Comparison of oxygen uptake kinetics during concentric and



Comparison of oxygen uptake kinetics during concentric and
J Appl Physiol
91: 2135–2142, 2001.
Comparison of oxygen uptake kinetics during concentric
and eccentric cycle exercise
Department of Kinesiology, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1;
Laboratoire des Sciences du Sport, Unité de Formation et de Recherche en Sciences et Techniques
des Activités Physiques et Sportives, 25030 Besançon Cedex; and 3Faculté des Sciences du Sport,
Université de Montpellier, 34090 Montpellier Cedex, France
Received 13 February 2001; accepted in final form 25 June 2001
muscle action; muscle fatigue; cycling; surface electromyography
THERE ARE CONFLICTING RESULTS about whether the rate of
adaptation of O2 uptake (V̇O2) at the onset of exercise is
dependent on relative work rate and the type of muscular action. Although some investigators have shown
adaptation of V̇O2 to be progressively slowed as the
work rate increases (11, 18, 22, 31), others have shown
a constancy at least until the work rate exceeds the
ventilatory threshold (VT) (15) or even above this work
rate (2). In the case of high work rates above VT but
below maximal V̇O2, an additional slow component is
Address for reprint requests and other correspondence: R. L.
Hughson, Dept. of Kinesiology, University of Waterloo, Waterloo,
ON, Canada N2L 3G1 (E-mail: [email protected]).
observed, normally developing ⱖ90 s after the start of
exercise (2, 31). There have been very few investigations of the kinetics of V̇O2 during eccentric compared
with concentric muscular actions. In their first study,
Knuttgen and Klausen (25) observed a very small or no
O2 deficit at the onset of eccentric exercise. In a subsequent investigation, Bonde-Petersen et al. (7) found
that the O2 deficit was not different from that incurred
when the same absolute V̇O2 was achieved during concentric cycling. However, these early studies were conducted with Douglas bag collections and without detailed analysis of the components of the dynamic
In eccentric exercise, the muscles are forcibly lengthened while activated. Walking and running downhill include large components of eccentric exercise, whereas
opposing the torque transmitted from a motor to the
pedals of a cycle ergometer may be a more “pure” form of
eccentric exercise (32). It is possible that the kinetics of
V̇O2 might differ between cycling and running, in part
because of different muscle recruitment patterns, including the proportion of eccentric vs. concentric muscle activation (10, 23). The physiological cost of eccentric muscle
activation is substantially less than that of concentric
muscle contraction (1, 24, 25, 32, 37), inasmuch as fewer
muscle fibers are recruited to perform the same work rate
during eccentric cycling (1, 5).
Electromyographic (EMG) activity has been studied
in conjunction with investigations of the V̇O2 slow component to determine whether there is an increase in
integrated EMG (iEMG) or in mean power frequency
(MPF). The increase in iEMG might reflect greater
total muscle fiber recruitment (36) as fibers fatigue.
Changes in MPF might indicate recruitment of a
greater proportion of fast-twitch fibers (17, 35).
Greater recruitment of fast-twitch fibers with time
during heavy exercise has been suggested to accompany the onset of the V̇O2 slow component, although
the results of research are contradictory (27, 35, 36).
The purpose of the present study was to investigate
the muscle activity patterns and the V̇O2 kinetics durThe costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
8750-7587/01 $5.00 Copyright © 2001 the American Physiological Society
Downloaded from by on September 29, 2016
Perrey, S., A. Betik, R. Candau, J. D. Rouillon, and
R. L. Hughson. Comparison of oxygen uptake kinetics during concentric and eccentric cycle exercise. J Appl Physiol 91:
2135–2142, 2001.—O2 uptake (V̇O2) kinetics and electromyographic (EMG) activity from the vastus medialis, rectus femoris, biceps femoris, and medial gastrocnemius muscles were
studied during constant-load concentric and eccentric cycling. Six healthy men performed transitions from baseline to
high-intensity eccentric (HE) exercise and to high-intensity
(HC), moderate-intensity (MC), and low-intensity (LC) concentric exercise. For HE and HC exercise, absolute work rate
was equivalent. For HE and LC exercise, V̇O2 was equivalent.
V̇O2 data were fit by a two- or three-component exponential
model. Surface EMG was recorded during the last 12 s of
each minute of exercise to obtain integrated EMG and mean
power frequency. Only in the HC exercise did V̇O2 increase
progressively with evidence of a slow component (phase 3),
and only in HC exercise was there evidence of a coincident
increase with time in integrated EMG of the vastus medialis
and rectus femoris muscles (P ⬍ 0.05) with no change in
mean power frequency. The phase 2 time constant was slower
in HC [24.0 ⫾ 1.7 (SE) s] than in HE (14.7 ⫾ 2.8 s) and LC
(16.7 ⫾ 2.2 s) exercise, while it was not different from MC
exercise (20.6 ⫾ 2.1 s). These results show that the rate of
increase in V̇O2 at the onset of exercise was not different
between HE and LC exercise, where the metabolic demand
was similar, but both had significantly faster kinetics for V̇O2
than HC exercise. The V̇O2 slow component might be related
to increased muscle activation, which is a function of metabolic demand and not absolute work rate.
ing concentric and eccentric cycle ergometry at the
same absolute and relative work rates. We hypothesized that, for a given mechanical work rate, the increase in V̇O2 would be more rapid during eccentric
exercise, because the total metabolic requirement was
reduced. Comparison of the V̇O2 kinetics between eccentric and concentric exercise that had the same metabolic requirement was expected to show the same rate
of increase to the steady state. We further hypothesized that a V̇O2 slow component would be observed
only when there was an increase in iEMG reflecting
greater motor unit recruitment as exercise continued.
J Appl Physiol • VOL
91 • NOVEMBER 2001 •
Downloaded from by on September 29, 2016
Subjects. Six healthy, well-motivated men volunteered to
participate in the study after they were informed of the
nature and possible inconveniences associated with the experiment. Each subject gave informed written consent on a
form approved by the Office of Research Ethics at the University of Waterloo. Subjects’ age, weight, and height were
(mean ⫾ SE) 25 ⫾ 1 yr, 74.2 ⫾ 2.8 kg, and 180 ⫾ 3 cm,
respectively. Subjects were physically active in a variety of
sports (running, cycling, and cross-country skiing) a minimum of four times per week.
Experimental protocol. Preliminary testing of all subjects
consisted of an incremental concentric exercise test to exhaustion to determine the VT and peak V̇O2. The initial work
rate of 60 W was increased by 30 W/min at a cycling frequency of 60 rpm. The VT was estimated by using a five-point
filter of V̇O2 and CO2 output (V̇CO2) to reduce breath-tobreath variability before application of a modified V-slope
method (21). V̇CO2 was examined as a function of V̇O2 under
the assumption that the VT corresponds to the break point in
the V̇CO2-V̇O2 relationship. The break point was selected by a
computer program that determined the V̇O2 above which
V̇CO2 increased faster than V̇O2, without hyperventilation.
Peak V̇O2 was taken as the average of the highest five
consecutive breaths attained in the last minute of exercise.
The VT and peak V̇O2 were used in choosing the individual
work rates for the step tests in concentric and eccentric
For the concentric bouts, the design of the study involved
testing below and above the VT with step changes in work
rate. On any test day, subjects completed at least two different step work rate tests. Always, the lowest intensities were
performed first. Thus, on a test day, a subject might perform
a moderate-intensity followed by a high-intensity concentric
exercise. Data collection commenced with 4 min at 15 W to
establish a baseline, then the subjects performed 6-min bouts
of higher constant-load exercise. The three different work
rates selected for concentric cycling were defined as light
concentric (LC), moderate concentric (MC), and heavy concentric (HC) exercise. The LC intensity was selected to
achieve a steady-state V̇O2 similar to that measured during
the eccentric exercise. Thus it was necessary to conduct the
LC tests on separate days after the eccentric tests. The MC
work rate elicited V̇O2 corresponding to ⬃90% of the VT. No
slow component rise in V̇O2 occurred at this intensity. The
HC exercise was performed at a work rate estimated to
require V̇O2 equal to ⬃70% of the difference (⌬) between the
subject’s VT and peak V̇O2, i.e., a value of VT ⫹ 0.7⌬.
The high eccentric (HE) exercise intensity was at the same
absolute work rate as the HC exercise and was preceded by 4
min at baseline (15 W) pedaling. During pilot tests with three
subjects, the peak work rate that could safely be achieved did
not elicit a noticeable increase in V̇CO2 relative to V̇O2, so it
was impossible to select the VT. At the highest exercise
intensities achieved in the incremental tests (⬎400 W), there
is a possibility of severe tissue damage and orthopedic problems with eccentric muscle actions (26). Three ⱖ20- to 30min training sessions on the eccentric ergometer were completed before the HE test sessions to minimize muscle
To reduce the noise in breath-by-breath data and enhance
the underlying response patterns, each subject performed
several repetitions of the exercise transitions under each
condition. Exercise transitions were performed three times
for the concentric bouts and four times for the eccentric
The order in which the conditions were performed was
randomized, except for LC exercise as noted. At least 2 days
separated an eccentric test from another test, either eccentric
or concentric.
Venous blood was drawn from a catheter inserted in the
back of the hand into heparinized syringes at rest and before
the onset and at the last minute of MC and HC exercise. The
samples were then placed in ice and analyzed after a short
delay. Whole blood samples were analyzed (at 37°C) for
plasma lactate concentration (StatProfile 9 Plus Blood GasElectrolyte Analyzer, Nova Biomedical Canada). Blood was
not sampled during the HE exercise in all subjects, inasmuch
as samples from the first two subjects confirmed previous
observations with the same ergometer (33) that lactate was
unaffected at the low metabolic demand of this exercise.
At the end of each exercise step, subjects were asked to
rate their perceived exertion (RPE) using the Borg scale,
ranging from 0 (nothing at all) to 10 (very, very heavy) (8).
Breath-by-breath gas exchange. Breath-by-breath ventilation and gas exchange were measured by using a portable
measurement system (Cosmed K4 b2, Rome, Italy). This system
consists of a facemask with a low-dead-space (70 ml) and a
low-resistance, bidirectional digital turbine (28-mm diameter) that was calibrated before each test with a syringe of
known volume (3,007 ml). Expired gases were sampled at the
mouth by an O2 analyzer with a polarographic electrode and
a CO2 analyzer with an infrared electrode. The O2 and CO2
analyzers were calibrated by using ambient air and precision-measured reference gas at the beginning of each session
according to procedures recommended by the manufacturer.
Heart rate (HR) was continuously calculated with each
ventilatory data point via a wireless Polar monitoring system.
Concentric and eccentric ergometers. The concentric exercise was performed on an electrically braked cycle ergometer
(Excalibur Sport, Lode, Groningen, The Netherlands). Seat
and handlebar positions were kept constant for individual
subjects during the study.
The eccentric test and training exercises were performed
on a standard Monark cycle ergometer modified for eccentric
exercise (33, 37). The pedals were driven in a reverse direction from normal cycling by an electric motor. The subjects
were instructed to keep the pedaling rate at 60 rpm and thus
to resist the tendency of the ergometer to increase the pedal
axle from 60 rpm. For this type of ergometer, the transition
from the baseline level to step work rate levels was fairly
rapid (typically ⬃3 s) and was followed by a fine tuning.
Because eccentric exercise on a cycle ergometer is not without risk due to high force development, an investigator was
constantly prepared to activate a brake and press an emergency stop button. There were no emergencies.
To minimize any extra energy expenditure required during
eccentric cycling as a function of movements to balance and
V̇ O 2 ⫽ A0 ⫹ A1关1 ⫺ e⫺共t ⫺ TD1兲/␶1兴 䡠 u1 ⫹ A2关1 ⫺ e⫺共t ⫺ TD2兲/␶2兴 䡠 u2
where u1 ⫽ 0 for t ⬍ TD1 and u1 ⫽ 1 for t ⱖ TD1, and u2 ⫽ 0
for t ⬍ TD2 and u2 ⫽ 1 for t ⱖ TD2.
J Appl Physiol • VOL
Data from the HC exercise were fit by a three-component
exponential model that had an additional component after
V̇ O 2 ⫽ A0 ⫹ A1关1 ⫺ e⫺共t ⫺ TD1兲/␶1兴 䡠 u1 ⫹ A2关1 ⫺ e⫺共t ⫺ TD2兲/␶2兴 䡠 u2
⫹ A3关1 ⫺ e⫺共t ⫺ TD3兲/␶3兴 䡠 u3
where u1 and u2 have the same meaning, and u3 ⫽ 0 for t ⬍
TD3 and u3 ⫽ 1 for t ⱖ TD3.
The overall time course of the response was determined
from the mean response time (MRT), which is calculated
from a weighted sum of TD and ␶ for each component (28).
MRT is equivalent to the time required to achieve ⬃63% of
the difference between A0 and the final V̇O2.
O2 deficit (O2Def) was calculated for exercise without evidence of a slow component as follows: O2Def ⫽ (A1 ⫹
A2) 䡠 MRT (38).
HR response was estimated during HE and LC exercise as
the equivalent to MRT from the time to achieve 63% of the
difference between the baseline and the steady-state or endexercise HR. In HE exercise, there was often an overshoot
before steady state.
Statistics. A one-way repeated-measures analysis of variance with multiple comparisons (Student-Newman-Keuls
post hoc test) was used to compare the differences in variables among conditions. A one-way repeated-measures analysis of variance was also used to determine the overall effect
of time on iEMG data with respect to the normalized baseline
at t ⫽ 1. The Friedman rank test was used for variables with
nonnormal distribution or unequal variance. Statistical significance was set at P ⬍ 0.05, and values are means ⫾ SE.
The peak V̇O2 achieved by the subjects during the
preliminary incremental exercise test was 59.7 ⫾ 1.4
ml 䡠 kg⫺1 䡠 min⫺1. The work rates utilized during the
constant-load tests were 207 ⫾ 11 W for MC exercise
and 317 ⫾ 14 W for HC and HE exercise. The work rate
required in LC exercise to achieve approximately the
same metabolic cost as in HE exercise was 62 ⫾ 7 W
(range 40–88 W). Steady-state V̇O2 was not different
between LC and HE exercise (1,212 ⫾ 79 and 1,169 ⫾
103 ml/min, respectively).
V̇O2 kinetics. The principal results of V̇O2 kinetics for
the concentric and eccentric exercises are presented in
Table 1, with an example of the V̇O2 response for a
single subject during each condition in Fig. 1. For
eccentric exercise, the V̇O2 response was best described
with two exponential terms because of an absence of
the slow component rise in V̇O2. During the baseline
period, V̇O2 (A0) was significantly lower in HE than in
the three concentric exercises (P ⬍ 0.05). The onset of
phase 2 (TD2) was significantly earlier in HC than in
any other exercise, whereas TD2 in HE exercise occurred earlier than in LC exercise (Table 1). The phase
2 time constant (␶2) was slower in HC than in HE and
LC exercise but was not different from MC exercise.
HC exercise resulted in higher amplitudes (A1 and A2)
than the other exercises, as did MC exercise compared
with LC and HE exercises. A2 was smaller in LC than
in HE exercise. HC exercise was characterized by a V̇O2
slow component at 123.7 ⫾ 3.5 s after exercise onset
(TD3, Table 1). Hence, HC exercise resulted in a slower
91 • NOVEMBER 2001 •
Downloaded from by on September 29, 2016
fix the body, we chose to use a rigid chair, rather than a
standard bicycle saddle, to support the subjects. The chair
was equipped with a firm back support, allowing maximum
relaxation of all nonworking muscles. This enabled the subjects to resist the heavy exercise loads more easily. The
subjects were seated with a hip angle of 125°. The subjects’
feet were positioned with adjustable rubber straps on pedals
placed so that the center of rotation of the ankle joint was
close to the axis of the pedal shaft. This minimized contractions of the gastrocnemius (GA) muscles.
EMG on leg muscles. EMG data of the rectus femoris (RF),
vastus medialis (VM), biceps femoris, and GA muscles were
collected twice for each condition on the right leg by using
bipolar surface Ag-AgCl electrodes with a surface cup diameter of 10 mm (200 Medi-Trace Mini, Graphics Control,
Buffalo, NY), with an intrapair distance of 35 mm. A common
reference electrode was placed on the tibial tuberosity of the
same leg. Before electrode placement, the skin areas were
shaved and cleaned with isopropyl alcohol swabs to reduce
skin impedance. Electrodes were always placed on the most
prominent site of the muscle belly as noted by visualization
and palpation by the same investigator. They were placed in
the same direction as the muscle fibers. Electrode wires then
were taped to the skin to reduce movement artifacts.
The myoelectrical signal was band-pass filtered (20–500
Hz), differentially amplified (gain 1,000 times, input impedance 2 M⍀, common mode rejection ratio ⬎90 dB), and
sampled at 1,024 Hz with a 12-bit analog-to-digital conversion board (DAS-16, MetraByte, Taunton, MA) for the last
12 s of each minute during exercise. After removal of any
bias, raw EMG signals were full-wave rectified and low-pass
filtered with a second-order Butterworth filter (cutoff frequency ⫽ 3 Hz) (39), producing a linear envelope. The resulting smoothed EMG signal was integrated (iEMG) over 10 s,
generating an indication of the total muscle activity of the
exercising muscle at each minute of exercise.
The digitized EMG was also processed by using a Hamming window function and 2,048-point fast Fourier transform to obtain mean power frequency (MPF) over 10 s.
Surface EMG signals present extensive interindividual variance because of electrode position or impedance of underlying
tissues. Therefore, iEMG and MPF were expressed relative
to the value corresponding to the 1st min of exercise (nonfatigued state). This technique does not allow for comparison
between test sessions but does allow for reliable tracking of
changes within a test.
Data analysis. Breath-by-breath data for V̇O2 from at least
three repetitions of an identical test condition were linearly
interpolated at 1-s intervals, time aligned, and ensemble
averaged to provide a single response for each subject. Two
mathematical models were employed to characterize the average-response curves using least-squares nonlinear regression techniques, in which the best fit was defined by minimization of the residual sum of squares. The model utilized to
describe the kinetic response provides an estimate of the
baseline (A0), amplitude terms (A1, A2, and A3), time constants (␶1, ␶2, and ␶3), and time delays (TD1, TD2, and TD3).
Data from the concentric tests below VT (LC and MC) and
eccentric tests (HE) were fit by a two-component exponential
Table 1. V̇O2 kinetic parameters during concentric
and eccentric exercise
A0, ml/min
A1, ml/min
TD1, s
␶1, s
A2, ml/min
TD2, s
␶2, s
A3, ml/min
TD3, s
␶3, s
O2Def, liters
MRT, s
837 ⫾ 40
818 ⫾ 48
704 ⫾ 70
542 ⫾ 38*
1.1 ⫾ 0.5
0.9 ⫾ 0.6
5.8 ⫾ 1.0
6.8 ⫾ 0.5
2,055 ⫾ 158 1,384 ⫾ 84*
16.4 ⫾ 0.7
21.4 ⫾ 0.2*
24.0 ⫾ 1.7
20.6 ⫾ 2.1
525 ⫾ 76
123.7 ⫾ 3.5
103.1 ⫾ 10.8
1.03 ⫾ 0.07
63.2 ⫾ 4.5
32.4 ⫾ 1.7*
782 ⫾ 30
120 ⫾ 27*†
0.6 ⫾ 0.3
5.5 ⫾ 0.6
310 ⫾ 46*†
22.1 ⫾ 0.6*
16.7 ⫾ 2.2*
418 ⫾ 21*†‡
223 ⫾ 43*†
⫺0.9 ⫾ 0.3*†‡
6.7 ⫾ 0.9
528 ⫾ 70*†
19.3 ⫾ 1.3*‡
14.7 ⫾ 2.8*
0.21 ⫾ 0.04†
29.8 ⫾ 2.2*
0.32 ⫾ 0.05†
26.4 ⫾ 2.8*
overall V̇O2 response than the other exercises, as indicated by a slower MRT and the presence of ␶3 (Table 1).
The O2 deficit was not different between HE and LC
conditions but was higher in MC exercise (Table 1).
HR kinetics. HR response (Fig. 2) was examined in
the two exercise conditions that elicited similar steadystate V̇O2 (i.e., LC and HE). There was a significantly
lower baseline HR and a greater increase in HR above
baseline during HE than during LC exercise (Table 2).
In HE exercise, there was a slight overshoot (⬃6 beats/
min) in the first seconds after the onset of exercise (Fig.
2, Table 2). The time to achieve 63% of the increase in
HR above baseline was significantly less during HE
than during LC exercise (Table 2). HR kinetics data for
HC exercise are not presented, inasmuch as there was
not a steady end-exercise value.
Fig. 2. Heart rate during final 2 min of baseline exercise at 15 W and
during 6 min of LC (E) and HE exercise (Œ). Values are averages of 6
subjects; see Table 2 for SE. bpm, Beats/min.
EMG data. The normalized iEMG was constant
across time periods for all muscles during each of the
MC and HE exercise tests (data for VM and RF muscles are shown in Fig. 3). In the HC tests, iEMG
increased significantly at minutes 5 and 6 from the
values at minutes 1, 2, and 3 in the VM and RF muscles
(Fig. 3), but not in the biceps femoris and GA, muscles.
The increase in iEMG between minutes 3 and 6 of the
HC exercise tests averaged 20 and 30% in VM and RF
muscles, respectively. There were no changes in MPF
as a function of time for any muscle during any of the
exercise tests (data not shown).
RPE and lactate. Significant differences in RPE were
noted among exercise conditions (Table 3), with the
highest RPE observed during HC exercise followed by
HE, MC, and LC exercise (each different from one
another, P ⬍ 0.05). Blood lactate concentration increased more at the end of HC than MC exercise
(Table 3).
We have found that the time course for the increase
in V̇O2 at the onset of eccentric exercise (HE), indicated
by ␶2, is similar to that for concentric exercise at the
same metabolic demand (LC). On the other hand, ␶2 for
V̇O2 was slower for HC than for HE or LC exercise.
Measurements of muscle activation patterns by norTable 2. Selected parameters for kinetic analysis of
HR during HE and LC exercises
Fig. 1. O2 uptake (V̇O2) for 1 subject during final 2 min of baseline
exercise at 15 W during heavy-intensity (HC, 330 W), moderateintensity (MC, 216 W), and low-intensity (LC, 70 W) concentric
exercise, and during high-intensity eccentric exercise (HE, 330 W)
for 6 min. Breath-by-breath data have been interpolated to 1-s
intervals and averaged across 3–4 repetitions for each exercise
J Appl Physiol • VOL
HR baseline, beats/min
Overshoot, beats/min
HR increase, beats/min
Time to 63%, s
77.9 ⫾ 2.6
14.6 ⫾ 2.4
24.9 ⫾ 2.2
70.1 ⫾ 2.0*
6.1 ⫾ 2.4
32.5 ⫾ 2.0*
13.8 ⫾ 1.3*
Values are means ⫾ SE; n ⫽ 6 subjects. HR, heart rate; LC, light
concentric; HE, high eccentric. * Significantly different from LC, P ⬍
91 • NOVEMBER 2001 •
Downloaded from by on September 29, 2016
Values are means ⫾ SE; n ⫽ 6 subjects. V̇O2, O2 uptake; A,
amplitudes of response; ␶, time constants; TD, time delay; MRT,
mean response time; O2Def, O2 deficit; NA, not available. * Significantly different from heavy concentric, P ⬍ 0.05. † Significantly
different from moderate concentric, P ⬍ 0.05. ‡ Significantly different from light concentric, P ⬍ 0.05.
malized iEMG demonstrated a significant increase
with time only for HC exercise where there was also
the appearance of a phase 3 response in V̇O2. There was
no change in iEMG with time during HE exercise, and
there was a phase 3 component of V̇O2, even though the
absolute work rate was the same as in HC exercise.
These results provide new information about the kinetics of V̇O2 during eccentric patterns of muscle activation. In an early study of eccentric cycling, Knuttgen
and Klausen (25) described a very rapid increase in V̇O2
that appeared to proceed immediately to steady state
without development of an O2 deficit. A subsequent
study from that group (7) in which O2 deficit measured
during HE exercise did not differ from that during
concentric exercise at the same metabolic rate is consistent with our finding of no difference in O2 deficit or
␶2 between LC and HE exercise. Our observation of
faster ␶2 in LC and HE than in HC exercise suggests
that achievement of more rapid adaptation at a lower
J Appl Physiol • VOL
Table 3. RPE and increase in lactate during
concentric and eccentric exercise
⌬Lac, mmol/l
9.0 ⫾ 0.8†
8.2 ⫾ 1.2
3.8 ⫾ 0.6†
2.1 ⫾ 0.4*
0.75 ⫾ 0.3†
5.7 ⫾ 0.4†
Values are means ⫾ SE; n ⫽ 6 subjects. RPE, rating of perceived
exertion; ⌬Lac, increase in lactate. * Significantly different from HC,
P ⬍ 0.05. † Significantly different from each other P ⬍ 0.05.
91 • NOVEMBER 2001 •
Downloaded from by on September 29, 2016
Fig. 3. Evolution of integrated electromyogram (iEMG) as a function
of time during MC, HC, and HE exercise for the rectus femoris (A)
and vastus medialis (B) muscles. Values are means ⫾ SE for 6
subjects obtained in 2 repetitions of each test normalized to the
individual values at 1 min under that test condition. *Different from
minutes 1, 2, and 3, P ⬍ 0.05.
metabolic demand is in agreement with some studies
(15, 31) but in contrast with others that reported no
difference in ␶2 across a wide range of work rates (2,
Eccentric exercise. Resisting the lengthening of a
muscle as force is applied is a common part of daily
activities such as normal walking, but it is especially
important during downhill walking or running (32). In
this study, we were able to focus on the eccentric
muscle action by application of force to the muscle as a
motor-driven cycle ergometer pedals backward. However, the maximum force that can be generated during
eccentric contractions is sufficiently high that muscle
damage can occur (26). Indeed, even moderate intensities of eccentric exercise will cause muscle soreness in
subjects unaccustomed to this type of exercise. Thus we
introduced training sessions to allow our subjects to
adapt and to avoid subsequent pain. It is possible that
the reason we observed no progressive increase in V̇O2
(the V̇O2 slow component, see below) during the HE
tests was that our subjects did not experience any
fatigue. The constant iEMG and MPF with time during
HE exercise are consistent with the absence of fatigue
during HE exercise. Eccentric exercise is accomplished
with lower muscle activation (5, 30) and a pattern of
activation (16) that is different from that of an equivalent concentric workload. The considerably lower
baseline values and the smaller increase in V̇O2 for a
given change in applied work rate during HE exercise
than during any of the concentric exercise tests indicate the greater efficiency of eccentric exercise. The
gain calculated as the change in V̇O2 per change in
work rate was ⬃2.4 ml 䡠 min⫺1 䡠 W⫺1 for the eccentric
exercise compared with 9–10 ml 䡠 min⫺1 䡠 W⫺1 for the
concentric exercise. The reduced effort to accomplish
the same absolute work rate with eccentric exercise
was reflected in the lower scores for RPE with HE than
with HC exercise. However, the higher RPE and HR
(see below) at the same metabolic rate for HE than for
LC exercise probably indicates the greater stimulation
of peripheral mechanoreceptors during eccentric exercise (20).
Incremental eccentric exercise was performed by
three of our subjects. The peak V̇O2 attained during
these tests was only 26–45% of peak during concentric
cycling to exhaustion, and none of the subjects displayed evidence of reaching VT. Safety considerations
and the ability to generate muscular power, rather
J Appl Physiol • VOL
our subjects also completed MC tests. The rationale for
this was that the MC work rate was more typical of the
demand investigated in many other studies, and the ␶2
and MRT values were consistent with previous findings (15, 28). There was clearly a trend for ␶2 to increase from LC (16.7 ⫾ 2.2 s) to MC (20.6 ⫾ 2.1 s) and
HC (24.0 ⫾ 1.7 s) exercise. The lack of significant
difference between each of these mean values might
have been a consequence of insufficient sample size to
isolate the effect. Engelen et al. (15) also suggested
that a small sample size might explain why some
investigators reported slower V̇O2 kinetics above than
below VT (31), whereas others did not (2). In recent
investigations, values for ␶2 that were 24 (10) and 40%
(35) slower in above-VT than in below-VT exercise
were reported as not statistically significant. The small
sample size (n ⫽ 7 in each study) and some betweensubject variation probably precluded detection of significance. The differences between studies highlight
the need for future investigations to be conducted carefully and with sufficient statistical power to isolate
physiologically significant differences in the time
course of adaptation of V̇O2 at the onset of exercise.
V̇O2 slow component in HC. During HC exercise, a
three-component exponential model was required to
achieve a satisfactory description of the V̇O2 kinetics.
This observation is consistent with the concept of the
V̇O2 slow component, in which there appears to be an
additional metabolic cost in addition to the expected
V̇O2 or that work rate (2, 31). The present experiments
show that this added cost is not a function of the work
rate per se, because there was no slow component in
the HE exercise. A potential explanation for the V̇O2
slow component is the metabolic cost of recruiting
additional muscle fibers to accomplish the work task as
some of the fibers initially recruited to complete the
exercise task fatigue. This explanation differs from the
proposal that recruitment of fast-twitch fibers to accomplish the greater work rate caused a higher metabolic cost (3), inasmuch as the newly recruited fibers
could be slow or fast twitch. Although many authors
suggest that it must be additional fast-twitch fibers
that contribute to the V̇O2 slow component, the rationale for this is not clear. The study of Crow and
Kushmerick (13), in which it was shown that mouse
fast-twitch fibers had a greater metabolic requirement
than slow-twitch fibers, is frequently cited as rationale
for the V̇O2 slow component (4, 9, 34). However, differences in metabolic cost for tension development in
human fast- vs. slow-twitch muscle are not clear. Recent in vitro studies of human muscle indicated a
greater metabolic cost for tension development in fasttwitch fibers (19). Coyle et al. (12) reported greater
efficiency during cycling across a range of constant
loads in trained cyclists with a higher percentage of
slow-twitch fibers. In contrast, Barstow et al. (4) found
during ramp incremental cycling that the metabolic
cost might actually be less in subjects with a high
percentage of fast-twitch fibers than in those with a
high percentage of slow-twitch fibers (9 vs. 11
ml 䡠 min⫺1 䡠 W⫺1). The same researchers reported that
91 • NOVEMBER 2001 •
Downloaded from by on September 29, 2016
than a cardiovascular limitation, determined the end
points in the tests.
During exercise, such as running, that involves eccentric and concentric muscle actions, it appears that
the magnitude of the V̇O2 slow component is reduced
when the proportion of concentric exercise is reduced
(6, 23). Our results that showed no V̇O2 slow component
during HE compared with HC cycling exercise might
provide a partial explanation for these observations
during running, as previously suggested (10, 23).
V̇O2 and HR kinetics in concentric and eccentric exercise. The two-component model adequately described
V̇O2 kinetics, except for the HC exercise, where a third
component was necessary. The amplitude of the phase
1 component was graded according to the total metabolic cost of the exercise. The onset of phase 2 (TD2)
occurred first in HC exercise, but phase 2 for HE
exercise also occurred before that for LC exercise. Thus
it appears that the magnitude of the muscle tension
influenced the return of deoxygenated blood to the
lungs. In the case of HE compared with LC exercise,
not only was there greater muscle tension and, therefore, the potential to cause a greater mechanical displacement of blood back to the heart, but also the HR
adapted more rapidly. Although we do not have information on cardiac stroke volume in the transition to
exercise, it is likely that stroke volume was maintained
at or even above baseline levels at the start of HE
exercise. Thus the more rapid increase in HR, where
there was even an overshoot during HE exercise, probably reduced the time required for blood to go to and
return from the working muscle, facilitating a shorter
time to the onset of phase 2. Because of the nonlinear
nature of HR control across the range of work rates
investigated, we did not attempt to fit exponential
curves to these data. The absolute HR was higher
during HE than during LC exercise, even though the
metabolic demand was similar. Similar results were
reported for downhill vs. uphill treadmill walking (14).
Greater muscle tension during HE than during LC
exercise might be expected to provide greater stimulus
to control of HR through activation of the motor cortex
as well as peripheral mechanoreceptor feedback.
The observation that ␶2 of the V̇O2 response was
significantly faster in HE and LC exercise than in the
HC test provides some interesting new data for a
long-standing debate about the effect of work rate on
V̇O2 kinetics. Progressive slowing of V̇O2 kinetics has
been reported across a range of work rates from very
light to near maximal (11, 18). However, these studies
reported results on the basis of a one-component model
rather than allowing for separate isolation of phases 1,
2, and 3 as required. When Casaburi et al. (11) explored the effect of different fitting models on their
results, they concluded that the early-phase response
did not differ substantially between low- and highintensity exercise. There are other studies in which
significant differences were reported for ␶2 across a
range of work rates, especially below vs. above VT (15,
31), as we found. As part of this experiment, which was
designed to focus on HE vs. HC and HE vs. LC exercise,
J Appl Physiol • VOL
The authors gratefully acknowledge the technical assistance of
Dave Northey and thank B. J. Whipp for discussions on the eccentric
exercise model.
This project was supported in part by the Natural Sciences and
Engineering Research Council of Canada and a grant to S. Perrey
from the Franche-Comté Regional Council.
1. Abott BC, Bigland B, and Ritchie JM. The physiological cost
of negative work. J Physiol (Lond) 119: 380–390, 1952.
2. Barstow TJ, Casaburi R, and Wasserman K. O2 uptake
kinetics and the O2 deficit as related to exercise intensity and
blood lactate. J Appl Physiol 75: 755–762, 1993.
3. Barstow TJ, Jones AM, Nguyen PH, and Casaburi R. Influence of muscle fiber type and pedal frequency on oxygen
uptake kinetics of heavy exercise. J Appl Physiol 81: 1642–1650,
4. Barstow TJ, Jones AM, Nguyen PH, and Casaburi R. Influence of muscle fibre type and fitness on the oxygen uptake/
power output slope during incremental exercise in humans. Exp
Physiol 85: 109–116, 2000.
5. Bigland-Ritchie B and Woods JJ. Integrated electromyogram
and oxygen uptake during positive and negative work. J Physiol
(Lond) 260: 267–277, 1976.
6. Billat VL, Richard R, Binsse VM, Koralsztein JP, and
Haouzi P. The V̇O2 slow component for severe exercise depends
on type of exercise and is not correlated with time to fatigue.
J Appl Physiol 85: 2118–2124, 1998.
7. Bonde-Petersen F, Knuttgen H, and Henrickson J. Muscle
metabolism during exercise with concentric and eccentric contractions. J Appl Physiol 33: 792–795, 1972.
8. Borg G, Ljunggren G, and Ceci R. The increase of perceived
exertion, aches and pain in the legs, heart rate and blood lactate
during exercise on a bicycle ergometer. Eur J Appl Physiol 54:
343–349, 1985.
9. Burnley M, Jones AM, Carter H, and Doust JH. Effects of
prior heavy exercise on phase II pulmonary oxygen uptake kinetics during heavy exercise. J Appl Physiol 89: 1387–1396,
10. Carter H, Jones AM, Barstow TJ, Burnley M, Williams CA,
and Doust JH. Oxygen uptake kinetics in treadmill running
and cycle ergometry: a comparison. J Appl Physiol 89: 899–907,
11. Casaburi R, Barstow TJ, Robinson T, and Wasserman K.
Influence of work rate on ventilatory and gas exchange kinetics.
J Appl Physiol 67: 547–555, 1989.
12. Coyle EF, Sidossis LS, Horowitz JF, and Beltz JD. Cycling
efficiency is related to the percentage of type 1 muscle fibers.
Med Sci Sports Exerc 24: 782–788, 1992.
13. Crow MT and Kushmerick MJ. Chemical energetics of slowand fast-twitch muscles of the mouse. J Gen Physiol 79: 147–
166, 1982.
14. Davies CT and Barnes C. Negative (eccentric) work. II. Physiological responses to walking uphill and downhill on a motordriven treadmill. Ergonomics 15: 121–131, 1972.
15. Engelen M, Porszasz J, Riley M, Wasserman K, Maehara
K, and Barstow TJ. Effects of hypoxic hypoxia on O2 uptake
and heart rate kinetics during heavy exercise. J Appl Physiol 81:
2500–2508, 1996.
16. Enoka RM. Eccentric contractions require unique activation
strategies by the nervous system. J Appl Physiol 81: 2339–2346,
17. Gerdle B, Henrikson-Larson K, Lorentzon R, and Wretling
M-L. Dependence of the mean power frequency of the electromyogram on muscle force and fibre type. Acta Physiol Scand 142:
457–465, 1991.
18. Hagberg JM, Nagle FJ, and Carlson JL. Transient O2 uptake
response at the onset of exercise. J Appl Physiol 44: 90–92, 1978.
19. Han YS, Proctor DN, Geiger PC, and Sieck GC. Reserve
capacity for ATP consumption during isometric contraction in
human skeletal muscle fibers. J Appl Physiol 90: 657–664, 2001.
20. Henriksson J, Knuttgen HG, and Bonde-Petersen F. Perceived exertion during exercise with concentric and eccentric
muscle contraction. Ergonomics 15: 537–544, 1972.
91 • NOVEMBER 2001 •
Downloaded from by on September 29, 2016
individuals with a high percentage of fast-twitch fibers
also had a larger V̇O2 slow component (3).
Changes in MPF have been investigated as a possible indicator of the proportion of fast-twitch fibers
contributing to muscle contraction (17). Recent evidence on changes in MPF during heavy exercise is
conflicting. Although an increase in MPF of the vastus
lateralis muscle that coincided with the slow component has been observed (34), other studies found no
change (27, 35). Our results from four different muscles also suggest no change in the proportion of fasttwitch fibers contributing, but the relatively high fitness level of our subjects might have reduced the
likelihood of recruiting fast-twitch fibers (27). Furthermore, caution should always be applied to interpretation of surface EMG results.
The progressive increase in iEMG for the RF and VM
muscles paralleling the increase in V̇O2 during HC
exercise supports the hypothesis that more fibers are
being recruited. In previous investigations of EMG
during higher-intensity exercise, progressive increases
(34, 36) or no change (27, 35) in EMG was found,
coinciding with the appearance of the V̇O2 slow component. It is possible that examination of different muscles (typically only the vastus lateralis in many investigations) or selection of exercise intensity as a
percentage of the difference between VT and peak V̇O2
might be responsible for some of the between-study
discrepancies. A recent investigation (34) confirmed
greater muscle fiber recruitment from the transverse
relaxation time (T2) of magnetic resonance imaging.
The concept of recruiting additional fibers and/or increasing motor unit firing rate to maintain the work
rate is consistent with previous investigations of fatigue during sustained contractions (29). This can explain the greater metabolic cost, inasmuch as it is
necessary to maintain cellular homeostasis after an
action potential, even when there is not normal contractile activity.
Conclusion. The kinetics of V̇O2 during the onset of
high-intensity eccentric cycling exercise were described by a two-component exponential model similar
to that required to describe the responses during lowto moderate-intensity concentric cycling. Consistent
with our hypothesis, the rate of increase in V̇O2 was
faster for HE and LC exercise, where the metabolic
demand was reduced, than for the higher-intensity
(HC) concentric exercise. Previous investigations that
employed Douglas bag techniques to study the V̇O2
kinetics at the onset of eccentric exercise could not
clearly extract this information on the time course of
the response (7, 25). In our comparison of muscle activation between HE and HC exercise, it became apparent that the V̇O2 slow component was not a function of
absolute work rate, but that it increased only when the
iEMG of the VM and RF muscles increased. This latter
observation suggests that fatigue of individual motor
units forced recruitment of additional motor units to
achieve the same work rate, increasing the total V̇O2.
J Appl Physiol • VOL
31. Paterson DH and Whipp BJ. Asymmetries of oxygen uptake
transients at the on- and offset of heavy exercise in humans.
J Physiol (Lond) 443: 575–586, 1991.
32. Pimental NA, Shapiro Y, and Pandolf KB. Comparison of
uphill and downhill walking and concentric and eccentric cycling. Ergonomics 25: 373–380, 1982.
33. Plante PD and Houston ME. Effects of concentric and eccentric exercise on protein catabolism in man. Int J Sports Med 5:
174–178, 1984.
34. Saunders MJ, Evans EM, Arngrimsson SA, Allison JD,
Warren GL, and Cureton KJ. Muscle activation and the slow
component rise in oxygen uptake during cycling. Med Sci Sports
Exerc 32: 2040–2045, 2000.
35. Scheuermann BW, Hoetling BD, Noble ML, and Barstow
TJ. The slow component of O2 uptake is not accompanied by
changes in muscle EMG during repeated bouts of heavy exercise
in humans. J Physiol (Lond) 531: 245–256, 2001.
36. Shinohara M and Moritani T. Increase in neuromuscular
activity and oxygen uptake during heavy exercise. Ann Physiol
Anthrop 11: 257–262, 1992.
37. Wells R, Morrissey M, and Hughson R. Internal work and
physiological responses during concentric and eccentric cycle
ergometry. Eur J Appl Physiol 55: 295–301, 1986.
38. Whipp BJ and Ward SA. Physiological determinants of pulmonary gas exchange kinetics during exercise. Med Sci Sports
Exerc 22: 62–71, 1990.
39. Winter DA and Scott SH. Technique for interpretation of
electromyography for concentric and eccentric contractions in
gait. J Electromyogr Kinesiol 4: 263–269, 1991.
91 • NOVEMBER 2001 •
Downloaded from by on September 29, 2016
21. Hughson RL, Green HJ, and Sharratt MT. Gas exchange, blood
lactate, and plasma catecholamines during incremental exercise in
hypoxia and normoxia. J Appl Physiol 79: 1134–1141, 1995.
22. Hughson RL and Morrissey MA. Delayed kinetics of V̇O2 in the
transition from prior exercise. Evidence for O2 transport limitation
of V̇O2 kinetics: a review. Int J Sports Med 11: 94–105, 1983.
23. Jones AM and McConnell AM. Effect of exercise modality on
oxygen uptake kinetics during heavy exercise. Eur J Appl
Physiol 80: 213–219, 1999.
24. Knuttgen HG, Bonde-Petersen F, and Klausen K. Oxygen
uptake and heart rate responses to exercise performed. Med Sci
Sports Exerc 3: 1–5, 1971.
25. Knuttgen HG and Klausen K. Oxygen debt in short-term
exercise with concentric and eccentric muscle contractions.
J Appl Physiol 30: 632–635, 1971.
26. Knuttgen HG, Patton JF, and Vogel JA. An ergometer for
concentric and eccentric muscular exercise. J Appl Physiol 53:
784–788, 1982.
27. Lucia A, Hoyos J, and Chicharro JL. The slow component of
V̇O2 in professional cyclists. Br J Sports Med 34: 367–372, 2000.
28. MacDonald M, Pedersen PK, and Hughson RL. Acceleration
of V̇O2 kinetics in heavy submaximal exercise by hyperoxia and
prior high-intensity exercise. J Appl Physiol 83: 1318–1325,
29. Milner-Brown HS and Stein RB. The relation between surface electromyogram and muscular force. J Physiol (Lond) 246:
549–569, 1975.
30. Moritani T, Muramatsu S, and Muro M. Activity of motor
units during concentric and eccentric contraction. Am J Phys
Med 66: 338–350, 1988.