Secondary Logo

Reliability of Power Output During Eccentric Sprint Cycling

Brughelli, Matt1,2; Van Leemputte, Marc1

Journal of Strength and Conditioning Research: January 2013 - Volume 27 - Issue 1 - p 76–82
doi: 10.1519/JSC.0b013e31824f2055
Original Research

Brughelli, M and Van Leemputte, M. Reliability of power output during eccentric sprint cycling. J Strength Cond Res 27(1): 76–82, 2013—The purpose of this study was to determine the reliability of power outputs during maximal intensity eccentric cycling over short durations (i.e., eccentric sprint cycling) on a “motor-driven” isokinetic ergometer. Fourteen physically active male subjects performed isokinetic eccentric cycling sprints at 40, 60, 80, 100, and 120 revolutions per minute (rpm) on 4 separate occasions (T1–T4). Each sprint lasted for 6 seconds, and absolute measures of mean power (MP) and peak power (PP) per revolution were recorded. Significant increases in MP and PP were observed between T1 and subsequent trials, but no significant differences were identified between T2, T3, and T4. The coefficient of variation (CV) and intraclass correlation coefficient (ICC) were calculated to reflect within-subject and between-session reliability of MP and PP at each cadence. The CV improved to below 10% for cadences of 60, 80, 100, and 120 rpm between T3 and T4, and the majority of ICC values improved to above 0.90. The remaining ICC values remained in the moderate range between T3 and T4 (i.e., 0.82–0.89). Coefficient of variation and ICC values for the 40 rpm cadence remained at unacceptable levels throughout the 4 trials and thus should be avoided in future investigations. The results of this study indicate that reliable power outputs may be obtained after 2 familiarization sessions during eccentric sprint cycling at cadences ranging from 60 to 120 rpm.

1Department of Biomedical Kinesiology, Faculty of Kinesiology and Rehabilitation Sciences, KU Leuven, Leuven, Belgium

2Sports Performance Research Institute New Zealand, AUT University, Auckland, New Zealand

Address correspondence to Marc Van Leemputte,

Back to Top | Article Outline


It is well known that skeletal muscles are capable of producing greater magnitudes of force during active lengthening (i.e., eccentric contractions) than active shortening (i.e., concentric contractions), especially at higher velocities. This capacity has been described as a potential stimulus for increasing muscle size, strength, and power (19,20). Consequently, several investigations have used chronic eccentric exercise and have reported beneficial adaptations in a variety of populations and settings (3,7,8,12).

Eccentric cycling ergometry provides an ideal model for investigating eccentric muscle contractions during a multijoint and multi-degree-of-freedom movement (25). Eccentric cycling ergometers use a motor that drives the cranks in the reverse direction over a range of cadences. Recently, Elmer et al. (10) investigated the contribution of each joint during submaximal eccentric cycling in the recumbent position at 60 revolutions per minute (rpm). As expected, the knee joint absorbed the majority of the negative power (i.e., 58% of 256 W), whereas the hip and ankle joints absorbed 29% and 10%, respectively. Thus, it is not surprising that chronic, submaximal, eccentric cycling has been shown to enhance quadriceps and gluteal size and strength (9,18,24). Even more impressive, these improvements have been reported with minimal muscle damage and soreness as the load and intensity were minimal at the beginning of training and progressively increased throughout the training periods (i.e., up to 12 weeks). The major limitation with this method is the lack of motivation of the subjects to perform up to 36 training sessions over a 12-week period, which may not be realistic in most sport and clinical settings. Motor-driven ergometers can also be adapted to allow for maximal intensity eccentric cycling for durations of less than 10 seconds (i.e., eccentric sprint cycling). Such an approach could allow for direct comparisons of torque and power during maximal eccentric and concentric cycling over a range of cadences and could lead to the development of effective and efficient training protocols.

Several articles have reported a learning effect in maximal power development after the first testing sessions during normal concentric cycling (1,6,21,23). Such a learning effect could confound results and lead to invalid interpretations of independent variables. Logically, valid evaluation of torque and power, pre- and posttraining, requires learning to be completed before experimental data collection. Thus, the purposes of the present study were to (a) determine if a learning effect occurred over 4 separate testing sessions during eccentric sprint cycling and (b) to determine how many familiarization sessions are needed to obtain reliable power outputs (i.e., within-subject reliability and between-day reliability) over a range of cadences. Such information is vital for the development of future acute and chronic experiments involving eccentric sprint cycling.

Back to Top | Article Outline


Experimental Approach to the Problem

A group of recreationally trained men performed eccentric sprint cycling on 4 separate occasions/trials (T1–T4). The phenomenon of a learning effect was identified as a significant difference (P > 0.05) in power output between subsequent trials. Within-subject and between-day reliability of power output were determined during T1–T2, T2–T3, and T3–T4. The number of familiarization sessions needed to obtain reliable power output was determined when the majority of coefficient of variation (CV) and intraclass correlation coefficient (ICC) values improved to less than 10% and greater than 0.90, respectively.

Back to Top | Article Outline


Fourteen healthy male subjects were recruited to participate in this study. None of the subjects had any prior experience with eccentric sprint cycling. All the subjects were university students and participated in multiple sports. The mean ± SD values for age, height, and mass were 22.3 ± 2.5 years, 181.5 ± 12.1 cm, and 83.3 ± 6.3 kg. The Human Research Ethics Committee of KU Leuven approved all procedures before commencing the study. All subjects provided their written informed consent before participating in the study.

Back to Top | Article Outline


Each subject performed multiple 6-second eccentric sprints on an isokinetic cycling ergometer on 4 separate occasions. The testing sessions were separated by a minimum of 2 days and a maximum of 5 days. All testing sessions were completed within 3 weeks for each subject. Efforts were made to ensure that all testing sessions were performed at the same time of day and that no strenuous physical activity was performed within 24 hours of testing. Each testing session began with a 5-minute warm-up of concentric cycling at 80 revolutions per minute (rpm) while maintaining approximately 80 W. The warm-up also included 2 concentric sprints at 80 rpm for 3 seconds. After the warm-up, the subject had 2 minutes to rest before commencing the maximal eccentric sprints. Each subject subsequently performed 6-second maximal eccentric sprints at 40, 60, 80, 100, and 120 rpm in random order. For each sprint, the cadence was increased from zero to the predetermined testing cadence. Once the testing cadence was reached, a countdown (i.e., “ready,” “set”) was followed by a load “Go” command. The subjects received strong verbal encouragement throughout the 6 seconds of testing, followed by clear instructions of when to stop. Data were collected throughout each 6-second eccentric sprint. The testing sessions ended with a 5-minute cooldown of concentric cycling at 80 rpm while maintaining approximately 80 W.

The subjects were given clear instructions to avoid backward or reverse direction cycling. They were instructed to perform eccentric cycling by resisting the pedals from the bottom vertical position until the top vertical position as the cranks rotated in the reverse direction. Several previous studies have investigated the acute and chronic effects of eccentric cycling at low–moderate intensities for prolonged durations dating back to 1952 (2). The novelty of the present study is that eccentric cycling was performed at maximum intensities for short durations, which we have termed “eccentric sprint cycling.”

Due to the inherent risks of performing eccentric sprint cycling on an isokinetic ergometer, an emergency switch was implemented so that the tester could immediately terminate any sprint at any time. There were no emergencies.

Back to Top | Article Outline

Isokinetic Ergometer

All maximum eccentric sprints were performed on a motor-driven isokinetic erogmeter as described by Koninckx et al. (15,16). A general race bicycle was mounted on a self-constructed isokinetic ergometer. The rear fork, chain, and derailleur of the bicycle were fixed onto a motor-driven flywheel body and hub (Shimano; Dura-Ace-FH-7700, Sakai City, Japan). The system was adapted to allow the flywheel to be driven in the reverse direction by a servo-controlled motor (SEW-Eurodrive, Bruchsal, Germany). A timing belt and 2 pulleys synchronized the movement of the cranks with the controlled movement of the motor axle. Torque was measured continuously by a torque transducer (Lebow, Troy, MI, USA) mounted on the outgoing axle of the motor. The position of the cranks was monitored with a sensor (Assemtech, Essex, United Kingdom) on the ergometer frame, and a magnet was attached to the left crank. The “zero angle” was identified as the left cranks reached the vertical downward position during the full revolution. Thus, the crank was continuously linked to the angle of the motor axle. Furthermore, a dSpace interface (ControlDesk; dSpace GmBH, Paderborn, Germany) was used for data streaming with a real-time control program. All subjects wore cycling shoes that could be clipped into the pedal surface (Look Keo-Sprint, Nevers, France) during each testing session. During each 6-second eccentric sprint, crank torque was measured continuously at 1,000 Hz. Mean torque (MT) was calculated for each complete revolution at each cadence. Afterward, only the revolutions that had a value of greater than 90% of the highest MT value (i.e., per revolution) were included in the data analysis. Mean torque was then multiplied by the preset cadence to obtain mean power (MP) per revolution. Likewise, peak torque was calculated and multiplied by the preset cadence to obtain peak power (PP). All data were exported to a personal computer and analyzed with a custom-made Matlab program (The Mathworks, Inc., Natick, MA, USA).

Back to Top | Article Outline

Statistical Analyses

Statistical analysis was carried out using a statistical software program Statistica 7 (Statsoft, Tulsa, OK, USA). Descriptive statistics have been expressed as mean ± SD. Systematic changes in the performance variables (i.e., MP and PP) over the 4 trials were determined with a repeated measures analysis of variance. When a significant difference was identified (p ≤ 0.05), a least significant difference post hoc was used to determine where the difference occurred.

To further examine reliability of the performance variables over the 4 trails, the data were log transformed (13) and thereafter used to determine typical error of measurement and ICC expressed as 95% confidence intervals. The ICC was used to determine between-day reliability (i.e., repeatability). It is generally considered that reliability is high for ICCs above 0.90, moderate for values between 0.80 and 0.89, and questionable for values below 0.80 (27). The typical error of measurement, expressed as CV, was used to determine within-subject reliability using the following formula:

CV = 100(eSD/100 − 1).

We considered a CV of <10% for good and 10–15% as acceptable reliability (11,26).

Back to Top | Article Outline


Overall mean ± SD values for MP and PP at each cadence are shown in Figure 1. From 40 to 100 rpm, MP and PP significantly increased 15.5–37.5% between T1 and T2, respectively (P < 0.05). At 120 rpm, MP and PP significantly increased 14.2% and 12.9% between T1 and T3, respectively (p < 0.05). Furthermore, there were no significant differences in MP or PP between T2, T3, and T4 for any of the cadences.

Figure 1

Figure 1

Coefficient of variation and ICC values between trials are shown in Tables 1 and 2. Coefficient of variation values improved to less than 10% between T3 and T4 for MP and PP between the cadences of 60 and 120 rpm, respectively. At 80 rpm, CV values also improved to below 10% from T2 and T3 for MP and PP. At 40 rpm, CV values did not improve through the subsequent trials. However, they remained in the acceptable range of 10.8 and 16.4% throughout the 4 trials. Intraclass correlation coefficient values improved to 0.90 or above for MP at cadences of 80–120 rpm, and for PP at cadences of 80 and 100 rpm between T3 and T4. All ICC values remained in the moderate range between T3 and T4. For 80 rpm, ICC values improved to above 0.90 from T2 to T3. Intraclass correlation coefficient values never reached above 0.90 at the slowest cadence of 40 rpm but never dropped below 0.80 throughout the 4 trials.

Table 1

Table 1

Table 2

Table 2

Back to Top | Article Outline


To our knowledge, this was the first study to investigate reliability of power outputs during eccentric sprint cycling. The first main finding of this study was that a significant learning effect occurred for PP and MP after the first testing session. The learning effect was the most prominent for the moderate cadences (i.e., 60, 80, and 100 rpm). The second main finding was that moderate to high reliability for MP and PP was obtained after 2 familiarization sessions during eccentric sprint cycling at cadences ranging from 60 to 120 rpm.

Learning effects are inevitable with any novel performance test or activity. The systematic changes in performance are often due to the exposure and experience of previous trials (14). Knowledge of a learning effect is vital for interpreting values before and after an intervention. Thus, the effects of an intervention should only be considered after appropriate learning has occurred. Previous studies have reported learning effects (i.e., MP and PP) between the first 2 trails during maximal intensity “concentric” cycling for short durations (<8 seconds). Martin et al. (21) reported a significant increase in maximum power in trained and nontrained cyclists after a single trail. The subjects performed 3- to 4-second maximal cycling sprints on an inertial-loaded cycling ergometer from a static starting position. Maximum power was defined as the apex of the MP–cadence curve. In a more recent study, Mendez-Villanueva et al. (23) measured both MP and PP during maximal intensity cycling sprints (6-second duration) from a static position on a front-access cycle ergometer. It was reported that mean values of MP and PP significantly increased between T1 and T2 by 4.9 and 5.6%. Both studies also reported stabilized power outputs between T2, T3, and T4. Thus, the learning effect was completed after a single trail. Similar findings have also been reported in studies using repeated maximal intensity concentric cycling bouts. Capriotti et al. (6) reported that MP significantly increased by approximately 11% during repeated maximum cycling (i.e., 10 sets of 7 seconds) between T1 and T2; afterward, MP values stabilized. The subjects performed a fatiguing protocol with a heavily loaded ergometer (i.e., 11.34 kg). McGawley and Bishop (22) also used a repeated maximal cycling protocol (5 sets of 6 seconds) with a wind-braked front-access cycling ergometer and reported significant increases of 5.2% between T1 and T2. The present study found that MP and PP significantly increased between T1 and T2 (15.5–37.5%) at cadences 40 to 100 rpm. After the first trial, MP and PP remained relatively constant. At 120 rpm, MP and PP significantly increased between T1 and T3 (14.2 and 12.9%) and afterward stabilized. The greatest learning occurred between the moderate cadences of 60, 80, and 100 rpm (18.9–37.5%). These learning effects were much greater in percent change in comparison with the previously mentioned studies using concentric sprint cycling (i.e., 4.9 to approximately 11%) (6,21,23). It is well known that a phenomenon occurs during maximal eccentric contractions in humans and animals called “neural inhibition” (4,5). Neural inhibition is a reduction in muscle activating and force production during maximum eccentric contractions, which is thought to be a mechanism to prevent the muscle-tendon unit from being injured. Neural inhibition has not been shown during concentric or isometric contractions. It is possible that neural inhibition reduced MP and PP during T1 in the present study. Logically, it could also be suggested that the increase from T1 to T2 was due to a learning effect that reduced neural inhibition.

Very few studies have assessed the reliability of maximal isokinetic cycling for short durations and over a range of cadences. In a test-retest study, Koninckx et al. (15) reported ICC between 0.94 and 0.98 for PP when measured over 6 cadences (i.e., 40–140 rpm) in trained cyclists. Each sprint was performed isokinetically for 5 seconds. Additionally, power outputs on the isokinetic ergometer showed a relative difference of less than 1.6% in comparison with a calibrated SRM system (SRM GmbH, Julich, Germany) mounted on the isokinetic ergometer. The comparison occurred during submaximal concentric cycling over a range of cadences (i.e., 80–120 rpm) and intensities (i.e., 200–400 W). Lands et al. (17) assessed torque and power outputs on an isokinetic ergomteter at cadences of 60, 90, and 120 rpm during sprints of 10 and 30 seconds. The subjects included healthy men and women. It was reported that CV values during the test-retest protocol were between 5.2 and 9.1%. Although the results are comparable with the present study, these studies did not investigate the influence of familiarization on reliability of MP or PP during subsequent trials.

As mentioned previously, a great learning effect occurred for MP and PP between T1 and T2. A straightforward analysis of these findings would imply that reliable values could be obtained after a single familiarization session. However, another key finding of the present study was that reliability (reflected by CV and ICC values) of MP and PP did not improve until after 2 familiarization sessions for the majority of cadences tested. The present study found CVs between 4.6 and 8.6% after 2 familiarization sessions at cadences of 60–120 rpm. This finding is similar to previous studies using short-duration concentric spring cycling (i.e., <10 seconds) over separate trials. Mendez-Villanueva et al. (23) reported values of 1.8% (MP) and 2.5% (PP) for PP and MP during 6-second maximal cycling sprints after 2 familiarization sessions. McGawley and Bishop (22) reported a similar value of 3.1% for PP during the initial 6-second sprint of a repeated sprint protocol. Studies using repeated sprint and incremental sprint cycling protocols have reported similar findings for MP and PP after 2 familiarization sessions (CV = 2.4 to 3.7%). Thus, it could be concluded that a minimum of 2 familiarization sessions are needed for good within-subject reliability during eccentric and concentric modes of sprint cycling. In regard to between-session reliability, ICC values improved in the present study to 0.89–0.96 after 2 familiarization sessions at 60–120 rpm. McGawley and Bishop (22) reported a similar ICC value of 0.86 for PP after 2 familiarization sessions. These ICC values provide further evidence that reliability improved after 2 familiarization sessions. Due to a loss of motivation after a series of subsequent trials, it has been suggested that excessive familiarization should also be avoided (14). Too many familiarization sessions could decrease performance and increase variability. Thus, the results of the present study suggest that 2 familiarization sessions of eccentric sprint cycling are optimal for learning, familiarization, and reliability.

It should be noted that only 1 familiarization session was needed to obtain reliable MP and PP at the moderate cadence of 80 rpm. This cadence provided the most reliable data with the fewest familiarization sessions. Because the previous studies measured MP and PP over a single continuous bout, and the present study measured MP and PP over at separate cadences over separate bouts, comparisons cannot be made about the various cadences. It is possible that this moderate cadence was the most comfortable and thus resulted in the better control of movement patterns and muscular coordination (i.e., intra- and intermuscular coordination). It should also be noted that CVs at the slowest cadence of 40 rpm did not improve throughout the subsequent trials. It is likely that 40 rpm is too slow of a cadence and thus too difficult to coordinate movement patters, for obtaining high absolute reliability. However, reliability did remain in the moderate and acceptable range over the 4 trials at 40 rpm (CV = 10.9–16.4%; ICC = 0.83–0.89).

In conclusion, the results of this study indicate that a learning effect occurs (i.e., power outputs) after a single trial of eccentric sprint cycling. This learning effect is most pronounced at the moderate cadences. The results of this study also provide evidence that 2 familiarization sessions are appropriate to obtain reliable power outputs during eccentric sprint cycling at moderate to high cadences (i.e., 60–120 rpm). At the slowest cadence of 40 rpm, reliability did not improve with subsequent trials and thus should be avoided.

Back to Top | Article Outline

Practical Applications

Eccentric muscle contractions are capable of producing greater magnitudes of force with less metabolic effort in comparison with concentric muscle contractions. Consequently, chronic eccentric exercise has been effectively used to induce gains in muscular strength and size with less overall voluntary effort. Logically, eccentric sprint cycling could prove to be a useful form of resistance training for athletes attempting to maintain or increase strength but avoid overtraining. The results of the present study suggest that power output during eccentric sprint cycling can be reliably measured after 2 familiarization sessions. Thus, future investigation with eccentric spring cycling is warranted. For instance, very little is currently known about the power-force-velocity relation during multijoint and multi-degree-of-freedom movements that involve eccentric muscle contractions. The great majority of research has focused on this relation during single-joint movements at the knee, elbow, and shoulder joints. A comparison of power output during eccentric and concentric sprint cycling could provide valuable information and insights into developing future training protocols. Based on the findings of the present study, it appears that the effects of chronic eccentric sprint cycling on power adaptations and injury prevention may be examined as long as sufficient test familiarization has been incorporated to establish a realistic baseline of performance.

Back to Top | Article Outline


1. Abbiss CR, Levin G, McGuigan MR, Laursen PB. Reliability of power output during dynamic cycling. Int J Sports Med 29: 574–578, 2008.
2. Abbott BC, Bigland R, Ritchie JM. The physiological cost of negative work. J Physiol 117: 380–390, 1952.
3. Askling C, Karlsson J, Thorstensson A. Hamstring injury occurrence in elite soccer players after preseason strength training with eccentric overload. Scand J Med Sci Sports 13: 244–250, 2003.
4. Babault N, Pousson M, Ballay Y, Van Hoecke J. Activation of human quadriceps femoris during isometric, concentric, and eccentric contractions. J Appl Physiol 91: 2628–2634, 2001.
5. Beltman JG, Sargeant AJ, van Mechelen W, de Haan A. Voluntary activation level and muscle fiber recruitment of human quadriceps during lengthening contractions. J Appl Physiol 97: 619–626, 2004.
6. Capriotti PV, Sherman WM, Lamb DR. Reliability of power output during intermittent high-intensity cycling. Med Sci Sports Exerc 31: 913–915, 1999.
7. Clark R, Bryant A, Culgan J. The effects of hamstring strength training on dynamic jump performance and isokinetic strength parameters. Phys Ther Sport 6: 67–73, 2005.
8. Dibble LE, Hale TF, Marcus RL, Droge J, Gerber JP, Lastayo PC. High intensity resistance training amplifies muscle hypertrophy and functional gains in persons with Parkinson's disease. Mov Disord 21: 1444–1452, 2006.
9. Elmer S, Hahn S, McAllister P, Leong C, Martin J. Improvements in multi-joint leg function following chronic eccentric exercise. Scand J Med Sci Sports. 2011, Mar 16, Epub ahead of print.
10. Elmer SJ, Madigan ML, LaStayo PC, Martin JC. Joint-specific power absorption during eccentric cycling. Clin Biomech (Bristol, Avon) 25: 154–158, 2010.
11. Fredberg U, Bolvig L, Andersen NT, Stengaard-Pedersen K. Ultrasonography in evaluation of Achilles and patella tendon thickness. Ultraschall Med 29: 60–65, 2008.
12. Gerber JP, Marcus RL, Dibble LE, Greis PE, Burks RT, Lastayo PC. Safety, feasibility, and efficacy of negative work exercise via eccentric muscle activity following anterior cruciate ligament reconstruction. J Orthop Sports Phys Ther 37: 10–18, 2007.
13. Hopkins WG. A new view of statistics. Available at: Accessed March 5, 2011.
14. Hopkins WG, Schabort EJ, Hawley JA. Reliability of power in physical performance tests. Sports Med 31: 211–234, 2001.
15. Koninckx E, Van Leemputte M, Hespel P. Effect of isokinetic cycling versus weight training on maximal power output and endurance performance in cycling. Eur J Appl Physiol 109: 699–708, 2010.
16. Koninckx E, van Leemputte M, Hespel P. Effect of a novel pedal design on maximal power output and mechanical efficiency in well-trained cyclists. J Sports Sci 26: 1015–1023, 2008.
17. Lands LC, Hornby L, Desrochers G, Iler T, Heigenhauser GJ. A simple isokinetic cycle for measurement of leg muscle function. J Appl Physiol 77: 2506–2510, 1994.
18. Lastayo PC, Ewy GA, Pierotti DD, Johns RK, Lindstedt S. The positive effects of negative work: Increased muscle strength and decreased fall risk in a frail elderly population. J Gerontol Biol Sci Med Sci 58: M419–M424, 2003.
19. LaStayo PC, Woolf JM, Lewek MD, Snyder-Mackler L, Reich T, Lindstedt SL. Eccentric muscle contractions: Their contribution to injury, prevention, rehabilitation, and sport. J Orthop Sports Phys Ther 33: 557–571, 2003.
20. Lindstedt SL, Lastayo PC, Reich TE. When active muscles lengthen: properties and consequences of eccentric contractions. News Physiol Sci 16: 256–261, 2001.
21. Martin JC, Diedrich E, Coyle F. Time course of learning to produce maximum cycling power. Int J Sports Med 21: 485–487, 2000.
22. McGawley K, Bishop D. Reliability of a 5 x 6-s maximal cycling repeated-sprint test in trained female team-sport athletes. Eur J Appl Physiol 98: 383–393, 2006.
23. Mendez-Villanueva A, Bishop D, Hamer P. Reproducibility of a 6-s maximal cycling sprint test. J Sci Med Sport 10: 323–326, 2007.
24. Meyer K, Steiner R, Lastayo P, Lippuner K, Allemann Y, Eberli F, Schmid J, Saner H, Hoppeler H. Eccentric exercise in coronary patients: Central hemodynamic and metabolic responses. Med Sci Sports Exerc 35: 1076–1082, 2003.
25. Perrey S, Betik A, Candau R, Rouillon JD, Hughson RL. Comparison of oxygen uptake kinetics during concentric and eccentric cycle exercise. J Appl Physiol 91: 2135–2142, 2001.
26. Stokes M. Reliability and repeatability of methods for measuring muscle in physiotherapy. Physiother Pract 1: 71–76, 1985.
27. Vincent WJ. Statistics in Kinesiology, 2nd ed. Champaign, IL: Human Kinetics, 1999.

active lengthening; isokinetic ergometer; familiarization; reproducibility; typical error of measurement

© 2013 National Strength and Conditioning Association