Med Sci Sports Exerc. 1998; 30(3): 442-9
Neuromuscular, metabolic, and kinetic adaptations for
skilled pedaling performance in cyclists
TAKAISHI, TETSUO; YAMAMOTO, TAKASHI; ONO, TAKASHI; ITO, TOMONORI; MORITANI, TOSHIO
Institute of Natural Sciences, Nagoya City University, Nagoya 467, JAPAN; Laboratory of Exercise Physiology and Biomechanics, School of
Physical Education, Chukyo University, Toyota 470-03, JAPAN; Graduate School, Aichi University of Education, Kariya 448, JAPAN; and
Laboratory of Applied Physiology, Graduate School of Human and Environmental Studies, Kyoto University, Kyoto 606, JAPAN
Submitted for publication October 1996. Accepted for publication September 1997.
Purpose: The purpose of this study was to clarify the reason for the difference in the preferred cadence between cyclists
and noncyclists. Methods: Male cyclists and noncyclists were evaluated in terms of pedal force, neuromuscular activity for
lower extremities, and oxygen consumption among the cadence manipulation (45, 60, 75, 90, and 105 rpm) during pedaling
at 150 and 200 W. Noncyclists having the same levels of aerobic and anaerobic capacity as cyclists were chosen from
athletes of different sports to avoid any confounding effect from similar kinetic properties of cyclists for lower extremities
(i.e., high speed contraction and high repetitions in prolonged exercise) on both pedaling performance and preferred
cadence. Results: The peak pedal force significantly decreased with increasing of cadence in both groups, and the value for
noncyclists was significantly higher than that for cyclists at each cadence despite the same power output. The normalized
iEMG for vastus lateralis and vastus medialis muscles increased in noncyclists with rising cadence; however, cyclists did not
show such a significant increase of the normalized iEMG for the muscles. On the other hand, the normalized iEMG for
biceps femoris muscle showed a significant increase in cyclists while there was no increase for noncyclists. Oxygen
consumption for cyclists was significantly lower than that for noncyclists at 105 rpm for 150 W work and at 75, 90, and 105
rpm for 200 W work. Conclusions: We conclude that cyclists have a certain pedaling skill regarding the positive utilization
for knee flexors up to the higher cadences, which would contribute to a decrease in peak pedal force and which would
alleviate muscle activity for the knee extensors. We speculated that pedaling skills that decrease muscle stress influence the
preferred cadence selection, contributing to recruitment of ST muscle fibers with fatique resistance and high mechanical
efficiency despite increased oxygen consumption caused by increased repetitions of leg movements.
In general, cyclists with long experience prefer to adopt
significantly higher pedaling cadences than noncyclists.
However, it has been well known that employing a higher
cadence, at which cyclists train and race, leads to poor
mechanical efficiency (2,3,9,22). The reason of this is
because oxygen consumption (˙VO2) during exercise
increases linearly or exponentially with increasing
cadence, owing to an increase in internal work for
repetitive limb movements(28) and an increase in oxygen
demand per pedal thrust to compensate for the lack of
pedaling skill at higher cadences(4).
We have been studying factors that decide preferred
cadence(26,27). In these studies, we have demonstrated
that the cadence at which neuromuscular fatigue is
minimized (estimated through electromyogram from the
vastus lateralis muscle) coincides with the preferred
cadence, rather than the most economical cadence. This
cadence was about 70 rpm for noncyclists and between 80
and 90 rpm for cyclists during moderate work intensity.
These findings imply that cyclists have acquired some kind
of pedaling skill that contributes to a shift of the cadence
for the minimal neuromuscular fatigue to a higher one in
the process of training.
In recent years, many researchers have investigated the
differences between cyclists and noncyclists in
measurements such as mechanical efficiency(17), peak
pedal force (23), muscle activity pattern(9,16,18,19,24),
and preferred cadence (18) under various experimental
conditions. For example, Marsh and Martin (18) compared
experienced cyclists with aerobically equally fit noncyclists
and found that preferred cadence for constant work
output is dependent on the level of physical fitness of the
subject. Their study is noteworthy because of some special
characteristics of subjects. However, as they have pointed
out themselves, it still remains unclear whether force-
velocity properties of the lower extremity muscles
adopted in one's training history has a significant effect on
preferred cadence. Meanwhile, it is expected that both
preference for pedaling cadence and muscle fatigue in the
lower extremity are strongly dependent on relative
intensity to the maximal working capacity in leg
muscles(12). So far as we know, there are no studies in
which anaerobic power for pedaling exercise was taken
into consideration. We suppose that skilled pedaling
performance in cyclists, which is effective for preventing
neuromuscular fatigue in working muscles at higher
cadences, should be examined
measurements obtained from cyclists and noncyclists with
similar physical characteristics and fitness.
The purpose of this study is to elaborate on adaptations to
pedaling exercise by clarifying pedaling skills in terms of
neuromuscular, metabolic, and kinetic differences
between cyclists and noncyclists of similar aerobic and
Med Sci Sports Exerc. 1998; 30(3): 442-9
anaerobic fitness. In this study, noncyclists were chosen
from athletes participating in sport teams at colleges to
avoid confounding from previous training, which might
give cyclists a kinetic advantage at lower extremities(i.e.,
relatively high speed of muscle contractions, high
repetitions in prolonged exercise) on both pedaling
performance and preferred cadence.
Subjects. Seven cyclists (25.9 (SD 5.1) years, cycling
experience of more than 4 yr; ranging 4 to 15 yr) and
seven noncyclists (22.1 (SD 1.9) years, an endurance
runner, two soccer players, a baseball player, two middle
distance swimmers, and a rugby player) participated in
this study. None of the subjects have suffered from
injuries for more than 2 yr that would effect the present
pedaling exercises. The study was approved by the
Institutional Review Board for Use of Human Subjects at
Nagoya City University. All experimental procedures were
explained in detail to each subject who then signed a
statement of informed consent in accordance with the
policy statement of the American College of Sports
experiments(experiment I, II, and III). On the first visit,
subjects performed experiment I and one session of
experiment II. On the second and third visits, they
performed two other sessions of experiment II,
respectively. On the fourth visit, they performed
experiment III. The explanation for the experiments are as
follows. Experiment I consisted of two different pedaling
trials at 150 and 200 W. In both trials, subjects were
required to pedal while randomly changing the cadences
every 40 s at 45, 60, 75, 90, and 105 rpm. Each trial was
preceded by a warm-up exercise consisting of 1 min of
unloaded pedaling and 2 min of pedaling at 100 W.
Experiment II consisted of five sessions of pedaling at five
different cadences (45, 60, 75, 90, and 105 rpm). Each
session was composed of a continuous 2 min of unloaded
pedaling, 3 min of pedaling at 100 W, 5 min of pedaling at
150 W, and 5 min of pedaling at 200 W, respectively.
However, the work rate for the first session of experiment
II, which was performed at the latter half of the first visit,
was continuously increased from 200 W by 25 W each
minute, and the session was utilized as a part of a step test
to the point of exhaustion. During the step test, which was
started at 225 W, the cadence was freely chosen by
subjects between 75 and 90 rpm. Both experiment I and
experiment II were carried out with an electromagnetically
braked ergometer, which has standard toe clips and can
control the power output regardless of cadence. Between
the second and third sessions and the fourth and fifth
sessions, a rest of 2 h or more was taken. After performing
all sessions in experiment II, each subject was asked which
cadence they preferred. Experiment III consisted of six or
seven sessions of very short (for 10 s) pedaling with
maximal effort against a different torque(ranging from 1.5
to 11.0 kp) on an electromagnetically
ergometer(Powermax V, Combi, Japan). The ergometer
has standard toe clips and displays the maximal cadence
for each session. The reliability of the cadence was
confirmed by using the electric signals from a goniometer
The subjects joined three different
attached to a leg. A sufficient rest of at least 10 min or
more was taken between the sessions.
Measurement of oxygen uptake. Measurements of
oxygen uptake(˙VO2) were made in all subjects for the step
exercise (experiment II). During the five sessions of
pedaling and following a step test to the point of
exhaustion for the first session, respiratory measurements
were obtained by our on-line computer system, which
consisted of a mass spectrometer (WSMR-1400, Westron,
Japan) and pneumotachograph connected to a respiratory
flow transducer. The analog signals of fractional
concentration of O2, CO2, and N2 from the mass
spectrometer and those from the flow transducer were
continuously digitized at 100 Hz by the computer system
(PC-9801 F III, NEC, Japan). The ˙VO2 carbon dioxide
production and expired ventilation were calculated every
20 s, and those data were stored on a floppy disk for
subsequent analysis. For the determination of˙VO2 for 150
W and 200 W pedaling, the mean ˙VO2 after the initial 3
min for each step was obtained because it would have
taken at least 2 min to reach a steady state of ˙VO2 at the
Measurement of EMG. Myoelectric signals during the five
sessions of pedaling (experiment I) were recorded by the
surface EMG technique. The EMG instrumentation used in
the present experiment has been fully described in our
previous study (26). Briefly, two miniature electrodes(Ag-
AgCl, 6-mm contact diameter, 4-cm interelectrode
distance) were placed over the belly of three muscles
(biceps femoris (BF), vastus lateralis (VL), and vastus
medialis (VM)), and a reference electrode was placed over
the anterior superior spine of the iliac crest. All electrode
placements were preceded by abrasion of the skin to
reduce the source impedance to less than 2 kΩ.
Myoelectric signals were amplified (AM-601G, Nihon
Koden, Japan) with band pass filtering (5-500 Hz) and
recorded on the two same digital recorders (RD-101T,
Teac, Japan). The recorded data were digitized at a
sampling rate of 1 kHz, and the iEMG was calculated using
the stable 20-s period at five different cadences for the
sessions using a Hewlett-Packard 98580C desktop
computer. The cadence was estimated by using the
electric signals from a goniometer attached to a leg and
judged to be stable when the difference from a given
cadence was less than 1 rpm.
Measurement of pedal force. To obtain information on
the pedaling technique of each subject, a modified pedal
with a standard toe clip was mounted to the right crank
for experiment I. The strain-gauged force transducer
equipped inside the pedal produced an analog signal that
indicated the magnitude of the force perpendicular to the
pedal (pedal force). Electric signals from a DC amplifier
(DPM-601A, Kyowa, Japan) was simultaneously recorded
with electromyographic signals on the data recorder. The
recorded data were digitized at a sampling rate of 1 kHz.
Peak force for pedaling in each subject at a certain
cadence was represented by the average of peak forces
for pedal thrusts during the same period for the iEMG
estimation at each cadence by the use of a Hewlett-
Packard 98580C desktop computer. The peak pedal force
for each cadence was determined by the average of the
Med Sci Sports Exerc. 1998; 30(3): 442-9
peaks of pedal force at the same period for calculating the
Data analysis. Because there was a large variance in the
absolute value of the iEMG between subjects, muscle
activity in individuals was normalized by setting the iEMG
at 45 rpm as the standard. The iEMG at each cadence
(consisting of 20-s samples) was divided by the subject's
iEMG at 45 rpm. Then, the value at each cadence was
expressed in percentages. These values were named
Statistics. Means and standard deviations for height, body
mass,˙VO2max, maximal anaerobic power, ˙VO2, peak pedal
force, and normalized iEMG were calculated. Student's t-
tests were conducted to compare the means for height,
body mass, ˙VO2max, and maximal anaerobic power
between the groups. Two-way ANOVAs were conducted to
examine the effect of cadence and cycling experience on
˙VO2, peak pedal force, and normalized iEMG. With regard
to normalized iEMG, data at 45 rpm as a standard were
excluded from measures for ANOVA. WhenF-ratio was
significant (P < 0.05) in ANOVA,post-hoc multiple
comparison using Tukey's procedure was conducted. The
difference in the mean values between groups (i.e., cyclist
and noncyclist) at each cadence was analyzed by
independent t-test. In all analyses, differences were
considered significant at P < 0.05.
Figure 1 shows a typical set of data indicating a highly
linear relationship (r = -0.999) between pedaling torque
and cadence obtained by experiment III and a resultant
torque-power curve (parabolic curve) obtained by the
relationship in one subject. As reported in the previous
studies (20,21), a linear relationship between cadence and
pedaling torque was obtained in all of our subjects with
significant (P < 0.01) correlation coefficients ranging from-
0.993 to -0.999. Maximal anaerobic power for each
subject was obtained at the peak of the torque-power
curve. The physical characteristics and fitness levels of the
present cyclists and noncyclists are shown inTable 1. The
mean values of height, body mass,˙VO2max, and maximal
anaerobic power were not significantly different between
the groups. On the basis of these results, we regarded that
the groups had the same levels of aerobic and anaerobic
Figure 1-Typical data set for pedal torque and cadence(
linear line ) and estimated power and cadence( parabolic
curve ) for single subject. Peak parabolic curve indicates
the maximal anaerobic power.
TABLE 1. Physical characteristics and fitness levels for
cyclists and noncyclists (means (Â±SD)).
Figure 2 shows the differences of peak pedal force in each
cadence at 200 W (A) and 150 W (B) for both groups. The
mean values of peak pedal force demonstrated a
hyperboliclike curve and significantly decreased according
to the increase of cadence from 45 rpm. For cyclists, there
were no significant differences between 90 and 105 rpm
at both 150 and 200 W. For noncyclists, there were no
significant differences between 75 and 90 rpm and
between 90 and 105 rpm at 150 W. As a result of
comparing peak pedal forces between the groups, the
mean value for noncylists was significantly higher than
that for cyclists at all cadences except for 45 rpm. The
differences of the normalized iEMG against cadence
manipulation is shown in Figure 3, A (200 W) and B (150
W), respectively. Similar to the result demonstrated in our
previous study (seeFig. 3 in (27)), the normalized iEMG for
VL and VM for noncyclists showed a quadratic curve with a
bottom at 60 rpm, and the values significantly increased
with the increase of cadence; however, cyclists did not
have a difference in values for the cadence manipulation.
On the other hand, the normalized iEMG for BF for cyclists
showed an abrupt increase, and the values were
significantly different between 75 and 90 rpm at 200 W,
whereas such an increase was not obtained for
noncyclists. The results of both preferred and disliked
cadences for pedaling at 200 W is illustrated between
panels A and B in Figure 3. Four of the noncyclists
preferred 60 rpm, and the other noncyclists preferred 75
rpm. On the other hand, one of the cyclists preferred 75
rpm and the other cyclists preferred 90 rpm. All the
subjects disliked to pedal at the cadence of 45 rpm. Figure
4 shows the differences of the˙VO2 in each cadence
condition at 200 W (A) and 150 W (B) for both groups.
Consistent with previous literature(2,3,18,25,26,27),˙VO2
demonstrated a quadratic curve with a bottom at 60
and/or 75 rpm, and ˙VO2 for both groups significantly
increased from 75 rpm according to the increase of
cadence. ˙VO2 for cyclists was significantly lower than that
for noncyclists at 105 rpm for pedaling at 150 W and at 75,
90, and 105 rpm at 200 W. ˙VO2 at the pedaling conditions
for 150 W ranged from 52.5 to 61.2% of ˙VO2max for cyclists
and 54.3 to 64.4% of ˙VO2max for noncyclists, respectively.
As for pedaling at 200 W, the values ranged from 70.7 to
75.8%˙VO2max for cyclists and 71.5 to 80.9% ˙VO2max for
Med Sci Sports Exerc. 1998; 30(3): 442-9
Figure 2-Relationship between means of peak pedal force
and cadence for cyclists (â–¡) and noncyclists (â–ª) during
pedaling at 200 W (A) and 150 W (B). Vertical lines
indicate standard deviations. Open asterisks and closed
asterisks show significant difference in values among
cadences for cyclists and noncyclists, respectively. Open
and closed asterisks, P < 0.05. SS indicates significant
difference in values between cyclists and noncyclists at
each cadence( P < 0.01).
Figure 3-Relationship between cadence and normalized
iEMG for cyclists ( open symbols ) and noncyclists( closed
symbols ) during pedaling at 200 W (A) and 150 W(B).
Vertical bars for standard deviation and asterisks to
show statistic differences were omitted for clear
presentation. Arrows inserted between A and B show
preferred and disliked cadences for subjects. N and C
indicate noncyclist and cyclist, respectively. Numerals
mean number of subjects who prefer or dislike to pedal
at the cadence.
Figure 4-The relationship between cadence and oxygen
consumption for cyclists (â–¡) and noncyclists (â–ª)
during pedaling at 200 W (A) and 150 W (B). ( Single
asterisks ( open and closed ), P < 0.05; double asterisks (
open and closed), P < 0.01). S and SS indicate significant
difference in values between cyclists and noncyclists at
each cadence (S, P < 0.05; SS, P < 0.01).
Figure 5 shows a comparison of rectified EMG for VL and
BF with one cyclist (A) and one noncyclist (B) during
pedaling at 200 W. As shown at the top of Figure 5A, the
peak of rectified EMGs for VL did not show a decrease in
the cadence manipulation despite the peak pedal force for
a pedal thrust being significantly decreased(Fig. 2). As
shown in the middle ofFigure 5A, the rectified EMG for BF
at 90 rpm for a cyclist showed a larger amplitude than that
at 60 rpm, whereas the activity of the muscle for a
noncyclist was comparatively smaller at both cadences.
The peak of the shifted rectified EMG at the bottom of the
figure clearly overlapped with the latter half of the
rectified EMG in cyclists. Such a small peak at the latter
half of the rectified EMG for BF was observed in six of the
seven cyclists and one of the seven noncyclists.
Med Sci Sports Exerc. 1998; 30(3): 442-9
Figure 5-Typical set of rectified EMG signals for VL (top), BF (middle), and shifted VL (bottom) for cyclist (A) and noncyclist
(B). The EMGs at the bottom were shifted by 500 and 333 ms, which are equal to duration for half-round of the crank at 60
and 90 rpm, respectively, to simulate muscle activities from the opposite leg.
Pedaling skill acquired by cycling training. In accordance
with our previous studies (26,27), preferred cadences with
cyclists (75 or 90 rpm) were higher than those with
noncyclists (60 or 75 rpm) in this study. Considering that
both groups had similar physical characteristics and fitness
(Table 1), difference in preferred cadences between
groups indicates that cyclists have acquired at least some
pedaling skill, which affects preferred cadence and
measurements, during the course of cycling training. First,
we examine the differences in pedaling skill between the
Ekblom and Goldberg (8) reported that the rates of
perceived exertions (RPE) for running and swimming were
lower than that for pedaling when they were compared at
the same ˙VO2 and suggested that RPE was strongly
related to feeling of strain in overcoming resistance in
working muscles compared with some central factor, i.e.,
perceived tachycardia and dyspnea. Their findings suggest
that difference in resistance for a pedal thrust, which
affects leg muscle strain, plays an important role in
deciding the preferred cadence for pedaling exercise.
Previously, we expected that noncyclists would not
possess any particular skill for effective transmitting of
actual leg force to the crank. Also, we assumed pedal force
would not decrease at higher cadences because in our
previous experiment(26) some noncyclists indicated that
they could not feel a decrease in pedal thrust load at
higher cadences. However, the peak pedal force for
noncyclists (Fig. 2) reached a minimum at the highest
cadence in the present study, as it did in Patterson and
Moreno(23). The result indicates that noncyclists also have
a pedaling skill to reduce pedal force at much higher
cadences (=105 rpm) than their preferred cadence (60 or
75 rpm), although they needed significantly higher peak
pedal force than cyclists at higher cadences.
We further examined iEMG to explain the difference in the
preferred cadences between the groups. Figure 3A
demonstrates that noncyclists required an increase in
muscle activity for VL and VM, which are crucial in
generating force for pedal thrust, at cadences higher than
90 rpm. The results for normalized iEMG and peak pedal
force suggest that noncyclists experienced significantly
more strain in knee extensors than cyclists. On the
contrary, cyclists had a significant increase in muscle
activity for BF against cadence manipulation. Such
differences in muscle activity provide manifestation of a
large difference in pedaling skill between the groups and
hint at adaptation to pedaling exercise.
A possible reason for the difference in peak pedal force is
pedaling skill for coordination between legs (6,9). Ericson
et al.(9) have reported that for positive knee flexion with
cyclists, EMG for knee flexors has a smaller peak during
the backstroke, between the bottom and top pedal
positions. Our result demonstrating the overlap in EMGs
for VL and BF (Fig. 5A) indicates that positive pedal lifting
was being performed to alleviate the load needed in
pushing down on the other pedal. The observed iEMG
Med Sci Sports Exerc. 1998; 30(3): 442-9
change in our cyclists strongly suggests that positive
utilization of BF contributes to alleviate the amount of
muscle activity and the intensity of muscle stress for VL
and VM up to the higher cadences.
Löllgen et al. (17) reported that the rate of peak pedal
force to the maximal force at each cadence gradually
increases, even if the absolute value of peak pedal force
decreases with rising cadence. Additionally, they showed
that the time-derivative force at 100 rpm was significantly
larger than at 80 rpm with the same subjects. The findings
imply that pedaling exercise at a higher cadence does not
necessarily lead to a decrease in effort required to push
the pedal down with knee extensors. This possibly elicits
feeling of more intensive muscle strain, although the peak
pedal force actually decreases with an increase in
cadence. In the present study, six of the seven cyclists
employed the pedaling technique to lift up the pedal with
the knee flexors. It is not clear whether this pedaling
technique is commonly acquired and employed by highly
trained cyclists because BF has a large variability in the
action patterns(16,24). However, when considering that
peak pedal force of cyclists at 75, 90, and 105 rpm were
observed to be about 86.8, 83.6, and 82.8% of those of
noncyclists, it seems that the pedaling technique
employed by our cyclists might contribute to a decrease in
feeling of muscle strain in extensors and greatly affect
preferred cadence selection.
On the other hand, Marsh and Martin (19) reported that
preferred cadence for constant work output is dependent
on the level of physical fitness of the subject with the
experienced cyclists and aerobically equally fit noncyclists.
In their study, the preferred cadence for cyclists and
noncyclists (endurance runners) were 85 and 92 rpm,
respectively. This result seems to conflict with the present
finding. When paying attention to the iEMG changing
patterns for the extensors (see Table 1 in (19)), it is
apparent that both groups of subjects in their study show
similar iEMG patterns for knee extensors as our cyclists,
although the runners would not have a pedaling technique
with regard to the knee flexors. Therefore, as they pointed
out themselves, preferred cadence similarity obtained in
their study is possibly explained by the similarity in
dynamic muscular training of the groups. It is supposed
that prolonged running training, which consists of high
frequent repetitions of leg movements, would cause
runners to acquire a skilled performance related to
efficient exertion of leg force by knee extensors.
The issue of how our findings translate to daily life should
be addressed. Himann et al. (14) reported that step
frequency for daily walking with young adult males was
119.7 (SD ± 10.1) steps·min-1. In addition, Zarrugh and
Radcliffe(30) demonstrated that the maximum step
frequency for very fast-paced walking was less than 140
steps·min-1(equivalent to 70 rpm for pedaling). We often
see that subjects pedal at a very low cadence (less than 50
rpm) until they are given a concrete indication of cadence
on their first visit to the laboratory. Therefore, if preferred
cadence is related to kinetic properties of leg muscles at a
familiar workload, it becomes clear why the majority of
our noncyclist subjects(accustomed to just walking)
preferred pedaling at 60 rpm, and why none of them
preferred 90 rpm. These findings strongly suggest that
athletic cycling needs completely different muscle usage
for its contraction velocity from bicycle riding in daily life.
Here, we should remember that cocontraction between
agonists and antagonists, which does not contribute to
generating pedal force, might have an effect on muscle
activity increase. Regrettably, the approach using
normalized iEMG has a limitation toward comparison of
the quantity of muscle activity between groups. However,
the same oxygen consumption for the groups at 45 rpm at
200 W (Fig. 4) supports that it is reasonable to set 45 rpm
as a standard for normalizing iEMGs. As shown inFigure 4,
the normalized iEMG with cyclists were significantly less
than those with noncyclists at higher cadences. This result
and prolonged training experience at higher cadences in
cyclists suggest that our cyclists should also have acquired
skilled performance related to efficient exertion of leg
force by knee extensors, more so than the endurance
Physiological advantage derived from skilled pedaling.
Next, we describe the physiological advantage derived
from such a skilled pedaling at higher cadences. Hagberg
(12) has reported that muscle fatigue due to blood flow
impairment occurs rapidly at contraction levels above 15-
20% MVC. In addition, it has been confirmed that peak
pedal force occurs between 10 and 20% MVC of hip and
knee extensors (seeFig. 3 in (17)) for high cadence
pedaling exercise corresponding to 70 to 100% ˙VO2max.
These two studies suggest that in the present study,
pedaling at 200 W was performed at force output intensity
closely related to the force at which muscle fatigue caused
by occluded blood flow occurs. Also, it would seem that
significantly lower peak pedal force for cyclists by 43.2,
51.5, and 46.9 N at 75, 90, and 105 rpm would positively
contribute to improved blood flow(10) and prevent leg
muscles from fatiguing due to lower intramuscular
pressure (13). Furthermore, it has been confirmed that a
decrease in force output leads to the recruitment of a
higher percentage of ST fibers (1,11), which have higher
oxidative capacity (15) and mechanical efficiency(5,29).
Therefore, the reduced workload for the extensors in our
cyclists might elicit a higher percentage of ST fibers and
result in significantly lower ˙VO2 at 75, 90, and 105 rpm,
relative to noncyclists. These findings prove that pedaling
at a higher cadence actually provides physiological
advantages in muscles, although it is accompanied by a
significant increase in oxygen consumption due to
increased internal work. Each of these findings regarding
muscle fiber recruitment and blood flow can well explain
the difference in the cadences for minimal neuromuscular
fatigue between cyclists and noncyclists(26,27).
Similarity in pedaling skill between the groups. Our result
for muscle activity is well reflected in the difference in
˙VO2 between the groups and between work intensities.
Although ˙VO2 for cyclists tended to be lower than that for
noncyclists, statistical significance was attained only at 105
rpm at 150 W. The reason for the similarity in˙VO2 for the
groups is explained by the relative intensity of work and
the iEMG changing patterns. It is well known that the
acquired kinetic properties of muscle are specifically
effective when the muscle is used at the same contraction
speed at which it had been trained (7). Coyle et al. (4)
reported that highly trained cyclists could continuously
Med Sci Sports Exerc. 1998; 30(3): 442-9
pedal for 1 h at the intensity that required higher than
85%˙VO2max. Our cyclists had not been as well trained as
those reported by Coyle et al. However, work intensity of
150 W that induced˙VO2 ranging from 52.5 to 61.2%
˙VO2max for our cyclists seemed to be much lower than the
intensity required in their daily training. In addition, Figure
3B demonstrates that the iEMG changing patterns for the
knee extensors were the same for the groups despite a
slight difference obtained for BF. These findings suggest
that work intensity of 150 W might be too low to induce
specific performance acquired by training for our cyclists.
Meanwhile, there were no differences in˙VO2 between the
groups at 45 and 60 rpm, although significant differences
in peak pedal forces were obtained at both cadences. The
result for 60 rpm is possibly derived from cycling
experience in daily life for noncyclists and/or lack of
statistical power. As for 45 rpm, it is speculated that the
difference in pedaling skill would be canceled by the
higher rate of muscle activity in the upper extremities and
the trunk muscles for gaining stability of the body on the
bicycle. In addition, the following comment from a subject
summarizes the state of ˙VO2 at 45 rpm, which we
commonly observed in cyclists. Generally we pedal
standing up when we climb hills at such a low cadence.
Therefore, our cyclists would not have acquired the
pedaling skill to decrease peak pedal force during heavy
pedaling while sitting on the saddle.
In conclusion, the differences in both muscle activity for
lower extremities and peak pedal force between the
groups provide experimental evidence that cyclists have
adapted to pedaling exercise at a higher cadence. This is
accomplished by acquiring a certain pedaling technique
for positive utilization of the knee flexors to decrease peak
pedal force and alleviate muscle activity for the knee
extensors. The difference in preferred cadence between
the groups can be well explained by the iEMG changing
patterns in the knee extensors. We speculated that
pedaling skills to decrease muscle stress, which can
influence the preferred cadence selection, contribute to
promoting the recruitment of ST fibers with higher
mechanical efficiency(5,29) and fatigue resistance(15)
despite increase of oxygen consumption. Thus, the
difference in the cadences for minimal neuromuscular
fatigue between cyclists and noncyclists (26,27) can be
The authors gratefully acknowledge the cooperation of Dr.
Yoshifumi Yasuda. We also express our gratitude to Mr. Jeffrey L.
Brown and Mr. Jiro Takai for their careful reading of manuscript.
Address for correspondence: Tetsuo Takaishi, Institute of Natural
Sciences, Nagoya City University, Mizuho-cho, Mizuho-ku, Nagoya
467, Japan. E-mail:email@example.com.
Ahlquist, L. E., D. R. Bassett, R. Sufit, F. J. Nagle, and D. P. Thomas. The effect of pedaling frequency on
glycogen depletion rates in type I and type II quadriceps muscle fibers during submaximal cycling exercise.
Eur. J. Appl. Physiol. 65:360-364, 1992.
BÃ¶ning, D., Y. Gonen, and N. Maassen. Relationship between work load, pedal frequency, and physical
fitness. Int. J. Sports Med. 5:92-97, 1984.
Coast, J. R. and H. G. Welch. Linear increase in optimal pedal rate with increased power output in cycle
ergometry. Eur. J. Appl. Physiol. 53:339-342, 1985.
Coyle, E. F., M. F. Feltner, S. A. Kautz, et al. Physiological and biomechanical factors associated with elite
endurance cycling performance. Med. Sci. Sports Exerc. 23:93-107, 1991.
Coyle, E. F., L. S. Sidossis, J. F. Horowittz, and J. D. Beltz. Cycling efficiency is related to the percentage of
type I muscle fibers. Med. Sci. Sports Exerc. 24:782-788, 1992.
Davis, R. R. and M. L. Hull. Measurement of pedal loading in bicycling: II. analysis and results. J. Biomech.
Duchateau, J. and K. Hainaut. Isometric or dynamic training: differential effects on mechanical properties
of a human muscle.J. Appl. Physiol. 56:296-301, 1984.
Ekblom, B. and A. N. Goldberg. The influence of physical training and other factors on the subjective rating
of perceived exertion.Acta Physiol. Scand. 83:399-406, 1971.
Ericson, M. O., A. Bratt, R. Nisell, U. P. Arborelius, and J. Ekholm. Power output and work in different
muscle groups during ergometer cycling. Eur. J. Appl. Physiol. 55:229-235, 1986.
10. Gotshall, R. W., T. A. Bauer, and S. L. Fahrner. Cycling cadence alters exercise hemodynamics. Int. J. Sports
Med. 17:17-21, 1996.
11. Hagan, R. D., S. E. Weis, and P. B. Raven. Effect of pedal rate on cardio-respiratory responses during
continuous exercise.Med. Sci. Sports Exerc. 24:1088-1095, 1992.
12. Hagberg, J. Muscle endurance and surface electromyogram in isometric and dynamic exercise. J. Appl.
Physiol. 51:1-7, 1981.
13. Hagberg, J. M., J. P. Mullin, M. D. Giese, and E. Spitzagel. Effect of pedaling rate on submaximal exercise
responses of competitive cyclists. J. Appl. Physiol. 51:447-451, 1981.
14. Himann, J. E., D. A. Cunningham., P. A. Rechnitzer, and D. H. Paterson. Age-related changes in speed of
walking. Med. Sci. Sports Exerc. 20:161-166, 1988.
Med Sci Sports Exerc. 1998; 30(3): 442-9 Download full-text
15. Ivy, J. L., R. T. Withers, P. J. Van Handel, D. H. Elger, and D. L. Costill. Muscle respiratory capacity and fiber
type as determinants of the lactate threshold. J. Appl. Physiol. 48:523-527, 1980.
16. Jorge, M. and M. L. Hull. Analysis of EMG measurements during bicycle pedalling. J. Biomech. 19:683-694,
17. LÃ¶llgen, H., T. Graham, and G. Sjogaard. Muscle metabolites, force, and perceived exertion bicycling at
varying pedal rates.Med. Sci. Sports Exerc. 12:345-351, 1980.
18. Marsh A. P. and P. E. Martin. The association between cycling experience and preferred and most
economical cadences. Med. Sci. Sports Exerc. 25:1269-1274, 1993.
19. Marsh, A. P. and P. E. Martin. The relationship between cadence and lower extremity EMG in cyclists and
noncyclists. Med. Sci. Sports Exerc. 27:217-225, 1995.
20. McCartney, N., G. J. F. Heigenhauser, and N. L. Jones. Power output and fatigue of human muscle in
maximal cycling exercise.J. Appl. Physiol. 55:218-224, 1983.
21. Nakamura, Y., Y. Mutoh, and M. Miyashita. Determination of the peak power output during maximal brief
pedalling bouts. J. Sports Sci. 3:181-187, 1985.
22. Pandolf, K. B. and B. J. Noble. The effect of pedaling speed and resistance changes on perceived exertion
for equivalent power outputs on the bicycle ergometer. Med. Sci. Sports Exerc. 5:132-136, 1973.
23. Patterson, R. P. and M. I. Moreno. Bicycle pedaling forces as a function of cadence and power output.
Med. Sci. Sports Exerc. 22:512-526, 1990.
24. Ryan, M. M. and R. J. Gregor. EMG profiles of lower extremity muscles during cycling at constant workload
and cadence. J. Elctromyogr. Kinsiol. 2:69-80, 1992.
25. Seabury, J. J., W. C. Adams, and M. R. Ramey. Influence of cadence and power output on energy
expenditure during bicycle ergometry.Ergonomics 20:491-498, 1977.
26. Takaishi, T., Y. Yasuda, and T. Moritani. Neuromuscular fatigue during prolonged pedaling exercise at
different cadences. Eur. J. Appl. Physiol. 69:154-158, 1994.
27. Takaishi, T., Y. Yasuda, T. Ono, and T. Moritani. Optimal pedaling rate estimated from neuromuscular
fatigue for cyclists.Med. Sci. Sports Exerc. 28:1492-1497, 1996.
28. Wells, R., M. Morrissey, and R. Hughson. Internal work and physiological responses during concentric and
eccentric cycle ergometer.Eur. J. Appl. Physiol. 55:295-301, 1986.
29. Wendt, I. R. and C. L. Gibbs. Energy production of rat extensor digitorum longus muscle. Am. J. Physiol.
30. Zarrugh, M. Y. and C. W. Radcliffe. Predicted metabolic cost of level walking. Eur. J. Appl. Physiol. 55:295-