Vertical jump coordination: Fatigue effects

Abstract

The aim of this study was to investigate the segmental coordination of vertical jumps under fatigue of the knee extensor and flexor muscles.
Eleven healthy and active subjects performed maximal vertical jumps with and without fatigue, which was imposed by requesting the subjects to extend/flex their knees continuously in a weight machine, until they could not lift a load corresponding to approximately 50% of their body weight. Knee extensor and flexor isokinetic peak torques were also measured before and after fatigue. Video, ground reaction forces, and electromyographic data were collected simultaneously and used to provide several variables of the jumps.
Fatiguing the knee flexor muscles did not reduce the height of the jumps or induce changes in the kinematic, kinetic, and electromyographic profiles. Knee extensor fatigue caused the subjects to adjust several variables of the movement, in which the peak joint angular velocity, peak joint net moment, and power around the knee were reduced and occurred earlier in comparison with the nonfatigued jumps. The electromyographic data analyses indicated that the countermovement jumps were performed similarly, i.e., a single strategy was used, irrespective of which muscle group (extensor or flexors) or the changes imposed on the muscle force-generating characteristics (fatigue or nonfatigue). The subjects executed the movements as if they scaled a robust template motor program, which guided the movement execution in all jump conditions. It was speculated that training programs designed to improve jump height performance should avoid severe fatigue levels, which may cause the subjects to learn and adopt a nonoptimal and nonspecific coordination solution.
It was suggested that the neural input used in the fatigued condition did not constitute an optimal solution and may have played a role in decreasing maximal jump height achievement.

Full text

APPLIED SCIENCES
Biodynamics
Vertical jump coordination: fatigue effects
ANDRE
´LUIZ FELIX RODACKI, NEIL E. FOWLER, and SIMON J. BENNETT
Department of Exercise and Sport Sciences, Manchester Metropolitan University, Staffordshire, UNITED KINGDOM; and
Departamento de Educac¸a˜o Fı´sica, Universidade Federal do Parana´, Jardim Botaˆnico, Curitiba, Parana´, BRAZIL
ABSTRACT
RODACKI, A. L. F., N. E. FOWLER, and S. J. BENNETT. Vertical jump coordination: fatigue effects. Med. Sci. Sports Exerc., Vol.
34, No. 1, 2002, pp. 105–116. Purpose: The aim of this study was to investigate the segmental coordination of vertical jumps under
fatigue of the knee extensor and flexor muscles. Methods: Eleven healthy and active subjects performed maximal vertical jumps with
and without fatigue, which was imposed by requesting the subjects to extend/flex their knees continuously in a weight machine, until
they could not lift a load corresponding to ~50% of their body weight. Knee extensor and flexor isokinetic peak torques were also
measured before and after fatigue. Video, ground reaction forces, and electromyographic data were collected simultaneously and used
to provide several variables of the jumps. Results: Fatiguing the knee flexor muscles did not reduce the height of the jumps or induce
changes in the kinematic, kinetic, and electromyographic profiles. Knee extensor fatigue caused the subjects to adjust several variables
of the movement, in which the peak joint angular velocity, peak joint net moment, and power around the knee were reduced and
occurred earlier in comparison with the nonfatigued jumps. The electromyographic data analyses indicated that the countermovement
jumps were performed similarly, i.e., a single strategy was used, irrespective of which muscle group (extensor or flexors) or the changes
imposed on the muscle force-generating characteristics (fatigue or nonfatigue). The subjects executed the movements as if they scaled
a robust template motor program, which guided the movement execution in all jump conditions. It was speculated that training programs
designed to improve jump height performance should avoid severe fatigue levels, which may cause the subjects to learn and adopt a
nonoptimal and nonspecific coordination solution. Conclusion: It was suggested that the neural input used in the fatigued condition
did not constitute an optimal solution and may have played a role in decreasing maximal jump height achievement. Key Words:
VERTICAL JUMPS, MOVEMENT STRATEGY, MOTOR CONTROL, TRAINING SPECIFICITY
Vertical jumping ability is a crucial skill in the per-
formance of several sports, such as volleyball, bas-
ketball, and football. The execution of this motor
task depends on the coordination of the segmental actions of
the human body, which is determined by the interaction
between the muscle forces (ultimately modulated by im-
pulses sent by the central nervous system) and the net
moments that have to be generated around the joints to
accomplish the mechanical demands of the task. Results of
kinematic and electromyographic studies have shown that
vertical jumping is performed according to a robust stereo-
typed pattern (6). It has been shown that the timing, se-
quence, and amplitude of the muscle activation and joint
movements are quite comparable, even when the movement
is performed by different subjects (22). It has also been
shown that some movement constraints (e.g., constraining
the trunk segment (12)) barely disrupt the pattern of vari-
ables used to describe coordination (e.g., muscle activation).
Rodacki et al. (28) studied coordination of vertical jump
under fatigue and also suggested the existence of a consis-
tent pattern, irrespective of the force-generating properties
of the muscles. Although the activation amplitude of the
knee extensor and flexor muscles increased at the end of
fatiguing exercises, the pattern of the electromyographic
traces remained similar to that observed before fatigue, and
the subjects performed the movement as if they scaled a
robust muscle activation pattern, which guided the execu-
tion of the movement, without considering the best available
muscle strength.
It was suggested that during maximal vertical jumps a
common drive exists that controls the agonist-antagonist
muscle pair activity as a single functional entity (27,37). In
that study (27), it was proposed that the modulation of this
common drive resulted in similar muscle activation (EMG)
between fatigued and nonfatigued jumps, but affected the
peak angular velocity and peak power around the joints
during the propulsive phase of the movement.
In a vertical jump simulation study, Bobbert and Van
Soest (7) demonstrated that although muscle strength deter-
mines the maximal jump height achievement, actual perfor-
mance depends on the control of the muscle properties. In
that study, neither increasing the muscle strength of the knee
0195-9131/02/3401-0105/$3.00/0
MEDICINE & SCIENCE IN SPORTS & EXERCISE
®
Copyright © 2002 by the American College of Sports Medicine
Submitted for publication October 2000.
Accepted for publication March 2001.
105

extensor muscles nor raising the strength of all muscles
resulted in jump height improvement, until the muscle ac-
tivation (control) was reorganized (reoptimized).
Other studies involving cycling (18), running (19), sprint-
ing (26), lifting (31), and continuous hopping exercises (8)
provide evidence that some compensatory mechanisms are
used to counterbalance the loss of the muscle force-gener-
ating properties because of fatigue. This is in agreement
with the arguments proposed by Van Ingen Schenau et al.
(33) that changes in muscle activation timing should be
accomplished to avoid deterioration of the performance
when the properties of the musculoskeletal system are
changed. Hence, under fatigue, defined as the inability of
the neuromuscular system to sustain the required or ex-
pected power output around a joint (11), compensatory
strategies may induce a reorganization of the movement
structure and a new coordination pattern may appear.
Experimentally, it is not known whether or how the
neuromuscular system reorganizes the pattern of maximal
countermovement jumps when a particular muscle group
(e.g., knee extensor or knee flexor muscles) has its force-
generating properties changed (increased or decreased).
Changing the force-generating properties of one component
of the agonist-antagonist pair is an attractive way to test
whether a common drive exists and how it influences move-
ment organization during dynamic conditions. If a common
drive guides the movement execution of maximal counter-
movement jumps, a unique stimulation signal would be used
and muscle activation timing may be consistent, and other
variables (e.g., peak angular velocity, peak net moment, and
the peak net power around the joints) may vary when one
component of the agonist-antagonist pair is fatigued (i.e.,
the muscle force-generating properties are reduced). For
instance, left- or rightward phase shifts in the timing of other
variables of the movement (e.g., peak angular velocity time)
would be expected to occur under fatigue, whereas the
muscle activation (EMG) would remain without large
variations.
It is possible that a decline in performance after fatigue may
be the result of (a) a change in coordination (i.e., changing the
neural input), (b) a change in the functional capacity of the
muscles to produce force (i.e., without changing the neural
input), or (c) the combination of these two factors. In the first
case, changes in both muscle activation and kinematics are
likely to occur, whereas the second case may be characterized
by a stable neural input (EMG), in which a different kinematic
output may emerge. In other words, does the neuromuscular
system adopt a new coordination pattern to account for local
muscular fatigue or does it do the same thing as suffer the
consequences of reduced muscular force? The aim of the
present study was to investigate whether and to what extent the
neuromuscular system (re)organizes and accommodates the
controls used in multisegment movements when different mus-
cle groups are fatigued. It has been hypothesized that fatigue
will decrease performance and will influence the magnitude
and the time of several kinematic and kinetic variables (as
speculated above), but will have little effect on muscle activa-
tion pattern (EMG). Studies analyzing how the neuromuscular
system adjusts the movement coordination pattern used during
vertical jumps under fatigue may provide valuable information
to understand motor control of multisegment movements.
METHODS
Eleven healthy male subjects (age, 23.1 ⫾4.8 yr; height,
183.4 ⫾6.1 cm; and body mass, 84.0 ⫾13.2 kg) engaged
in various sports (six volleyball players, three rugby players,
and two multiple sports) and with previous experience in
vertical jumping were informed of the procedures involved
in this study and gave their informed consent to act as
subjects, in accordance with the ACSM policy statement
regarding the use of human subjects, informed consent, and
approval by an ethics committee.
Experimental procedures. Subjects reported to the
laboratory for two test sessions, which were separated by at
least 3 d. The knee extensor muscle group was fatigued in
the first session, whereas the flexor muscle group was fa-
tigued in the second session. Fatigue was imposed by re-
questing the subjects to extend and flex both their knees in
a knee flexor/extensor weight machine (PowerSport
®
,Mid
Glamorgan, United Kingdom) using a self-selected pace,
until they could no longer lift a load. The load used for each
subject corresponded to ~50% (knee extensors) and 40%
(knee flexors) of their body mass. The knee extensors were
fatigued in an upright sitting posture, whereas the knee
flexor muscles were fatigued in a prone posture (biceps
curl). In both exercise modes, the subjects were allowed to
stabilize themselves by holding on either side of the seat,
and verbal encouragement was given. On average, the sub-
jects were not able to continue the knee extension and
flexion movements after 27 ⫾9 and 18 ⫾8 repetitions,
respectively.
Isokinetic strength measurements. The isokinetic
peak torques of the right knee extensor (PT
EXT
) and flexor
(PT
FLEX
) muscles were measured using a Cybex
®
dyna-
mometer (Cybex International, Medway, MA) in each ses-
sion. In both sessions, before the commencement of the
isokinetic assessment, the subjects were also allowed to
perform a set of five submaximal contractions in the isoki-
netic dynamometer. After finishing the warm-up on the
isokinetic dynamometer, a 3-min rest interval was imposed
before the subjects performed a maximal test of five max-
imal voluntary repetitions to represent the nonfatigued con-
dition. The isokinetic measurement in the fatigued condition
took place at the end of each session, immediately after the
countermovement jumps assessment in the fatigued condi-
tion (see below). The interval between the end of the coun-
termovement jumps and the initiation of the isokinetic test in
the fatigued condition (60–90 s) was caused by the necessity
to walk the subjects through a distance of 15 m and to set the
isokinetic dynamometer.
During the isokinetic measurements, the subjects were
seated in an upright posture (90 degrees between the trunk
and the thigh) and secured across the pelvis and trunk by a
four-point belt. Peak torque was defined as the highest
torque achieved during five successive (without pause)
106
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

maximal concentric knee extension (from 90 degrees of
knee flexion to 180 degrees of knee extension) and flexion
(from 180 degrees of knee extension to 90 degrees of knee
flexion) movements. The hamstrings to quadriceps torque
ratio (H/Q) was defined as the quotient between PT
FLEX
and
PT
EXT
. The range of motion and the movement velocity (60
deg·s
⫺1
) of the concentric contractions were preset by the
experimenter. This movement velocity was selected to rep-
resent the muscle force-generating properties. Gravity cor-
rection was performed according to manufacturer’s guide-
lines. Verbal encouragement and visual feedback of the
torque traces were given during the testing.
Vertical jumps. After the first isokinetic testing, the
subjects were prepared for the electromyographic and kine-
matic assessments (see below), received explanation and
demonstration of the countermovement jumps, and per-
formed three warm-up trials, which were followed by a set
of three maximal countermovement jumps. During the
countermovement jumps, subjects bent their knees to a
freely chosen angle, which was followed by a maximal
vertical thrust. The effect of the arms was minimized by
requesting the subjects to keep their hands and arms crossed
against the chest. Kinematic, force, and electromyographic
data were recorded simultaneously for three maximal jumps,
but only the jump that resulted in the greatest flight time was
further analyzed to represent the subject’s best performance
in the nonfatigued condition (CMJ
1
). Force data were used
to determine the flight time of each jump in all conditions.
The second set of three maximal countermovement jumps
took place immediately after the end of the fatiguing exer-
cises and dismounting the knee extensor/flexor weight ma-
chine, which was positioned close to the force platform—
approximately 3 m. The interval between the end of the
fatiguing exercises and the countermovement jumps in the
fatigued condition was kept as short as possible (⬍10 s).
The subjects were instructed to follow the same procedures
used in CMJ
1
. Again, three maximal countermovement
jumps with simultaneous kinematic, force, and EMG data
were recorded and the best performance jump selected to
represent fatigued conditions (CMJ
2 EXT
and CMJ
2 FLEX
,
respectively).
Kinematic assessment. Reflective marks were
placed on the right side of the subject’s body to match with
the following sites: 1) fifth metatarsal joint, 2) lateral mal-
leolus, 3) lateral femoral epicondyle of the knee, 4) the most
prominent protuberance of the greater trochanter, and 5)
neck at the level of the fifth cervical vertebrae. The subjects
were filmed (100 Hz) using a two-dimensional kinematic
optoelectric system (ELITE, BTS, Milan, Italy) and the
coordinates of the marker points were filtered using spline
functions. These filtered body markers defined the position
of the foot (FOT), shank (SHA), thigh (THI), and upper
body (TRU), and were used to calculate joint angular dis-
placement, velocity, and acceleration of the ankle (ANK),
knee (KNE), and hip (HIP). Figure 1 provides visual infor-
mation of the four-segment model and also shows the joint
angle conventions.
Electromyographic assessment. Surface electro-
myographic signals were recorded from gastrocnemius late-
ralis (GAS) (over the area of the greatest muscle bulk on the
lateral calf), soleus (SOL) (over the lateral edge, where the
muscle protrudes below the GAS), vastus lateralis (VL)
(over the area of the greatest muscle bulk just lateral to the
rectus femoris on the distal half of the thigh), rectus femoris
(RF) (over the midpoint between the anterior superior iliac
spine and the patella superior border), semitendinosus (ST)
(midway on a line between the ischial tuberosity and the
medial epicondyle of the tibia), and gluteus maximus (GLU)
FIGURE 1—The four-body segment model and the joint angle con-
vention. The muscles soleus (SOL), gastrocnemius (GAS), semitendi-
nosus (ST), vastus lateralis (VL), rectus femoris (RF), and gluteus
maximum (GM) are indicated.
VERTICAL JUMP COORDINATION: FATIGUE EFFECTS Medicine & Science in Sports & Exercise姞
107

(over the bulkiest part of the middle of the muscle belly).
The electrode placement sites followed the recommendation
of Acierno et al. (1).
Electromyographic signals were obtained using dispos-
able bipolar surface electrodes (Bio-tabs MSB
®
(MIE Med-
ical Research Ltd., Leeds, UK) Ag/AgCl, with leadoff area
2.75 cm
2
), placed with center-to-center distance of 1.5 cm
and border-to-border distance of 1.0 cm. Reference elec-
trodes (3M Red Dot
®
, type 2237 (3M Company, St. Paul,
MN), Ag/AgCl with foam tape and solid gel) were used at
the most distant point possible away from the electrode sites
(approximately 10–15 cm), toward the most distal point of
the segments. Because of the fast and explosive character-
istic of the movement, all sites were covered with straps of
adhesive tape to prevent disconnection and reduce move-
ment artifacts. All test sites were identified and prepared by
the same experimenter.
The electromyographic signals were preamplified in sub-
miniature amplifiers before transmission via FM radio te-
lemetry (459 MHz, with channel bandwidth of 1000 Hz) to
a recording device no farther than 3 m away. The miniature
preamplifiers (33 ⫻21 ⫻9 mm) provided a gain of 1000,
bandwidth of 15 kHz, noise of less than ⫺52 dB, common
mode rejection ratio of ⫺102 dB, and input impedance
greater than 10
8
⍀. The raw electromyograms were pro-
cessed into a linear envelope (EEMG). The EEMGs were
calculated using the MYO-DAT
®
5.0 EMG analysis pack-
age software (MIE Medical Research Ltd., Leeds, UK) by
applying a second-order low-pass filter set at 6 Hz fre-
quency. Electromyographic data were sampled at 200 Hz.
Because of technical limitations, it was not possible to
record the electromyographic signals using a sampling fre-
quency higher than 200 Hz. Therefore, the relatively low
sampling frequency used in this study (200 Hz) may not
have allowed the high-frequency components to be recorded
adequately, which may constitute a violation of the sam-
pling theorem. In order to assess the impact of using such
resolution, we performed an assessment in which the elec-
tromyographic sampling frequency was set at 800 Hz. These
data were used to generate a second data set, in which the
sampling frequency was reduced to 200 Hz. Then, both data
sets (800 and 200 Hz) were processed (rectified, filtered,
and normalized with respect to magnitude). The average
root mean square difference was 6.0 ⫾0.5% (GAS, 5.9;
VM, 6.6; VL, 5.5). The differences in the detection of
initiation (ON) and peak (PK) instants between the two data
series were small (ON, 4 to 7 ms; PK, 5 to 8 ms, respec-
tively). Therefore, and despite constituting a certain limita-
tion, the EMG data collected at 200 Hz can be used to
represent the muscle activation pattern without obscuring
relevant aspects of the biological significance of the move-
ment coordination analyzed in this study.
Kinetic assessment. A force platform (Kistler
®
,
model 9281B, Kistler Instruments, Winterthur, Switzerland)
synchronized with the kinematic and electromyographic
measurements and sampling at 1000 Hz provided force-time
traces. The kinematic analysis was combined with the
ground reaction forces to calculate net moments at the ankle,
knee, and hip joints. The moment of inertia of each segment
was estimated by using the Drillis and Contini (10) equa-
tions. Net powers around the joints were also calculated by
multiplication of the net moments and joint angular veloc-
ities. The net impulse was calculated by integrating the
force-time curves of the vertical component of the ground
reaction forces during the positive phase of the movement
(see below). Extension moments were considered positive at
all joints. All kinetic data were normalized with respect to
body weight (BW).
Definition of variables and data analysis. The
times at which initiation of extension (IEX) and peak an-
gular velocity (PAV) of the ankle (ANK), knee (KNE), and
hip (HIP) joints were determined. IEX was defined as the
first instant (frame) after a joint reaches its deepest flexion
angle, whereas PAV was defined as the instant at which the
greatest joint angular velocity is achieved during the pro-
pulsive phase of the movement. The difference in time at
which IEX of each joint occurred was used to determine the
relative timing and the sequential relationship between ad-
jacent segments (28). In other words, if the movement of a
proximal segment precedes the movement of its distal coun-
terpart (i.e., a proximal-to-distal order), the difference be-
tween the IEX of these joints would be negative. Because
the IEX was determined kinematically (from the joint angle
position data), it is estimated that a mean error of 5 ms
(ranging from 0 to 9 ms) may have occurred.
In order to analyze the movement sequence and temporal
organization, the first and the last data points corresponding,
respectively, to the beginning and take-off instants were
used to define the movement duration, i.e., the contact time
phase (CT) duration. The CT was fractioned in three phases:
the negative phase (NEG), the transient phase (TR), and the
positive phase (POS). The initiation of the NEG phase was
calculated using the ground reaction force traces and was
determined as the instant before the vertical force decreases
continuously for a period longer than 0.02 s (two frames).
The end of the NEG and the beginning of the POS phases
was defined as the first instant in which the vertical velocity
of the body mass center is positive. The vertical velocity
body mass center was calculated by integrating the net
vertical ground force data (2). The end of the POS phase was
determined with the help of the force data and was defined
as the instant in which the subjects lost contact with the
force platform, i.e., the instant in which the vertical com-
ponent of the ground reaction forces was zero—the take-off
instant (TO). At the end of the NEG phase and at the
beginning of the POS phase, there is a phase of transition
(TR) in which there are no large changes in the knee joint
angular velocity. For analysis purposes, the TR phase was
determined as the period in which the knee angular velocity
ranged between ⫹30 deg·s
⫺1
and ⫺30 deg·s
⫺1
in relation to
the deepest knee flexion angle (IEX) (28).
To better compare the characteristics of the electromyo-
graphic signals, the EEMG traces were normalized with
respect to the signal magnitude. For each muscle, the highest
electromyographic signal value obtained during the perfor-
mance of each trial was used as reference and set at 100%.
108
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

In the next sections, the processed electromyographic sig-
nals (rectified, filtered, and normalized with respect to mag-
nitude) are referred to as EMG. The ON and PK activation
of each muscle were also examined. ON was arbitrarily
considered as the first instant in which the EMG traces were
equal to or greater than 20% of PK, which was defined as
the highest muscle activation obtained during the POS
phase. The use of different criteria to determine the instant
in which muscle activation was initiated (ON) may produce
different results. Cross-correlation analysis (bidirectional
phase shift of 15 lags; 75 ms) were applied on the EMG
traces between CMJ
1
and the fatigued conditions (CMJ
2 EXT
and CMJ
2 FLEX
) to account for phase shifts between these
signals.
Muscle activation was quantified during the POS phase of
the movement by dividing the integrated EMG signal (the area
under the muscle activation curve) by the duration of the
movement phase. Changes in muscle activation were ex-
pressed as a mean percentage difference of the initial values
(CMJ
1
). Therefore, a positive value indicates greater muscle
activity in the fatigued condition than in the nonfatigued con-
dition, and a negative value indicates the opposite.
Knee joint stiffness was determined by calculating the
coefficient of linear regression of the moment-angle rela-
tionship from the last 15 degrees of the NEG phase to the
deepest excursion of the knee joint. This analysis only
considered the final 15 degrees of the NEG phase because
no differences were found before this point.
At-test was performed to compare several variables used
to describe coordination in the nonfatigued condition
(CMJ
1
) between the jumps performed in the first and second
sessions. Because no significant differences were found (P
⬍0.05), all variables representing this condition were col-
lapsed, and an ensemble average was used to represent the
nonfatigued condition (CMJ
1
). To compare the kinematic,
kinetic, electromyographic, and joint stiffness data between
conditions, a number of one-way ANOVAs with repeated
measures were applied. When significant differences were
detected, a Newman-Keuls post hoc test was applied. Effect
size (ES) was also calculated. A Kolgomorov-Smirnov test
was applied and confirmed data normality. All statistical
analyses were performed in the Statistica
®
package soft-
ware, version 5.5 (StatSoft, Inc., Tulsa, OK) and the signif-
icance level was set at P⬍0.05.
RESULTS
Isokinetic peak torques. No significant differences in
peak torque in the nonfatigued condition (PT
EXT
and PT-
FLEX
) were found between the first and second two sessions
(0.29 and 0.18 Nm·BW
⫺1
, respectively; P⬎0.05). The
fatigue protocol used in this study was proven to success-
fully reduce the peak torque of the knee extensor (PT
EXT
⫽
0.29 ⫾0.03 Nm·BW
⫺1
) and flexor (PT
FLEX
⫽0.17 ⫾0.02
Nm·BW
⫺1
) muscles by 14.2% and 12.6%, respectively (P
⬍0.05; ES ⫽1.43 and 1.2, respectively). When these
muscle groups acted as antagonists, the extensor and flexor
peak torques remained unchanged (1.4% and 1.1%, respec-
tively; P⬎0.05) in comparison to the initial condition
(ISO
1
). The H/Q found in ISO
1
(0.61 ⫾0.05) increased by
15.2 ⫾6.1% (P⬍0.05; ES ⫽1.51) after fatiguing the knee
extensor muscles, and decreased by 11.4 ⫾7.6% (P⬍0.05;
ES ⫽1.41) after fatiguing the knee flexor muscles.
Vertical jump performance. In the nonfatigued con-
dition there was no significant difference (P⬎0.05) be-
tween the jumps performed in the first and second sessions
(0.33 ⫾0.06 m and 0.33 ⫾0.07 m, respectively). Fatigue
reduced the ability of the subjects to jump as high as in the
nonfatigued condition only in CMJ
2 EXT
. On average, the
subjects were able to jump 86.1 ⫾7.0% (P⬍0.05; ES ⫽
0.74) and 93.8 ⫾4.0% (P⬎0.05; ES ⫽0.16) of the
maximal height jumped in CMJ
1
(0.33 ⫾0.06 m) during
CMJ
2 EXT
and CMJ
2 FLEX
, respectively.
Kinematics and kinetics. Fatigue did not change the
total duration of the countermovement jumps between CMJ
1
and the fatigued conditions (CMJ
2 EXT
and CMJ
2 FLEX
)(P⬎
0.05). In addition, the duration of the NEG, TR, and POS
phases of the movement remained unaltered (P⬎0.05; ES
⬍0.22) in comparison with CMJ
1
, irrespective of which
muscle group was fatigued (CMJ
2 EXT
and CMJ
2 FLEX
).
The magnitude and the time of the kinematic and kinetic
variables analyzed during the NEG phase of the movement
(Figs. 2 and 3) were similar in both fatigue conditions
(CMJ
2 EXT
and CMJ
2 FLEX
) and did not differ significantly
(P⬎0.05). In CMJ
2 EXT
, the knee joint was less flexed
FIGURE 2—Angle-angle representation of the ankle-knee (top) and
knee-hip (bottom) joint angular displacement before (CMJ
1
) and after
fatiguing the knee extensor (CMJ
2 EXT
) and flexor (CMJ
2 FLEX
)
muscles.
VERTICAL JUMP COORDINATION: FATIGUE EFFECTS Medicine & Science in Sports & Exercise姞
109

during its deepest position (P⬍0.05; ES ⫽0.78) and the
excursion of the body mass center was shallower (P⬍0.05;
ES ⫽0.74) than in CMJ
1
. The angular displacement of the
ankle, knee, and hip joints across the experimental condi-
tions are shown in Figure 2.
The delays between IEX
HIP
–IEX
KNE
and IEX
KNE
–IEX-
ANK
indicated the existence of a proximal-to-distal order, in
which the hip was consistently the first joint to extend and
was followed by knee and ankle joint extensions (Table 1),
irrespective of the fatigue condition. The differences be-
tween IEX
KNE
and IEX
ANK
revealed a variable pattern
where, in some cases, ankle extension preceded knee exten-
sion. The statistical analyses of these delays suggest that the
proximal-to-distal order was not influenced by muscle
group fatigue (P⬎0.05; ES ⫽0.23).
Figure 3 shows the joint angular velocity of the ankle,
knee, and hip. Peak positive joint angular velocity was
reduced at the knee and hip joints (P⬍0.05; ES ⫽0.50),
when the knee extensor muscles were exercised (CMJ
2 EXT
).
The peak knee positive angular velocity occurred earlier
(P⬍0.05; ES ⫽1.31) in CMJ
2 EXT
than in CMJ
1
, and no
significant differences (P⬎0.05) were found in other joints.
Fatiguing the knee flexor muscles (CMJ
2 FLEX
) did not
change the ankle, knee, or hip peak joint positive angular
velocity magnitude or time (P⬎0.05).
Fatiguing neither the knee extensor (CMJ
2 EXT
) nor the
flexor muscles (CMJ
2 FLEX
) decreased the net peak power,
in comparison with the nonfatigued condition (CMJ
1
)(P⬎
0.05; ES ⫽0.34 and 0.36, respectively). The net peak power
of the ankle and knee joints occurred 25.7% (P⬍0.05; ES
⫽0.90) and 18.9% (P⬍0.05; ES ⫽1.04) earlier in CMJ
2
EXT
than in CMJ
1
. No significant changes in the temporal
characteristics of the net peak power around the hip joint
were detected (P⬎0.05; ES ⫽0.35) between CMJ
1
and
CMJ
2 FLEX
. The net impulse determined during the positive
phase in CMJ
1
was reduced by 8.5% in CMJ
2 EXT
(P⬍0.05;
ES ⫽0.38), whereas no significant changes were found in
CMJ
2 FLEX
(P⬎0.05; ES ⫽0.15).
Figure 4 shows the average stiffness of the knee joint.
Knee joint stiffness was calculated during the final 15 de-
grees of knee flexion, at the end of the NEG phase. Fatigu-
ing the knee flexor muscles (CMJ
2 FLEX
) did not change the
stiffness of the knee, which was similar to that observed in
CMJ
1
(P⬎0.05; ES ⫽0.46). On the other hand, knee joint
stiffness increased (P⬍0.05; ES ⫽1.2) during the final
part of the NEG phase of the movement in CMJ
2 EXT
.
Electromyographic analyses. Muscle activation of
the knee extensor muscles (VL and RF) increased by 39.0%
(P⬍0.05; ES ⫽1.4) in CMJ
2 EXT
and by 18.8% (P⬍0.05;
ES ⫽1.2) in CMJ
2 FLEX
in relation to CMJ
1
during the POS
phase of the movement. Muscle activation of the biarticular
flexor muscles of the knee (ST) increased by 29.6% (P⬍
0.05; ES ⫽1.0) in CMJ
2 EXT
and increased by 25.4% (P⬍
0.05; ES ⫽1.2) in CMJ
2 FLEX
in comparison with the
control condition. The activation of the SOL, GAS, and GM
muscles in CMJ
2 EXT
and CMJ
2 FLEX
remained without large
variations and did not differ significantly (P⬎0.05; ES ⫽
0.3) in relation to CMJ
1
. ON and PK did not differ signif-
icantly (P⬎0.05; ES ⫽0.2) in comparison with CMJ
1
,in
all muscles and fatigue conditions (CMJ
2 EXT
and CMJ
2
FLEX
) (Figs. 5 and 6). Cross-correlation analysis (Table 2)
showed that the highest correlation occurred at approxi-
mately zero phase lag, indicating that no significant phase
shifts occurred between the experimental (CMJ
2 EXT
and
CMJ
2 FLEX
) and the control (CMJ
1
) conditions.
DISCUSSION
Isokinetic peak torques and H/Q ratio. The equiv-
alent isokinetic peak torque of the knee extensor and flexor
muscle groups between the first and second sessions in the
nonfatigued condition suggested a good reproducibility of
the measurements when the performers were tested on dif-
ferent days. The H/Q ratio of approximately 0.6 found in the
nonfatigued condition is in agreement with other studies
using low-speed testing (13). The protocol used to fatigue
the knee extensor/flexor muscles was proven as a successful
way to manipulate (increase or decrease) the H/Q ratio.
FIGURE 3—Ensemble average of the ankle (top), knee (middle), and
hip (bottom) angular velocity before (CMJ
1
) and after fatiguing the
knee extensor (CMJ
2 EXT
) and flexor (CMJ
2 FLEX
) muscles. The stan-
dard deviation (ⴞ1 SD) of CMJ
1
is represented. The take-off instant
is indicated by the dotted vertical line, and the arrows highlight the peak
angular velocity time.
110
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Although the decrease in the peak torque of the knee
extensor and flexor muscles (14.2 and 12.6%, respectively)
may not be considered as “large,”these significant declines
were interpreted as satisfactory, since changes imposed after
strengthening training program (viewed as changes in the
opposite direction, i.e., gains in peak torque) rarely surpass
these values. It can be speculated that the changes in H/Q at
the time of performing maximal countermovement jumps in
the fatigued condition were even greater than that revealed
by the isokinetic peak torque testing because of the rela-
tively large interval (60 to 90 s) allowed between the coun-
termovement jumps and the isokinetic strength assessments.
The discussion and the results of this study are derived from
the fatigue level achieved in this experiment, and the reader
should bear in mind that different fatigue protocols and
other fatigue levels may invoke different responses.
FIGURE 4—The knee joint stiffness in
nonfatigued (CMJ
1
) and fatigued (CMJ
2
EXT
and CMJ
2 FLEX
) conditions. The neg-
ative (NEG), transient (TR), and positive
(POS) phases are indicated.
TABLE 1. Kinematic and kinetic characteristics of countermovement jumps before (CMJ
1
) and after fatiguing the knee extensor (CMJ
2
EXT) and flexor (CMJ
2
FLEX) muscles.
a
CMJ
1
CMJ
2
EXT
Variation
(%) CMJ
2
FLEX
Variation
(%)
NEG duration (ms) 615 ⫾82 605 ⫾132 –1.6 623 ⫾123 ⫹1.3
TR duration (ms) 47 ⫾943⫾6 –8.5 48 ⫾8⫹2.1
POS duration (ms) 319 ⫾47 325 ⫾50 ⫹1.8 328 ⫾76 ⫹2.8
Displacement, BMC (% standing still) 68.8 ⫾10.9 60.7 ⫾11.6
b
–11.7 67.3 ⫾132 –2.1
Impulse (N䡠s
–1
䡠BW
–1
) 300.7 ⫾69 279.0 ⫾72.4
b
–7.2 289.7 ⫾67.5 –3.6
Knee stiffness (Nm䡠BW
–1
䡠deg
–1
䡠10
–3
) 14.4 ⫾5.8 21.4 ⫾13.1
b
⫹48.6 17.1 ⫾7.6 ⫹18.7
Ankle
IEX joint angle (deg) 90.6 ⫾5.3 91.3 ⫾6.6 ⫹0.7 89.9 ⫾4.7 –0.7
IEX time (ms) 268 ⫾42 284 ⫾70 ⫹5.9 281 ⫾66 ⫹4.8
PAV, POS (deg䡠s
–1
) 594 ⫾141 587.3 ⫾120 –1.1 584.2 ⫾125 –1.7
PAV time, POS (ms) 67 ⫾20 90 ⫾36
b
⫹34.3 73 ⫾31 ⫹12.0
Peak power, POS (W䡠BW
–1
) 2.01 ⫾0.58 1.89 ⫾0.46 –5.9 1.95 ⫾0.38 –2.9
Time to peak power (ms) 70 ⫾20 88 ⫾28
b
⫹25.7 77 ⫾28 ⫹10.0
Knee
IEX joint angle (deg) 89.5 ⫾12.4 91.5 ⫾10.4
b
⫹2.2 85.2 ⫾10.3 –4.8
IEX time (ms) 313 ⫾49 321 ⫾55 ⫹2.5 324 ⫾83 ⫹3.5
PAV, POS (deg䡠s
–1
) 700.6 ⫾119 640.3 ⫾106
b
⫹8.6 683.3 ⫾121 –2.5
PAV time, POS (ms) 66 ⫾19 91 ⫾34
b
–37.8 77 ⫾31 ⫹16.6
Peak power, POS (W䡠BW
–1
) 1.65 ⫾0.59 1.52 ⫾0.64 –7.8 1.61 ⫾0.68 –2.4
Time to peak power (ms) 116 ⫾21 138 ⫾35
b
⫹18.9 123 ⫾24 ⫹6.0
Hip
IEX joint angle (deg) 65.4 ⫾12.7 69.9 ⫾15.7 ⫹5.5 70.0 ⫾13.8 ⫹7.0
IEX time (ms) 387 ⫾35 392 ⫾46
b
⫹1.3 397 ⫾59 ⫹2.5
PAV, POS (deg䡠s
–1
) 472.7 ⫾61 437.1 ⫾67
b
–7.5 446.3 ⫾63 –5.5
PAV time, POS (ms) 74 ⫾21 69 ⫾41 –6.7 84 ⫾30 ⫹13.5
Peak power, POS (W䡠BW
–1
) 1.42 ⫾0.44 1.27 ⫾0.52 –10.5 1.26 ⫾0.42 –11.2
Time to peak power (ms) 204 ⫾61 202 ⫾71 –1.0 209 ⫾66 ⫹2.4
Delay between IEX
HIP
–IEX
KNE
(ms) 74 ⫾13 71 ⫾11 –4.0 73 ⫾14 –1.3
Delay between IEX
KNE
–IEX
ANK
(ms) 45 ⫾40 37 ⫾67 –17.7 43 ⫾68 –4.4
CT, contact time; NEG, negative phase; POS, positive phase; BMC, body mass center; IEX joint angle, initiation of the joint extension; IEX time, time to initiation of the joint extension;
PAV, peak angular velocity; PAV time, time to peak angular velocity.
a
The percentage of variation is expressed in relation to the changes in the fatigued condition (CMJ
2
) in relation to the initial condition (CMJ
1
). The time of the peak expressed
in relation to take-off instant, which was set as zero.
b
Significant differences between experimental conditions (P⬍0.05).
VERTICAL JUMP COORDINATION: FATIGUE EFFECTS Medicine & Science in Sports & Exercise姞
111

Vertical jump performance. Vertical jump height
achievement in the CMJ
1
(0.33 m) was comparable to other
studies, in which subjects jumped using an equivalent tech-
nique (e.g., Hortoba´gyi et al. (20)). The pronounced de-
crease in the maximal countermovement jump performance
that occurred by fatiguing the knee extensor muscles sug-
gests that this group possesses greater importance than its
antagonist counterpart. This is not surprising if one takes
into account that the positive work done by the knee exten-
sors (49% of the net amount of work done) is much greater
than that reported around the ankle (23%) and hip joints
(28%) (21). Additionally, some studies (23,27) have esti-
mated that the knee flexor muscle group exerts a moment of
force that accounts for 11 to 17% of the resultant joint
moment, which is substantially less than that generated by
the knee extensors (the main force generators).
Kinematics and kinetics. Fatiguing the knee flexor
muscles did not change significantly the kinematic or ki-
netic variables at any joint level, during either the negative
or positive phases of the movement. Despite most flexor
muscles of the knee being biarticular (hamstrings), and
assuming that fatigue affected the ability of these muscles to
generate torque at both joints they span, the initiation of the
trunk segment extension remained relatively unaltered. In
part, this emphasizes the arguments that the gluteus maxi-
mus, which was not fatigued in this study, is the strongest
hip joint extensor and performs most work necessary to
extend the hip joint, while the hamstrings “make negligible
contributions to the joint angular accelerations”(25).
Fatiguing the knee flexor muscles did not produce sig-
nificant changes at the knee joint level in any variables
assessed in this study (see Figs. 2 and 3). It is apparent that
the level of fatigue imposed on the knee flexor muscles (ⱖ
12.6% decrease in PT
FLEX
) was not adequate to compel the
neuromuscular system to reorganize the segmental motion.
It is possible that after “fatiguing,”these biarticular muscles
were still able to exert a large enough flexor moment about
the knee to help reduce the extensor moment at this joint to
FIGURE 5—The electromyographic sig-
nals from soleus (SOL), gastrocnemius
(GAS), and vastus lateralis (VL). The
take-off instant is indicated. The initiation
of the push-off phase of the nonfatigued
condition (CMJ
1
) is represented. The mus-
cle activation was normalized with respect
to the maximal activation during the
movement in each jump condition. The
standard deviation of the countermove-
ment jumps performed in the CMJ
1
is
represented. CMJ
2 EXT
and CMJ
2 FLEX
are the knee extensor and flexor muscles
in the fatigued condition, respectively.
112
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

delay the initiation of joint extension. Therefore, it can be
speculated that the hamstrings muscles were still able to
play a role (fine-tune regulation) at the knee joint level
under the fatigue level used in this study.
Fatiguing the knee extensor muscles caused the subjects
to adjust several kinematic and kinetic variables of the
movement, which included a reduced joint angular displace-
ment of the knee joint, a decreased knee and hip peak joint
angular velocity, and an increased knee joint stiffness. In
addition, peak knee joint angular velocity and peak knee
power during the positive phase of the movement occurred
earlier when the knee extensor muscles were fatigued than
in the control condition.
Increased joint stiffness in stretch-shortening cycle has
been considered as an efficient way to potentiate the positive
phase of countermovement jumps (15). In this study, the
increased stiffness detected at the end of the negative phase
may have contributed to the smaller knee joint angular
displacement and to keep a short coupling time between the
eccentric and the concentric phases of the movement (9).
The potentiation that occurs after fatigue, i.e., increased
reuse of the stored elastic energy and reflex sensitivity that
accompany increased joint stiffness (15), may explain the
similar duration of the movement phases between experi-
mental and control countermovement jumps (36). This may
also account for the similar temporal structure of movement,
which remained relatively unaltered after fatiguing the knee
extensor muscles.
Rather than influencing the early stages of the movement
(NEG phase), where most variables did not change with
fatigue (see Figs. 2 and 3), the decline in the ability of the
extensor muscles to generate force was more reflected dur-
ing the final part of the movement, as the take-off ap-
proached. This is in agreement with previous studies
(24,32), in which concentric contractions have been re-
ported as more affected by the loss of the force-generating
properties of the contractile components than eccentric con-
tractions. Since each step in the chain of the events for
muscle contraction could be a site for fatigue, it is difficult
to determine the exact mechanisms that account for the
FIGURE 6—The electromyographic sig-
nals from rectus femoris (RF), semitendi-
nosus (ST), and gluteus maximum (GM).
The take-off instant is indicated. The ini-
tiation of the push-off phase of the nonfa-
tigued condition (CMJ
1
) is represented.
The muscle activation was normalized
with respect to the maximal activation
during the movement in each jump condi-
tion. The standard deviation of the coun-
termovement jumps performed in the
CMJ
1
is represented. CMJ
2 EXT
and CMJ
2
FLEX
are the knee extensor and flexor
muscles in the fatigued condition,
respectively.
VERTICAL JUMP COORDINATION: FATIGUE EFFECTS Medicine & Science in Sports & Exercise姞
113

reduced ability of the muscles to generate force (e.g., central
and/or peripheral fatigue; see Fitts (14)). Hence, rather than
debating the possible sources and causes of fatigue, the
present study focuses on the consequences of fatigue on
movement coordination.
The present study revealed an early occurrence (leftward
shift) in the time of the peak angular velocity around the
knee and ankle joints after fatiguing the knee extensor
muscles in comparison with the control countermovement
jumps. Similar shifts were also observed in the net peak
power around the knee joint. These findings will be dis-
cussed with the help of the electromyographic data.
Electromyography. The EMG traces found in the con-
trol countermovement jumps are qualitatively similar to
those reported in the literature (e.g., Bobbert and Van Ingen
Schenau (6) and Viitasalo et al. (35)). The increased acti-
vation that occurred immediately after the respective series
of fatiguing exercises (see Table 2) is in agreement with
other studies (26,35) and has been attributed to several
factors such as the recruitment of new motor unit pools (38)
and firing rate synchronization (4). These mechanisms have
been generally interpreted as an attempt by the neuromus-
cular system to compensate for the failure to produce the
same force output.
It is difficult to determine whether these adjustments in
muscle activity constitute a physiological and/or cognitive
strategy of the neuromuscular system to accommodate
changes in muscle force-generating properties. However,
the adjustments in muscle activation that are generally ob-
served after a period of training (17) suggest that changes in
muscle activation magnitude during maximal explosive
movements are likely to be a physiological response of the
neuromuscular system rather than a voluntary or cognitive
adjustment. The results presented by Psek and Cafarelli
(27), in which a common drive controls the activation of an
agonist-antagonist pair as a functional entity, is also further
evidence that muscle activation magnitude regulation of
movements involving production of high force level is me-
diated by a spinal cord mechanism. A comprehensive dis-
cussion of this issue is beyond the scope of this study.
Hamstring coactivation is an important factor in main-
taining knee joint stability (3) and has been shown to in-
crease in high-velocity contractions as a protective mecha-
nism (30). In countermovement jumps, where the knee peak
positive joint angular velocity is high (700 to 1000 deg·s
⫺1
(6,35)), strong antagonistic activity is expected during the
final part of the movement. If the subjects do not decelerate
the knee joint before full extension (via hamstrings cocon-
traction), the considerable amount of rotational energy
achieved at the final part of the push-off phase would expose
the soft tissues to damaging hyperextension. Despite playing
a role stabilizing the knee joint, the increased hamstring
coactivation detracts from the resultant (extensor) moment
and may explain the early peak angular velocity and peak
power that occurred after fatiguing the knee extensor
muscles.
It is not known if the greater hamstrings activation that
occurred when the knee flexor group was fatigued is an
attempt of the neuromuscular system to sustain the rota-
tional energy of the trunk segment as high as possible.
Increasing the rotational energy around the hip joint would
increase the amount of energy that could be transferred to
the knee joint via rectus femoris (6). This would also explain
the high activation of the rectus femoris and the small
(nonsignificant) decrease in hip joint angular velocity dur-
ing the push-off phase of the movement when the knee
extensor muscles were fatigued.
Setting aside the differences in the magnitude of the
electromyographic signals, which were interpreted in the
present study as a physiological response, the traces re-
corded after fatigue were similar to those recorded during
the control condition. The electromyographic traces in the
fatigued condition were an amplified scale of the traces
observed without fatigue, where the temporal characteristics
remained relatively unchanged, irrespective of which mus-
cle group was fatigued (see Figs. 5 and 6). This is in
agreement with the speed-sensitive hypothesis (16), which
suggests that faster limb movements are produced by in-
creasing the magnitude of the neural pulse, whereas the
duration of the pulse remains relatively constant. The sim-
ilarities between the electromyographic traces and the ab-
sence of a significant phase shift between jump conditions
indicates the existence of a stereotypical neural input, irre-
spective of which muscle had its force-generating property
changed. These findings support the hypothesis that a stable
coordination pattern exists. Therefore, changes in the kine-
matics and performance of the movement are likely to be the
result of changes in the muscle force-generating properties
TABLE 2. Average (⫾SD) changes in muscle activation amplitude, phase shift, and the cross-correlation values of six muscles assessed before (CMJ
1
) and after fatiguing the
knee extensor (CMJ
2
EXT) and flexor (CMJ
2
FLEX) muscles.
a
Muscle
CMJ
2
EXT CMJ
2
FLEX
Muscle
Activation
(% CMJ
1
)
Phase
Shift (ms)
Cross-
Correlation
Muscle
Activation
(% CMJ
1
)
Phase
Shift (ms)
Cross-
Correlation
SOL 4.1 ⫾3.3 5 ⫾1.9 0.94 ⫾0.03 3.5 ⫾3.1 2 ⫾0.7 0.97 ⫾0.02
GAS 6.1 ⫾4.0 12 ⫾1.6 0.95 ⫾0.04 5.2 ⫾4.7 6 ⫾1.7 0.96 ⫾0.04
VL 39.2 ⫾9.5
b
13 ⫾1.7 0.96 ⫾0.02 18.5 ⫾5.0
b
14 ⫾1.6 0.98 ⫾0.03
RF 39.0 ⫾6.6
b
15 ⫾2.9 0.93 ⫾0.04 19.3 ⫾6.3
b
14 ⫾2.0 0.93 ⫾0.05
ST 29.5 ⫾5.1
b
10 ⫾1.8 0.93 ⫾0.05 25.1 ⫾7.0
b
9⫾1.9 0.94 ⫾0.03
GM 5.2 ⫾2.9 6 ⫾1.0 0.97 ⫾0.07 6.2 ⫾4.3 8 ⫾1.3 0.98 ⫾0.02
a
The variation in muscle activation is expressed in percentage and is relative to the nonfatigued condition (% CMJ
1
). Positive cross-correlation indicates a leftward phase shift, whereas
a negative value indicates a rightward phase shift.
b
Significant differences (P⬍0.05).
114
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

(e.g., contractile component failure) rather than changes in
the temporal characteristics of the neural input.
The findings of the present study are consistent with
simulation (7) and experimental (5) studies, where the ex-
ecution of explosive movements was proven to rely on
preprogrammed muscle stimulation patterns, which cannot
be promptly changed, irrespective of the muscle properties.
Bobbert and Van Soest (7) showed that muscle strength
gains were not reflected in jump height achievement until
the control of the muscle properties were reoptimized. These
findings suggest the existence of a link between the capa-
bility of the muscles to generate force and the set of com-
mands (neural input) used by the neuromuscular system to
perform the movement and achieve maximal performances.
In this study, fatiguing a muscle set (knee extensors/flexors)
did not cause the neuromuscular system to reorganize the
temporal characteristics of the controls (neural input). This
is in agreement with the arguments of Van Zandwijk et al.
(34) that maximal movements require a unique set of control
signals yielding maximal performances.
It has been suggested that control reorganization occurs only
after a period of practice, where the subjects are allowed to
repeatedly solve the task requirements and learn how to control
their changed muscle properties to improve their jump height
performance. This may not be the case of maximal vertical
jumps, which are not always practiced continuously under
fatigue conditions (e.g., during a volleyball game). In other
activities, in which the movement is frequently practiced under
fatigue circumstances, as a natural consequence of the long
duration and the nature of the activity, reorganization is likely
to occur to prevent potential disruption of the movement pat-
tern and performance (33). Indeed, several studies have re-
ported muscle stimulation and/or segmental movement pattern
changes under fatigue (cycling (18), running (19), sprinting
(26), and lifting (31)). Note that most of these movements are
repeatedly performed under fatigue conditions (e.g., running
and sprinting) during training and/or competition situations.
When trained under these circumstances, the subjects may be
familiar with the coordinative exigencies required to accom-
modate the force strength decline and may perform the move-
ment “optimally,”i.e., using a coordination strategy that would
allow them to make use of the best available muscle strength.
The long duration of the lifting movement (0.7 s) studied by
Sparto et al. (31) may have allowed the subjects to use the
information of the mechanical capacities of the muscles to
restructure the motor command at low level so that the kine-
matic pattern could be relatively adapted during early stages of
the movement. This is not the case of countermovement jumps,
where the ballistic characteristic and the short duration of the
propulsive phase do not allow the subjects to tune the control
signals as they perform the movement (29).
In the case of the study performed by Bonnard et al.
(8), it can be speculated that the submaximal nature of the
task (hopping) and the fact that this continuous exercise did
not involve extensively all major muscles of the lower limbs
allowed the subjects to compensate for the effects of fatigue
and sustain the height of the jumps. Van Zandwijk et al. (34)
demonstrated that the control signals used in maximal and
submaximal vertical jump performances are strongly re-
lated, but differ with respect to the activity of the biarticular
muscles, which are modulated differently (temporal shifting
and amplitude changes).
Although the coordination strategy has been shown to be
constant even after fatigue, only the transient (acute) effects
of fatigue were examined in the present study. It is not
known whether the neuromuscular system would reoptimize
the neural input when the movement is continually practiced
under fatigue conditions (i.e., under chronic fatigue condi-
tions). The simulation study performed by Bobbert and Van
Soest (7) provided some insight into the mechanisms of
adaptation that occur when the movement controls (i.e., the
neural input) are changed. They showed that increases in
performance after changing the muscle force-generating
properties (i.e., increased muscle strength) were achieved
only when the neural input was reoptimized. Bobbert and
Van Soest (7) suggested that improvements in performance
would occur after a period of practice if the subjects have
the opportunity to learn and adjust their coordination pattern
(i.e., neural input) to the new force-generating properties of
the muscles (i.e., a chronic adaptation). If the same effect
exists under fatigue conditions, then repeatedly practicing
countermovement jumps under the acute effects of fatigue
may induce the subjects to adopt a new coordination strat-
egy that may reinforce a coordination pattern (i.e., a chronic
adaptation) that is not optimal for maximal performances in
a nonfatigued state. As a consequence, training programs
inducing repetitive use of control strategies, in which the
properties of the muscles are not taken into account or are
not specific, may produce unsatisfactory results or even
reduced jump height performance. Further studies are nec-
essary to investigate experimentally whether and how the
neuromuscular system reorganizes/reoptimizes the neural
input and the movement pattern after a period of practice
(training) in which the capacity of the muscles to generate
force is changed (increased or decreased).
CONCLUSION
In conclusion, it has been shown that vertical jump perfor-
mance is affected by fatigue of the knee extensor muscles, but
not by fatigue of knee flexors. Despite the change in effective
muscle force and therefore jump height, there was no change in
the temporal characteristics of the muscle activation pattern as
indicated by surface EMG. This suggests that the same move-
ment strategy was followed before and after fatigue. On the
basis of the arguments proposed by Bobbert and Van Soest (7)
that the neural input has to be adjusted to take into account the
muscle force-generating properties to produce maximal perfor-
mances, it is speculated that the coordination strategy (neural
control) used after fatigue was no longer optimal for the muscle
strength available.
Address for correspondence: Andre´ Luiz Felix Rodacki, Univer-
sidade Federal do Parana´, Setor de Cieˆ ncias Biolo´ gicas, Departa-
mento de Educac¸a˜ o Fisica, R. Corac¸a˜ o de Maria, 92 - BR116, km
95, Jardim Botaˆnico, Curitiba, PR, Brazil CEP 80215-370; E-mail:
rodacki@ufpr.br.
VERTICAL JUMP COORDINATION: FATIGUE EFFECTS Medicine & Science in Sports & Exercise姞
115

REFERENCES
1. ACIERNO, S. P., R. V. BARATTA, and M. SOLOMONOV.A Practical
Guide to Electromyography for Biomechanists. New Orleans:
Louisiana Medical Centre, Bioengineering Laboratory, Louisiana
State University, 1995, pp. 4–8.
2. BACA, A. A comparison of methods for analysing drop jump
performance. Med. Sci. Sports Exerc. 31:437–442, 1999.
3. BARATTA, R., M. SOLOMONOW,B.H.ZHOU,D.LETSON,R.CHUI-
NARD, and R. D’AAMBROSIA. Muscular co-activation: the role of the
antagonist musculature in maintaining knee stability. Am. J. Sports
Med. 16:113–122, 1988.
4. BIGLAND-RITCHIE, B. EMG/force relations and fatigue of human
voluntary muscle contractions. Exerc. Sci. Sports Rev. 9:75–117,
1981.
5. BOBBERT, M. F., and E. D. DEBRUIN. Training of muscle strength
and control in vertical jumping. Proceedings of the 3rd Annual
Congress-European College of Sport Science, Manchester, UK,
1998. Liverpool, UK: Centre for Health Care Development, 1998,
p. 15.
6. BOBBERT, M. F., and G. J. VAN INGEN SCHENAU. Co-ordination in
vertical jumping. J. Biomech. 21:249–262, 1988.
7. BOBBERT, M. F., and A. J. VAN SOEST. Effects of muscle strength-
ening on vertical jump height: a simulation study. Med. Sci. Sports
Exerc. 26:1012–1020, 1994.
8. BONNARD, M., A. V. SIRIN,L.ODDSON, and A. THORSTENSSON.
Different strategies to compensate for the effects of fatigue re-
vealed by neuromuscular adaptation processes in humans. Neuro-
sci. Lett. 166:101–105, 1994.
9. BOSCO, C., J. TIHANYI,F.LATTERI,G.FEKETE,P.APOR, and H.
RUSKO. The effect of fatigue on store and re-use of elastic energy
in slow and fast types of human skeletal muscle. Acta Physiol.
Scand. 128:109–117, 1986.
10. DRILLIS, R. J., and R. CONTINI.Technical Report No. 1166–03.
New York: New York University, 1966, pp. 3–4.
11. EDWARDS, R. H. T. Human muscle function and fatigue. In: Human
Muscle Fatigue: Physiological Mechanisms, R. Porter and J.
Whelan (Eds.). London: Pitman Medical, 1981, pp. 1–18.
12. ELORANTA, V. Effect of postural and load variation on the co-
ordination of the leg muscles in concentric jumping movement.
Electromyogr. Clin. Neurophysiol. 36:59–64, 1996.
13. FIGONI, S. H., C. B. CHRIST, and B. H. MASSEY. Effects of speed,
hip and knee angle and gravity on hamstring to quadriceps torque
ratios. J. Orthop. Sports Phys. Ther. 9:287–291, 1988.
14. FITTS, R. H. Cellular mechanisms of muscle fatigue. Physiol. Rev.
74:49–94, 1994.
15. GOLLHOFER, A., P. V. KOMI,M.MIYASHITA, and O. AURA. Fatigue
during stretch-shortening exercises: changes in mechanical per-
formance of human skeletal muscle. Int. J. Sports Med. 8:71–78,
1987.
16. GOTTLIEB, G. L., D. M. CORCOS, and G. C. AGARWAL. Strategies for
the control of single mechanical degree of freedom voluntary
movements. Behav. Brain Sci. 12:189–210, 1989.
17. HAKKINEN, K., M. ALEN, and P. V. KOMI. Effect of explosive type
strength training on isometric force and relaxation-time, electro-
myographic and muscle fibre characteristics of leg extensor mus-
cles. Acta Physiol. Scand. 125:587–600, 1985.
18. HAUTIER, C. A., L. M. ARSAC,K.DEGHDEGH,J.SOUQUET,A.BELLI,
and J. LACOUR. Influence of fatigue on EMG/force ratio and
co-contraction in cycling. Med. Sci. Sports Exerc. 32:839–843,
2000.
19. HEISE, G. D., D. W. MORGAN,H.HOUGH, and M. CRAIB. Relation-
ship between running economy and temporal EMG characteristics
of bi-articular leg muscles. Int. J. Sports Med. 17:128–133, 1996.
20. HORTOB´
AGYI, T., N. J. LAMBERT, and W. P. KROLL. Voluntary and
reflex responses to fatigue with stretch-shortening exercise. Can.
J. Sport Sci. 16:142–150, 1991.
21. HUBLEY, C. L., and R. P. WELLS. Work-energy approach to deter-
mine individual joint contributions to vertical jump performance.
Eur. J. Appl. Physiol. 50:247–254, 1983.
22. JACOBS, R., and G. J. VAN INGEN SCHENAU. Control of an external
force in leg extensions in humans. J. Physiol. 457:611–626, 1992.
23. KELLIS, E., and V. BALTZOPOULOS. The effects of antagonist mo-
ment on the resultant knee joint moment during isokinetic testing
of the knee extensors. Eur. J. Appl. Physiol. 76:253–259, 1997.
24. MACINTYRE, D. L., R. W. DARLENE,D.M.LYSTER,I.J.SZASZ, and
D. C. MCKENZIE. Presence of WBC, decrease strength and delayed
soreness in muscle after eccentric exercise. J. Appl. Physiol. 80:
1006–1003, 1996.
25. PANDY, M. G., and F. E. ZAJAC. Optimal muscular co-ordination
strategies for jumping. J. Biomech. 24:1–10, 1991.
26. PINNIGER, G. J., J. R. STEELE, and H. GROELLER. Does fatigue
induced by repeated dynamic efforts affect hamstring muscle
function? Med. Sci. Sport Exerc. 2:647–653, 2000.
27. PSEK, J. A., and E. CAFARELLI. Behaviour of coactive muscles
during fatigue. J. Appl. Physiol. 74:170–175, 1993.
28. RODACKI, A. L. F., N. E. FOWLER, and S. J. BENNETT. Multi-segment
coordination: fatigue effects. Med. Sci. Sports Exerc. 33:1157–
1167, 2001.
29. SCHMIDT,R.A.Motor Control and Learning: A Behavioural
Emphasis. Champaign, IL: Human Kinetics, 1982, pp. 48–75.
30. SNOW, C. J., J. COOPER,A.O.QUANBURY, and J. E. ANDERSON.
Antagonist co-contraction of knee extensors during constant ve-
locity muscle shortening and lengthening. J. Electromyogr. Kine-
siol. 5:185–192, 1995.
31. SPARTO, P. J., M. PARNIANPUOR,T.E.REINSEL, and S. SIMON. The
effect of fatigue on multi-joint kinematics, co-ordination, and
postural stability during a repetitive lifting test. J. Orthop. Sports
Phys. Ther. 25:3–12, 1997.
32. TESCH, P. A., G. A. DUDLEY,M.R.DUVOISIN,B.M.HATHER, and
H. T. HARRIS. Force and EMG signal patterns during repeated
bouts of concentric and eccentric muscle actions. Acta Physiol.
Scand. 138:263–271, 1990.
33. VAN INGEN SCHENAU, G. J., A. J. VAN SOEST,F.J.M.GABRE ¨
ELS, and
M. W. I. M. HORSTINK. The control of multi-joint movements relies
on detailed internal representations. Hum. Movement Sci. 14:531–
538, 1995.
34. VAN ZANDWIJK, J. P., M. F. BOBBERT,M.MUNNEKE, and P. PAS.
Control of maximal and sub-maximal vertical jumps. Med. Sci.
Sports Exerc. 32:477–485, 2000.
35. VIITASALO, J. T., K. H¨
AM¨
AL¨
AINEM,H.MONONEN,A.SALO, and J.
LAHTINEN. Biomechanical effects of fatigue during continuous
hurdle jumping. J. Sports Sci. 11:503–509, 1993.
36. VIITASALO, J. T., A. SALO, and J. LAHTINEN. Neuromuscular func-
tioning of athletes and non-athletes in the drop jump. Eur. J. Appl.
Physiol. 78:432–440, 1998.
37. WEIR, J. P., D. A. KEEFE,J.F.EATON,R.T.AUGUSTINE, and D. M.
TOBIN. Effect of fatigue on hamstring co-activation during isoki-
netic extensions. Eur. J. Appl. Physiol. 78:555–559, 1998.
38. WITTEKOPFT, G., E. SCHAAF, and H. TAUBENHEIM. Use of electro-
myography for quantification of local muscular fatigue following
a known strength-endurance load. In: Biomechanics VI-A,E.As-
mussen and K. Jorgensen (Eds.). Baltimore: University Park Press,
1978, pp. 185–193.
116
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org