Volume 21, Issue 1 , Pages 18-24, February 2011
The effect of cadence on timing of muscle activation and mechanical output in cycling: On the activation dynamics hypothesis
Article Outline
Abstract
The purpose of this study was to examine the activation dynamics hypothesis, which states that, in cycling, the pattern between muscle activity and crank position shifts in regard to its angle in the crank cycle with increasing cadence to maintain invariant positioning of the mechanical output. We measured surface EMG of six muscles, and by means of force measurements at the crank and inverse dynamics calculated hip, knee, and ankle joint dynamics during cycling at five cadences (60–100
rpm) at 75% of maximal power in trained cyclists. The joint dynamics (net muscle moment and power) showed a consistent positive phase shift with increasing cadence. The phase shift in muscle activation patterns was highly variable amongst subjects and was, on average, close to zero. Our results are in contradiction with the activation dynamics hypothesis.
Keywords: Cycling, Cadence, Mechanics, Electromechanical delay, Relative phase, EMG
1. Introduction
In cycling, timing of neuromuscular activation seems critical for the execution of an effective pedaling movement. Only the production of a force perpendicular to the crank, i.e., creating a crank torque, is effective in propulsion. At the same time, it seems impossible for a cyclist to generate such a perpendicular force, at least of considerable magnitude, at all times during the crank cycle (Gruben et al., 2003). A particular timing of force and torque production during the crank cycle is rather evident, with the maximal torque produced around the middle of the down stroke (e.g., Ettema et al., 2009, Kautz and Hull, 1993, Sarre et al., 2004). Marsh and Martin (1995) discovered that the peak of neuromuscular activity for a selection of lower extremity muscles occurred progressively earlier with increases in cadence for all the muscles investigated except one. They suggested that this phenomenon occurred “in order to contribute peak force at the same position in the crank cycle”. The rationale behind this idea is that the crank position, by affecting muscle length, affects the power and torque production capabilities of muscle. As stated by Neptune et al. (1997), onset of muscle activation must occur earlier in the crank cycle with increasing cadence to allow for electromechanical delay (EMD). Neptune et al. (1997) aptly named the idea the activation dynamics hypothesis. EMD can be defined as the time between neuromuscular activation, the electrical outcome, and force, the mechanical outcome (Li and Baum, 2004, Sarre and Lepers, 2005), and is regarded as being fairly constant (Li and Baum, 2004, Sarre and Lepers, 2007). Ingen Schenau et al. (1995) found EMD to approximate 90ms in most of the leg muscles during cycling, as did Vos et al. (1991). Even though EMD values may be different for individual muscles, one should be able to observe time delays between neuromuscular activation and resulting force close to EMD values previously reported for the rationale behind the activation dynamics hypothesis to be considered valid. Neptune et al. (1997) found that six of the eight muscles they investigated displayed neuromuscular activity shifting earlier in the crank cycle with increased cadence. Baum and Li, 2003, Bieuzen et al., 2007, Li and Baum, 2004, and Sarre and Lepers, 2005, Sarre and Lepers, 2007 have all corroborated their results. The activation dynamics hypothesis is not undisputed, though. In contrast to the findings presented above, Chapman et al. (2006) found that none of the five muscles they investigated were influenced by an increase in cadence with respect to timing of neuromuscular activity. However, they only measured lower leg muscles, which may have affected the outcome. As noted in Baum and Li (2003), proximal muscles tend to display more pronounced changes than distal muscles. In addition, Chapman et al. (2006) used intramuscular EMG, which, as they themselves noted, may have contributed to the discrepancy between their findings and previous research.
The activation dynamics hypothesis postulates that neuromuscular activity occurs earlier in the crank cycle with increased cadence for force to occur at the same place in the crank cycle regardless of cadence. However, Sarre and Lepers (2007), although finding support for the activation dynamics hypothesis with regard to neuromuscular activation, also found that peak torque occurred later in the crank cycle with increased cadence. In line with this, both Ettema et al. (2009) and Samozino et al. (2007) found a shift in power production later in the crank cycle with increases in cadence.
To the best of the authors’ knowledge, only Sarre and Lepers (2007) have so far addressed the activation dynamics hypothesis by measuring both surface EMG and torque at the pedal in the same study, investigating them under identical circumstances. This was not their main focus, though. Other studies concerning the activation dynamics hypothesis have measured one of the variables (i.e., neuromuscular activation or resulting force) and compared it with previous studies on the other (Ettema et al., 2009, Sarre and Lepers, 2005), consequently not measuring the two variables under identical circumstances, and sometimes not even under similar circumstances (e.g., Sarre and Lepers, 2005, Sarre et al., 2004). As both neuromuscular activation and resulting muscle force are essential to the activation dynamics hypothesis, they should both be included when attempting to investigate the validity of this principle. Also, Sarre and Lepers (2007) investigated mainly muscles around the knee. A wider range of muscles covering the functions of all the lower leg joints should be studied to shed more light on the changes in neuromuscular coordination in the lower extremities with increased cadence. In addition to a wider muscle range, the current research employed a different mathematical technique, using data from the entire crank cycle, to illustrate the time shifts in neuromuscular activity and joint power. However, it was not the purpose of this study to establish a new standard for studying activation dynamics.
The aim of the current research was to investigate how neuromuscular activation and joint power and moment react to increases in cadence with respect to time phase shifts in relation to crank cycle position, addressing the activation dynamics hypothesis put forward by Neptune et al. (1997). According to this hypothesis, timing of surface EMG shifts in a negative direction, while timing of joint mechanics (joint power and moment) does not shift.
2. Methods
2.1. Subjects
Nine male competitive cyclists (mean±SD age 22±11.5years, height 182.9±6.6cm, and body mass 76.2±9.4kg) participated in the study. All subjects were recruited from the sports program at a local high school, including coaches who were active cyclists themselves. Mean±SD self reported competition experience and training hours per year were 4.4±2.9years and 500±75h, respectively. The study was approved by the regional committee for research ethics. All subjects signed an informed consent form prior to the experiment, and were made aware that they could withdraw from the study at any point without providing a reason.
2.2. Protocol
Before entering the main experiment, all subjects performed an incremental bicycle test. This test was used to locate the power output at which the subjects would pedal in the main experiment. Both seat and handlebar positions were adjusted according to the preferred position for each subject. The subjects started pedaling at a power output of 0W, which was increased by 25W every second minute. Heart rate (see below for equipment details) was noted as an average of the last 10s of each 2-min trial. The test ended when the subjects could no longer sustain the required power output. The highest heart rate achieved during this test was considered maximum heart rate. The target power for the main experiment was chosen by locating the work rate (with 25W increments) that resulted in a heart rate closest to 75% of maximum heart rate.
The main experiment consisted of a warm-up session and five pedaling sessions. The subjects warmed up for 5min pedaling at a power output of 100W at a freely chosen cadence. After warming up, they completed a total of five pedaling sessions at the power output corresponding to 75% of individual maximum heart rate, as found in the preliminary session, each lasting approximately 3.5min. The pedaling sessions were completed consecutively at pedaling rates of 60, 70, 80, 90, and 100rpm, respectively, with no rest in between. The cadence sequence was counterbalanced between the subjects (i.e., half of the group cycled from 60 to 100rpm, the other from 100 to 60rpm) to eliminate any systematic order effect. After 1min of pedaling, kinetic and surface electromyography (sEMG) data were collected three times for 15s, each with 45s between. The subjects were not informed when data collection occurred. All subjects were instructed to maintain a constant handlebar grip during the entire experiment, as well as a constant trunk position. A new pedaling session was started when the subjects reached the desired cadence, and they were instructed to stay within ±1rpm. Visual feedback on cadence was available on a computer screen located directly in front of the subjects. Seat and handlebar configurations were adjusted to the subjects’ preferred position, as registered in the preliminary session. Directly after the end of each pedaling session, heart rate was noted.
2.3. Equipment, data collection and analysis
All measurements were performed on a Velotron computer controlled bicycle ergometer, using Velotron Coaching Software version 1.5 (RacerMate, Seattle, WA, USA). This bicycle ergometer automatically adapts resistance to varying pedaling rate, so that the preset power output remains constant. Together with the Velotron Coaching Software version 1.5 it also allows for continuous visual feedback of pedaling rate. The bicycle was equipped with clipless racing pedals, allowing the subjects to wear their own shoes. Heart rate was monitored during all sessions (Polar T31 heart rate transmitter belt and a Polar FS1 heart rate monitor, Polar Electro, Kempele, Finland). Heart rate was noted at the end of each pedaling session and was then averaged, providing a mean heart rate for the entire main experiment for each subject. Percentage values were calculated according to the respective maximum heart rate for each subject.
All dynamics and sEMG data were recorded using Qualisys Track Manager version 2.0.365 (Qualisys, Gothenburg, Sweden), an internally synchronized motion analysis system that can collect kinematic, force, and EMG data. The data were further processed in Matlab 7.5 (Mathworks, Natick, MA, USA). All kinematic signals were recorded at a sample rate of 500Hz, using 10 ProReflex motion capture cameras (Qualisys, Gothenburg, Sweden), located in a 360° circle around the Velotron ergometer bicycle. Spherical reflective markers were placed bilaterally on the subjects’ hip (greater trochanter), knee (lateral epicondyle), and ankle (lateral malleolus) for calculation of hip, knee, and ankle joint angles in the sagittal plane. For hip angle analysis, the angle between the thigh and the horizontal was used. This was sufficient as the trunk position was constant for all attempts and we considered only changes in hip angle and compared movement ranges, not absolute values. In addition, two markers were placed on extensions of both pedals (in the form of a metal strip) in the sagittal plane of the bicycle (Ettema et al., 2009). The distance between the markers was 20cm, each located 10cm from the pedal axis. The ankle joint was defined by the angle between the knee, the ankle, and the pedal axis (Ettema et al., 2009).
Analogue signals (pedal forces and sEMG) were recorded at 1000Hz. Both pedals were equipped with two load cells (capacity 250kg per cell, Revere Model 9363, Breda, The Netherlands), able to detect normal (perpendicular) and shear (parallel) forces (resolution <0.5N) (Ettema et al., 2009). By applying full normal and shear forces of known magnitude, the load cells were calibrated. sEMG data were recorded from the right leg through a Bagnoli 16-channel desktop EMG system (Delsys, Boston, MA, USA), using bipolar electrodes (DE-2.1 single differential surface EMG sensors, bar dimension 10mm×1mm, bar spacing 10mm). The muscles included in the experiment were rectus femoris (RF), biceps femoris (BF), vastus medialis (VM), gastrocnemius (Gas), tibialis anterior (TA), and soleus (Sol). The ground electrode (3M Red Dot, St. Paul, MN, USA) was placed on the right patella. The skin over the muscle belly was shaved, lightly abraded, and cleaned with isopropanol alcohol (Isopropanol prima, Arcus Produkter AS, Oslo, Norway) to minimize noise. The electrodes were prepared with a thin coat of electrolytic gel (SignaGel, Parker Laboratories, Inc., Fairfield, NJ, USA), and fixed on the skin over the muscle belly, via Delsys adhesive skin interfaces, perpendicular to the muscle fiber direction. All wires from the electrodes were secured to the skin with adhesive tape to prevent movement artifacts in the signal. The subjects wore two thin, modified stockings (94% nylon, 6% spandex), one covering the thigh and one covering the lower leg, to help keep the electrodes and wires in place over an extended period of time. All subjects stated that the stockings in no way hampered their cycling technique or were uncomfortable in any other way. Raw sEMG signals were band-pass filtered (10–450Hz, 8th order, zero lag Butterworth), rectified, and smoothed with a low-pass filter (15Hz, 8th order, zero lag Butterworth). sEMG measurements were averaged across the three data collections for each pedaling session, resulting in a single sEMG envelope for each muscle recorded for each subject, containing one signal trace for each cadence. sEMG magnitudes were not normalized.
After acceleration artifacts were corrected for (see Ettema and Huijing, 1994), pedal forces (normal and shear) were converted to crank forces (normal and shear) through rotation of the coordination system, as done in Ettema et al. (2009). The right and left pedal coordination systems were at all times oriented in opposite directions. The vector sum of the forces from the right and the left pedal was used for further analysis of dynamics on the crank system. Using a 5-point differentiating filter, continuous crank velocity was calculated from crank angles. The data set was reduced to 360 samples (1 sample per crank degree) when the average crank cycle was calculated by performing an averaging interpolation on all included crank cycles expressed against crank angle. For each pedaling session the cadences calculated from the three data collections were averaged, providing a mean cadence for each trial for each subject. These values were again averaged to provide a group mean cadence for each cadence trial.
Power was calculated as the product of crank velocity and normal crank force. The power outputs calculated from the three data collections for each cadence were also averaged, providing a mean power output for each cadence for each subject.
To test the activation dynamics hypothesis, it is essential that the phase of the mechanical output at the muscular level is established. This was obtained through inverse dynamics calculations for a linked system of rigid segments (thigh, leg, and foot) (e.g., Hull and Jorge, 1985), providing net muscle moments and power for all three joints. This method is a standard procedure that is well described in the literature by e.g., Kautz and Hull (1993) and presented in detail for our set-up by Ettema et al. (2009). In short, the moments at the joints are calculated from the forces measured at the pedal, movements of the relevant body segments, and inertial estimates (mass and moment of inertia) of these segments, by applying Newton’s inertial laws. Parameters to calculate masses of the segments and their moments of inertia were taken from Soest et al. (1993). Pedal forces and moments contain both a muscular and a non-muscular (inertial) component (Kautz and Hull, 1993), which makes the interpretation of these variables complicated. In contrast, joint moments and power contain only the muscular component, as these variables are calculated through the method of inverse dynamics, separating the two components.
Due to the unique movement in cycling, the relative phase shift could be calculated through a circular cross-correlation. As opposed to a regular cross-correlation, where the amount of overlapping data gradually decreases with increasing phase, the current method estimates the correlation over the same amount of data throughout the whole process. When shifting the second data set relative to the first by n samples, the last n samples that do not overlap with the first data set are included in front. In equation:
(1)Using this method, the phase resulting in the highest correlation was considered true phase, providing phase shift in angle. The phase shift was calculated between each adjacent set of cadences (i.e., between 60 and 70rpm, 70 and 80rpm, etc.) and the cumulative sum was averaged to obtain total phase shift relative to the lowest cadence (60rpm). The time delay related to the phase shift was obtained according to Ettema et al. (2009) and Li and Baum (2004), being the slope of the cadence-phase relationship, and is thus defined as:
(2)2.4. Statistical analysis
The difference between heart rate at target power output and calculated 75% of maximum heart rate was tested using a t-test. Analysis of joint dynamics was done conducting a two-way ANOVA for repeated measures to assess the effect of the variable factor (power, moment, velocity) and the joint factor (hip, knee, ankle) on phase shift (t, Eq. (2)). Statistics were performed in SPSS for Windows 14.0 (SPSS Inc., Chicago, IL, USA). Mauchly’s Test of Sphericity showed that sphericity could be assumed. Statistical significance was established at a level of p<0.05.
3. Results
The cadences employed by the subjects during the main experiments amounted to 60.6±1.0, 70.5±0.7, 80.4±0.8, 90.3±1, and 100.2±1rpm for the targeted 60, 70, 80, 90, and 100rpm conditions, respectively. The subjects performed the tests at a heart rate of 68.4±4.4% of maximal heart rate (195.3±13.3 beats per minute), somewhat lower than planned. The work rate was 175±25W.
Fig. 1 shows, as an example, joint moments, joint power, and rectified EMG data for one subject in polar diagrams. Joint moment curves shift with increasing cadence to a later stadium in the crank cycle. Also joint power traces show some shifts but these are less clear for this subject. The sEMG traces show variable results with regard to timing. BF, VM, Gas, and Sol show a shift, generally opposite of what is shown for joint moments, but the other muscles do not show a clear shift in any direction. Note that the data in Fig. 1 does not necessarily reflect the average outcome over all subjects. Considering all subjects, in all but one case the sEMG traces showed a particular (individual) pattern independent of cadence (as in Fig. 1) that only shifted in time, leading to cross-correlation values of 0.81 or higher. Only for TA, the cross-correlation values for two adjacent cadences were rather low (⩽0.665) in four out of 36 cases. In these cases, visual inspection did not indicate a change in activation pattern, only in amplitude. In one subject, Sol showed a changing sEMG pattern (not only phase shift) from 80 to 90rpm. In that case, it was impossible to regard the changes as a time shift and these data were not considered in the statistical analysis (but the result is depicted in Fig. 3).
Fig. 1.
Polar diagrams of joint moments, joint powers, and smoothed rectified sEMG with increasing cadence, relative to the crank cycle, for one subject. Negative values are indicated by dotted lines. Black dot in first polar diagram represents crank axis of rotation.
Fig. 3.
Phase shift as a function of cadence for sEMG of all muscles in all subjects. Dotted line for Sol is the one subject who showed a changing sEMG pattern. These data were excluded from further analysis.
Fig. 2 shows the shift in angle as a function of cadence for joint moment, velocity, and power. The slopes, i.e., time delays t, are presented in Table 1. All mean slopes were significant, and only in a very few cases did the individual curves not show a significant linear correlation (i.e., the t values at hand were not significantly different from zero). Together with the small group standard errors of the mean, these findings indicate that the group of subjects behaved in a very stereotyped manner regarding the cadence dependent phase shift.
Fig. 2.
Phase shift for (A) joint moment, (B) joint velocity and crank velocity, and (C) joint power, as a function of cadence. Vertical bars indicate s.e.m. Joint figure legend is valid for all graphs.
Table 1. Time shift t, mean and (SD), of the crank cycle in seconds for power, moment, and velocity at the hip, knee, and ankle joint, as well as related variables measured at the crank system.
| Variable | Joint | |||
|---|---|---|---|---|
| Hip | Knee | Ankle | Crank | |
| Power | 0.066 | 0.074 | 0.059 | 0.085 |
| (0.018) | (0.029) | (0.038) | (0.027) | |
| (Eff. force) | ||||
| Moment | 0.142 | 0.099 | 0.186 | 0.141 |
| (0.031) | (0.035) | (0.026) | (0.023) | |
| Velocity | 0.033 | 0.028 | 0.062 | 0.058 |
| (0.014) | (0.006) | (0.038) | (0.043) | |
The two-way ANOVA revealed a significant effect of both the variable factor (p<0.001) and the joint factor (p=0.006) on the size of the time shifts, as well as an interaction (p<0.001). Joint moment showed a considerably larger time shift than power and angular velocity. Joint moment and velocity differed significantly for all joints (p=0.001; p=0.014, respectively) in the same order. However, time shifts in joint power were not significantly different between joints.
The sEMG results show a different picture. The individual cadence-phase shift data are shown in Fig. 3. In many cases (22 out of 54), no significant linear regression correlations were obtained between cadence and phase. Furthermore, the regression slopes showed a considerable inter-individual variation around a group average relatively close to zero, not exceeding 0.04s; none of the sEMG phase shifts were significant. The average values were RF −0.024, Gas −0.034, BF −0.024, TA −0.012, VM −0.001, and Sol −0.01s, respectively.
4. Discussion
The aim of this study was to investigate the activation dynamics hypothesis, posing that, with increasing cadence, muscle activity occurs earlier in the crank cycle so that the timing of the mechanical outcome is constant and compensates for the effects of electromechanical delay. The main findings of the present study reject this hypothesis: we found a distinctive and very consistent phase shift of the mechanical outcome, irrespective of the variable that was considered to determine the phase shift; muscle activation of six major muscles in the lower extremity showed a wide inter-individual variation with regard to phase shift. Two subjects showed a negative phase for all muscles, one of which was clearly consistent (t=−0.073±0.008). Only for this subject, the sEMG results are distinctively in agreement with the activation dynamics hypothesis. However, all individuals (including this one subject) showed a positive phase shift for the mechanical outcome variables. In other words, for none of the subjects a shift of the sEMG profile occurred to maintain a constant mechanical profile of the crank cycle. Our EMG results contradict earlier findings (e.g., Baum and Li, 2003, Bieuzen et al., 2007, Li and Baum, 2004, Marsh and Martin, 1995, Neptune et al., 1997). This may be due to methodology differences. These earlier studies considered particular events of the EMG pattern, e.g., onset, offset, and burst peak. By using cross-correlation, we employed the muscle activity signal over the entire crank cycle. Although timing of onset and offset of muscle activation is informative, it does not provide the full picture on timing of muscle activation. Furthermore, these calculations are sensitive for calculation methods and noise (Hodges and Bui, 1996, Hug and Dorel, 2009).
Even though our results suggest that the activation dynamics hypothesis should be rejected, the electromechanical delay (EMD) could still explain the combination of phase shifts of muscle activation and the mechanical outcome (i.e., joint moments). In such a case, the difference between sEMG and joint moment phase shifts should be similar to EMD values reported in the literature (around 70–100ms; Ingen Schenau et al., 1995, Vos et al., 1991). A rough estimate based on the data in Table 1 and about −0.02s shifts for most muscles indicates values of about 0.16, 0.12, and 0.2s for hip, knee, and ankle muscles. It should be noted that any effects of possible altered coordination are ignored in this rationale. Still, these values, except possibly the knee, are too high to be explained by EMD.
The sEMG magnitudes were, as depicted in Fig. 1, generally higher for higher cadences for all subjects across all muscles (with the exception of BF displaying a varied response). One possible explanation is that at higher cadences the activation period is shorter, creating a need for a higher electrical activation burst to obtain the required average level of active state, a mechanism closely associated with EMD (see Soest and Casius, 2000). Secondly, when cadence increases, muscle velocity also increases. If we may assume that the lower cadences used in this study are close to optimal for power, then the higher cadences will result in less power at the same activation level. In such case, the level of activation must increase with increasing cadence. However, this explanation is less likely as the optimal cadence for power may be relatively high, i.e., 90–100rpm (Kohler and Boutellier, 2005; see also Soest and Casius, 2000).
The effects of cadence on joint power time shifts are in strong agreement with a previous study (Ettema et al., 2009), despite the fact that a different type of ergometer was used. These two ergometers have a different effect on the choice of cadence at a particular work rate, which may be related to different control strategies (Leirdal and Ettema, 2009). This strengthens the notion that the cadence dependent phase shift that was described in these studies is universal for cycling. Furthermore, this phase shift is not due to the increasing effect of inertial forces at the lower extremity (see e.g., Kautz and Hull, 1993), nor does any phase shift occur locally to counteract the effect of inertial forces such that a constant propulsion profile at the crank is obtained: both joint and crank dynamics show considerable phase shifts. Thus, the question arises if the phase shifts express a cadence dependent (functional) alteration in cycling technique and coordination. The difference in behavior between joints (net moment and velocity, Fig. 2) suggests that this indeed is the case. An interesting finding is that power time shifts in a similar fashion with cadence at all three joints as well as at the crank, which also was the case in Ettema et al. (2009). It may be hypothesized that the different shifts in joint moments and velocities must occur to obtain the same power output, the coordination of which (i.e., relative timing between joints and crank) is independent of cadence. Such a consistent phase shift of the power crank cycle could be considered an invariant characteristic of cycling coordination (see e.g., Gruben et al., 2003); irrespective of cadence, power is produced in the same coordination pattern, be it with a shift in phase. However, even if it were confirmed that the relative timing of power is an invariant characteristic in cycling, the consequence of the phase shift is that the involved muscles must perform in a variant manner to establish this. Obviously, shortening speed and time of contraction are affected by cadence. However, the length range that power is produced at apparently differs with cadence because it is produced at a shifting crank position and therefore at different joint angles. New research is needed to identify the specific mechanisms behind the phase shifts as well as to provide answers as to what extent muscle performance is affected.
References
- . Lower extremity muscle activities during cycling are influenced by load and frequency. J Electromyogr Kinesiol. 2003;13(2):181–190
- . Muscle activation during cycling at different cadences: effect of maximal strength capacity. J Electromyogr Kinesiol. 2007;17(6):731–738
- . Leg muscle recruitment in highly trained cyclists. J Sports Sci. 2006;47(2):115–124
- . Frequency response of rat gastrocnemius medialis in small amplitude vibrations. J Biomech. 1994;27:1015–1022
- . The effects of cycling cadence on the phases of joint power, crank power, force and force effectiveness. J Electromyogr Kinesiol. 2009;19(2):e94–e101
- . The control of foot force during pushing efforts against a moving pedal. Exp Brain Res. 2003;148(1):50–61
- . A comparison of computer-based methods for the determination of onset of muscle contraction using electromyography. Electroencephalogr Clin Neurophysiol. 1996;101(6):511–519
- . Electromyographic analysis of pedaling: a review. J Electromyogr Kinesiol. 2009;19(2):182–198
- . A method for biomechanical analysis of bicycle pedalling. J Biomech. 1985;18(9):631–644
- . The control of mono-articular muscles in multijoint leg extensions in man. J Physiol. 1995;484(1):247–254
- . A theoretical basis for interpreting the force applied to the pedal in cycling. J Biomech. 1993;26(2):155–165
- . The generalized force-velocity relationship explains why the preferred pedaling rate of cyclists exceeds the most efficient one. Eur J Appl Physiol. 2005;94(1–2):188–195
- . Freely chosen pedal rate during free cycling on a roller and ergometer cycling. Eur J Appl Physiol. 2009;106(6):799–805
- . Electromechanical delay estimated by using electromyography during cycling at different pedaling frequencies. J Electromyogr Kinesiol. 2004;14(6):647–652
- . The relationship between cadence and lower extremity EMG in cyclists and noncyclists. Med Sci Sports Exerc. 1995;27(2):217–225
- . The effect of pedaling rate on coordination in cycling. J Biomech. 1997;30(10):1051–1058
- . Why does power output decrease at high pedaling rates during sprint cycling?. Med Sci Sports Exerc. 2007;39(4):680–687
- . Neuromuscular function during prolonged pedalling exercise at different cadences. Acta Physiol Scand. 2005;185(4):321–328
- . Cycling exercise and the determination of electromechanical delay. J Electromyogr Kinesiol. 2007;17(5):617–621
- . Cadence, power output and mechanical torque during pedaling. J Hum Mov Stud. 2004;47:133–142
- . Which factors determine the optimal pedaling rate in sprint cycling?. Med Sci Sports Exerc. 2000;32(11):1927–1934
- . The influence of the biarticularity of the gastrocnemius muscle on vertical-jumping achievement. J Biomech. 1993;26(1):1–8
- . Electromechanical delay during knee extensor contractions. Med Sci Sports Exerc. 1991;23(10):1187–1193
David McGhie received his MSc in 2008 at the Human Movement Science Programme at The Norwegian University of Science and Technology in Trondheim, Norway. He is currently working at the same programme on his PhD in biomechanical analyses on artificial turf.
Gertjan Ettema received his MSc (1986) and PhD (1990) in Human Movement Science from Vrije Universiteit, Amsterdam. From 1991 to 1997 he worked at the University of Queensland, Brisbane. Since 1998, he serves as professor at the Human Movement Science Programme at The Norwegian University of Science and Technology in Trondheim, Norway. His research interests are motor control and biomechanics in human movement, particularly ski jump, cross-country skiing, and bicycling.
PII: S1050-6411(10)00084-2
doi:10.1016/j.jelekin.2010.04.007
© 2010 Elsevier Ltd. All rights reserved.
Volume 21, Issue 1 , Pages 18-24, February 2011
