompany fatigue in middle-distance time trial exercise. However, to maximize performance by varying energy expenditure, central neural control is suggested. Specifically, anaerobic energy resources seem to be important in regulating pacing strategy.
This study was supported by The Netherlands Organization for Applied Scientific Research TNO and the NOC*NSF.
1. Ansley, L., E. Schabort, A. St. Clair Gibson, M. I. Lambert, and T. D. Noakes. Neural regulation of pacing strategies during 4 km time trials. Med. Sci. Sports Exerc.
2. Bangsbo, J., T. E. Graham, B. Kiens, and B. Saltin. Elevated muscle glycogen and anaerobic energy production during exhaustive exercise in man. J. Physiol.
4. Borg, G. A. V. Psychophysical bases of perceived exertion. Med. Sci. Sports Exerc.
5. Ebenbichler, G., J. Kollmitzer, M. Quittan, F. Uhl, C. Kirtley, and V. Fialka. EMG fatigue patterns accompanying isometric fatiguing knee-extensions are different in mono-and bi-articular muscles. Electroencephal. Clin. Neurophysiol.
6. Foster, C., M. Schrager, A. C. Snijder, and N. N. Thompson. Pacing strategy and athletic performance. Sports Med.
7. Foster, C., A. C. Snijder, N. N. Thompson, M. A. Green, M. Foley, and M. Schrager. Effect of pacing strategy on cycle time performance. Med. Sci. Sports Exerc.
8. Foster, C., J. de Koning, F. Hettinga, J. Lampen, et al. Pattern of energy expenditure during simulated competition. Med. Sci. Sports Exerc.
9. Foster, C., J. de Koning, F. Hettinga, et al. Effect of competitive distance on energy expenditure during simulated competition. Int. J. Sports Med.
10. 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.
11. Gerdle, B., S. Karlsson, A. G. Crenshaw, and J. Friden. The relationships between EMG and muscle morphology throughout sustained static knee extension at two submaximal force levels. Acta Physiol. Scand.
12. Hawley, J. A., and T. Reilly. Fatigue revisited. J. Sports Sci.
13. Hunter, A. M., A. St. Clair Gibson, M. I. Lambert, and T. D. Noakes. Electromyographic (EMG) normalization method for cycle fatigue protocols. Med. Sci. Sports Exerc.
14. Hunter, A. M., A. St. Clair Gibson, M. Lambert, et al. EMG Amplitude in maximal and submaximal exercise is dependent on signal capture rate. Int. J. Sports Med.
15. van Ingen Schenau, G. J., J. J. de Koning, and G. de Groot. The distribution of anaerobic energy in 1000 and 4000 meter cycling bouts. Int. J. Sports Med.
16. van Ingen Schenau, G. J., W. M. M. Drossers, T. G. Welter, A. Beelen, G. de Groot, and R. Jacobs. The control of mono-articular muscles in multi-joint leg extension in man. J. Physiol.
17. Jorge, M., and M. L. Hull. Analysis of EMG measurements during bicycle pedalling. J. Biomechanics
18. 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.
19. de Koning, J. J., M. F. Bobbert, and C. Foster. Determination of optimal pacing strategy in track cycling with an energy flow model. J. Sci. Med. Sport
20. 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.
21. Lambert, E. V., A. St. Clair Gibson, and T. D. Noakes. Complex systems model of fatigue: integrative homeostatic control of peripheral physiological systems during exercise in humans. Br. J. Sports Med.
22. Medbø, J. I., and I. Tabata. Anaerobic energy release in working muscle during 30s to 3 min of exhausting bicycling. J. Appl. Physiol.
23. Noakes, T. D. Challenging beliefs: ex Africa simper aliquid novi. Med. Sci. Sports Exerc.
24. Noakes, T. D. Maximal oxygen uptake: "classical" versus "contemporary" viewpoints: a rebuttal. Med. Sci. Sports Exerc.
25. Noakes, T. D., A. St. Clair Gibson, and E. V. Lambert. 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.
26. 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.
27. St. Clair Gibson, A., M. I. Lambert, and T. D Noakes. Neural control of force output during maximal and submaximal exercise. Sports Med.
28. St. Clair Gibson, A., E. J. Schabort, and T. D. Noakes. Reduced neuromuscular activity and force generation during prolonged cycling. Am. J. Physiol.
29. Taylor, A. D., R. Brooks, 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.
30. Tucker, R., L. Rauch, Y. X. Harley, and T. D. Noakes. Impaired exercise performance in the heat is associated with an anticipatory reduction in skeletal muscle recruitment. Pflugers Arch.