J Appl Physiol 117: 1132–1140, 2014.
First published September 11, 2014; doi:10.1152/japplphysiol.00561.2014.
Between-muscle differences in the adaptation to experimental pain
X François Hug,1,2 Paul W. Hodges,1 Wolbert van den Hoorn,1 and Kylie Tucker1,3
1
The University of Queensland, NHMRC Centre of Clinical Research Excellence in Spinal Pain, Injury and Health,
School of Health and Rehabilitation Sciences, Brisbane, Australia; 2University of Nantes, Laboratory “Motricité, Interactions,
Performance” (EA 4334), UFR STAPS, F-44000, Nantes, France; and 3The University of Queensland, School of Biomedical
Sciences, Brisbane, Australia
Submitted 26 June 2014; accepted in final form 10 September 2014
quadriceps; noxious stimulation; electromyography; elastography; supersonic shear imaging
THE EFFECTS of experimentally induced pain on the motor
system have been extensively studied during single-joint tasks.
The observed adaptations (e.g., changes in motor unit discharge characteristics, gross muscle activity, and/or mechanical behavior) are thought to reduce stress (defined as force per
unit area) within the painful region to protect from further pain
and/or injury (10, 16).
Supersonic shear imaging (SSI) is an elastography technique
that quantifies the shear elastic modulus (stiffness) of a localized area of tissue. This technique provides reliable measurement of muscle shear elastic modulus (15), with a strong linear
relationship between shear elastic modulus and muscle stress
during isometric contractions (2, 4) and passive stretching (12,
17). Therefore, SSI provides a unique opportunity to quantify
Address for reprint requests and other correspondence: François Hug, Centre
of Clinical Research Excellence in Spinal Pain, Injury and Health, School of
Health and Rehabilitation Sciences, The Univ. of Queensland, St. Lucia, QLD
4072, Australia (e-mail: francois.hug@univ_nantes.fr).
1132
if stress within muscle tissue is modified during acute pain.
Taking advantage of this technique, we have previously shown
no systematic reduction in stress within a painful muscle
[vastus lateralis (VL) (24) or vastus medialis (VM) (11)]
during isometric tasks. In these studies (11, 24) participants
were instructed to match a knee extension torque before and
during pain. To maintain torque a decrease in stress (i.e., force)
in VL or VM (i.e., the painful muscle) would have necessitated
compensation (i.e., increased stress/force) with other knee
extensor muscles. Whether the central nervous system (CNS)
can differentiate the activation of the individual heads of the
quadriceps muscle is a matter of ongoing debate. Therefore, the
absence of deloading painful tissues during pain (11, 24) may
be a consequence of a limited potential to redistribute force
between synergist muscles in the previous investigations.
Although there is some indirect evidence for differential
activation of the quadriceps muscles during submaximal contractions (3, 13, 14, 21), these differences are subtle (3, 13)
and/or limited to compensations between the biarticular rectus
femoris (RF) muscle and the vastii (14, 21). Further, the high
common drive reported between VL and VM (18) leads to the
assumption that a dissociation of activity between these muscles might be difficult or might only occur with high cost to the
CNS. As the volume of RF represents only ⬃16% of the total
volume of the vastii muscles (9), RF would only have the
capacity to successfully counterbalance minor changes in force
produced by the vastii. Together, the inability to dissociate VL
and VM activation, and the limited capacity for RF to compensate for a substantial reduction in force output from the
vastii, may explain why no systematic decrease in stress within
VL or VM was observed when pain was induced in these
muscles while maintaining isometric knee extension torque
(11, 24). However, because RF is controlled more independently from the vastii than the vastii muscles are from each
other (21), it is more likely that pain in RF could be associated
with systematic reduction in stress in this painful muscle and
compensation by the vastii. Studying pain adaptation when
pain is induced into one vastii muscle or RF during an isometric leg extension may allow us to determine whether the
absence of deloading painful tissues during pain (11, 24) is a
consequence of the limited potential to redistribute force between some synergist muscles. Ultimately, this work will
provide insight into the developing theories of pain adaptation
that predict a change in muscle stress (10). In a broader motor
control context, these results will provide further evidence as to
whether the individual heads of the quadriceps muscle can be
independently recruited. This is particularly relevant as it is
generally assumed that the altered function of a single muscle
can be compensated within the same muscle group.
8750-7587/14 Copyright © 2014 the American Physiological Society
http://www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
Hug F, Hodges PW, van den Hoorn W, Tucker K. Betweenmuscle differences in the adaptation to experimental pain. J Appl
Physiol 117: 1132–1140, 2014. First published September 11, 2014;
doi:10.1152/japplphysiol.00561.2014.—This study aimed to determine whether muscle stress (force per unit area) can be redistributed
between individual heads of the quadriceps muscle when pain is
induced into one of these heads. Elastography was used to measure
muscle shear elastic modulus (an index of muscle stress). Electromyography (EMG) was recorded from vastus lateralis (VL), vastus
medialis (VM), and rectus femoris (RF). In experiment I (n ⫽ 20),
participants matched a knee extension force, and thus any reduction of
stress within the painful muscle would require compensation by other
muscles. In experiment II (n ⫽ 13), participants matched VL EMG
amplitude and were free to vary external force such that intermuscle
compensation would be unnecessary to maintain the experimental
task. In experiments I and II, pain was induced by injection of
hypertonic saline into VM or RF. Experiment III aimed to establish
whether voluntary drive to the individual muscles could be controlled
independently. Participants (n ⫽ 13) were asked to voluntarily reduce
activation of VM or RF while maintaining knee extension force.
During VM pain, there was no change in shear elastic modulus
(experiments I and II) or EMG amplitude of VM (experiment II). In
contrast, RF pain was associated with a reduction in RF elastic
modulus (experiments I and II: ⫺8 to ⫺17%) and EMG amplitude
(experiment II). Participants could voluntarily reduce EMG amplitude
of RF (⫺26%; P ⫽ 0.003) but not VM (experiment III). These results
highlight between-muscle differences in adaptation to pain that might
be explained by their function (monoarticular vs. biarticular) and/or
the neurophysiological constraints associated to their activation.
Between-Muscle Differences in Pain Adaptation
MATERIALS AND METHODS
Participants
Twenty healthy volunteers (means ⫾ SD: age 29 ⫾ 5 yr, weight
72 ⫾ 10 kg, height 176 ⫾ 11 cm; 4 females) participated in
experiment I. Thirteen healthy volunteers (age 27 ⫾ 6 yr, weight 71 ⫾
16 kg, height 175 ⫾ 13 cm; 5 females) participated in experiments II
and III. Participants provided informed written consent. The local
ethics committee approved the experiment (The University of
Queensland/CPP Nantes Ouest IV), and all procedures adhered to the
Declaration of Helsinki.
1133
Hug F et al.
used to calculate the shear wave velocity (Vs) along the principal axis
of the probe using a time-of-flight estimation. Considering a linear (1)
and elastic (5) behavior, the muscle shear elastic modulus (␮) is
calculated as follows:
␮ ⫽ ␳Vs2
where ␳ is the density of muscle (1,000 kg/m3). Maps of the shear
elastic modulus were obtained at 1 sample/s.
Using B-mode ultrasound imaging, the optimal transducer location
was determined for each muscle (RF and VM) as a region with a
muscle thickness of at least 1.5 cm and that avoided major hypoechoic
regions related to aponeurosis and tendon. The transducer was aligned
approximately in parallel with the muscle fiber direction of VM
(obliquus region). Because of the complex architecture of RF, the
ultrasound transducer was aligned with the shortening direction of the
muscle, consistent with previous work on RF (3). Previous experiments have demonstrated that muscle architecture (i.e., fusiform vs.
pennate muscle) does not influence the reliability of the estimation of
changes in muscle stress using SSI (4). The optimal transducer
locations were marked on the skin using a waterproof marker to guide
repeated placement throughout the trials.
Surface electromyography. For experiments II and III, myoelectrical activity was recorded from the right (test) leg with surface EMG
electrodes on VM, VL, and RF. For each muscle, a pair of selfadhesive Ag/AgCl electrodes (Blue sensor N-00-S, Ambu, Denmark)
was attached to the skin with an interelectrode distance of 20 mm
(center to center) (location shown in Fig. 2). Prior to electrode
application, the skin was cleaned with abrasive gel (Nuprep, D.O.
Weaver) and alcohol. The ground electrode (half of a Universal
Electrosurgical Pad, 3M Health Care) was placed on the skin over the
tibia of the right leg. EMG data were amplified 1,000 times, band-pass
filtered between 20 and 500 Hz (Neurolog, Digitimer), and sampled at
2,000 samples/s using a Power1401 Data Acquisition System with
Spike2 software (Cambridge Electronic Design).
Experimental pain. For experiments I and II, pain was induced by
a 0.5-ml, single-bolus injection of hypertonic saline (5% NaCl) using
a 25 G ⫻ 24 mm needle, into either the distal portion of RF (PainRF)
or VM (PainVM). Pain intensity was rated on an 11-point numerical
rating scale (NRS; anchored with “no pain” at 0 and “maximum
imaginable pain” at 10). After completion of the contractions during
pain, participants drew a surface representation of the region of pain
directly onto their leg, and a photograph was taken.
Measurements
Ergometer. Participants were seated on a plinth with their back and
upper legs supported. Their torso was reclined by 10° from upright
and their right knee (test leg) flexed to 65° from the horizontal. A
standard support strap was placed around the pelvis to minimize
changes in body position throughout the experimental task, and
participants crossed their arms over their chest. Isometric knee extension force was measured with a six-axis force sensor (measuring range
for Fz 500 N, resolution 0.01% of measuring range, Poitiers, France).
Signals were sampled at 2,000 samples/s (Power1401 Data Acquisition System, Cambridge Electronic Design, UK, or Bagnoli 16,
Delsys, Natick, MA).
Shear elastic modulus measurements. An Aixplorer ultrasound
scanner (version 6; Supersonic Imagine, Aix-en-Provence, France),
coupled with a linear transducer array (4 –15 MHz, SuperLinear 15-4,
Vermon, Tours, France) was used in shear wave elastography mode
(musculoskeletal preset). As originally described by Bercoff et al. (1),
the technique consists of a transient and remote mechanical vibration
generated by radiation force induced by a focused ultrasonic beam
(i.e., “pushing beam”). Each pushing beam generates a remote vibration that results in the propagation of a transient shear wave. Then, an
ultrafast ultrasound imaging sequence is performed to acquire successive raw radiofrequency data at a very high frame rate. A onedimensional cross correlation of successive radiofrequency signals is
Protocol
Experiment I. Two maximal isometric voluntary knee extensions
were performed for 3 s, separated by 90 s for recovery. The maximum
force was considered the best performance (maximum voluntary
contraction, MVC). The experimental task required participants to
match a target force set at 20% of MVC during short (12 s) constantforce isometric contractions. This contraction intensity was chosen for
two reasons. First, associated with the relatively short duration of the
contractions, the moderate force level limited the potential for neuromuscular fatigue that could interfere with interpretation of the data.
Second, this enabled direct comparison with a previous pain experiment that served as a basis for this series of experiments (11).
For the baseline condition, participants performed 10 ⫻ 12-s
contractions with a 30-s recovery period between contractions (Fig. 1).
Then, muscle pain was induced into either the distal portion of RF
(PainRF) or VM (PainVM) (Fig. 2A). When pain intensity was reported
as at least 2/10, another 3 ⫻ 12 s isometric knee extensions was
performed. The task was stopped if pain intensity dropped below 2/10.
Approximately 10 min after complete resolution of pain, the protocol
(including 10 baseline and 3 pain contractions) was repeated, but with
hypertonic saline injected into the second muscle (either RF or VM).
The order of pain location was randomized (PainRF was the first
condition for 9/20 participants).
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
Here we used SSI to determine whether muscle stress could
be redistributed between individual heads of the quadriceps
muscle when pain is induced in one of these heads. This study
involved three experiments. The first (experiment I) required
participants to match an isometric knee extension force at 20%
of maximum voluntary contraction force (MVC), and thus any
reduction of stress within the painful muscle would require
compensation by other synergist muscles. This experiment
aimed to determine whether stress is reduced within RF or VM
during homonymous nociceptive stimulation. We hypothesized
that a decrease in RF stress would be observed during nociceptive stimulation of this muscle whereas stress in VM would
not be reduced when it is the painful muscle. The second
experiment (experiment II) explored whether stress within RF
or VM could be reduced, independently of VL, during nociceptive stimulation of these muscles. Participants matched VL
EMG amplitude (rather than knee extension force). Thus they
were free to vary external force such that intermuscle compensation would be unnecessary to maintain the experimental task.
We hypothesized that RF muscle activity and stress would
decrease during RF pain, but VM muscle activity and stress
would remain unchanged during VM pain. In the third experiment (experiment III; performed without pain) we investigated
whether participants could voluntarily reduce the activation of
VM or RF (with EMG feedback) while maintaining the knee
extension force at 20% of MVC.
•
1134
Between-Muscle Differences in Pain Adaptation
•
Hug F et al.
During contractions, the ultrasound transducer was placed over the
distal region of RF (for both PainRF trials and their corresponding
baseline trials) or over the distal portion of VM muscle belly (for both
PainVM trials and their corresponding baseline trials).
Experiments II and III. These experiments were performed within
the same session. The MVC procedures were identical to those
described for experiment I. To determine the target amplitude of VL
EMG for experiment II, participants first performed 3 ⫻ 12 s isometric
reference knee extensions (30 s recovery) at 20% of MVC. The RMS
EMG of VL was calculated in real time over a 200-ms window
(Spike2, Cambridge electronic) and averaged across the three isometric knee extensions. The experimental task required participants to
match the RMS EMG activity recorded during the preceding reference
contractions (Fig. 1). For the baseline condition, participants performed 6 ⫻ 12-s isometric knee extensions at the VL RMS EMG
target with a 30-s recovery period between contractions. In the pain
trials, muscle pain was induced as described above, into either the
distal portion of RF (PainRF) or VM (PainVM) (Fig. 2B). When the
participants reached a pain intensity of at least 2/10, 6 ⫻ 12-s
isometric knee extensions were performed to an intensity that matched
the VL RMS EMG target. The task was stopped if pain intensity was
reported below 2/10. Approximately 10 min after complete cessation
of pain, the protocol (including 6 ⫻ 12-s baseline and 6 ⫻ 12-s pain
contractions) was repeated, with pain induced in the second muscle
(either RF or VM). The order of pain location was randomized
(PainRF was the first condition for 7/13 participants). The second
series of 6 ⫻ 12 s contractions (no pain) was considered as both the
washout of the first pain condition and the baseline of the second pain
condition. A final set of 6 ⫻ 12-s isometric knee extensions was
performed ⬃10 min after complete cessation of pain. This condition
was considered as the washout of the second pain condition (Fig. 1).
During this experiment, the ultrasound transducer was relocated for
each contraction, alternatively on VM and RF (the first recorded
muscle was randomized) such that the shear elastic modulus of both
muscles could be assessed in each condition.
After the completion of experiment II, additional tasks were performed to test the participants’ ability to voluntarily reduce activity of
the RF or VM muscles while maintaining the knee extension force at
20% of MVC (experiment III). First, participants performed 6 ⫻ 12-s
isometric knee extensions (30 s recovery) at 20% of MVC (i.e., force
feedback), and the mean RF and VM RMS EMG amplitudes were
calculated. Then, participants performed two trials of 3 ⫻ 20-s
isometric knee extensions (30 s recovery between the contractions and
3 min recovery between the 2 trials) during which they were asked to
voluntarily reduce either RF (referred to as “RF trials”) or VM
(referred to as “VM trials”) EMG amplitude while maintaining the
knee extension force within ⫾5% of the target force. The order of the
RF and VM trials was randomized. Feedback of both EMG amplitude
of the targeted muscle (EMG RMS calculated online over a 200-ms
window using Spike2 software, Cambridge Electronic Design) and
force was provided to the participants. Participants were instructed to
attempt to reduce the activity of the targeted muscle while maintaining
force by altering their motor strategy, using any means possible, while
remaining in the seated position. For each trial, the lowest EMG
amplitude produced by the targeted muscle (VM or RF) over a 3-s
window, when the force was maintained within ⫾5% of the target
force, was calculated.
Data Processing
All the data were processed using Matlab (version R2012a, The
Mathworks).
Shear elastic modulus. For experiments I and II, muscle shear
elastic modulus was analyzed from the middle 8 s (8 images) of the
12-s contractions. SSI recordings were exported in mp4 format and
sequenced in png. Image processing converted the colored map into
shear elastic modulus values. Before analysis of the shear elastic
modulus, the map was inspected for artefacts. If artefacts were present
in any image, the region of interest (ROI) was reduced in size to
remove the area of artefact from all images within that contraction.
The average areas of the ROIs used for analysis were ⬇15 and ⬇20
mm2 for RF and VM, respectively. The shear elastic modulus was
averaged across the ROI and between the eight images to provide a
representative value for each contraction. For each location, the
average value calculated over the baseline contractions and over the
pain contractions was considered for further analysis.
Surface electromyography. EMG signals were rectified and lowpass filtered (4 Hz, Butterworth filter, 2nd order). EMG amplitude was
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
Fig. 1. Experimental design. For experiment
I (A), the task required the participant to
match their knee extension force for 12 s to
a target force set at 20% of maximum voluntary contraction (MVC). For the experiment II (B), the task required the participant
to match a RMS EMG activity level of vastus lateralis (VL) for 12 s corresponding to
the activity measured at 20% of MVC. For
both experiments, the baseline conditions
were followed by injection of hypertonic
saline into either the rectus femoris (RF) or
vastus medialis (VM). Then, participants
performed another 3 (experiment I) or 6
(experiment II) 12-s contractions with pain
followed by a baseline condition (experiment
II). WO, washout. *These contractions performed to determine the target amplitude of
VL EMG.
Between-Muscle Differences in Pain Adaptation
•
Hug F et al.
1135
RESULTS
Experiment I
Fig. 2. Reported regions of pain and approximate location of the ultrasound
transducer, EMG electrodes, and saline injections. The area of reported pain
for each participant is shown for PainRF and PainVM. Rectangles indicate the
approximate location of the ultrasound transducer. Arrows indicate the location
of saline injection for experiments I (A) and II (B). Circles indicate the
approximate location of the surface EMG electrodes (experiment II).
quantified over the same 8-s period as that used for shear elastic
modulus data (experiment II). These values were normalized to the
maximal EMG amplitude calculated over a 500-ms window during
MVC.
Statistical Analysis
Data distributions consistently passed the Shapiro-Wilk normality
test and are therefore reported as means ⫾ SD throughout the text and
the figures. Significance was set at P ⬍ 0.05. Cohen’s d values (SD of
the difference used for standardization) are reported as measures of
effect size, with 0.2, 0.5, and 0.8 as small, moderate, and large effect,
respectively (6).
Experiment I. Three participants did not complete all conditions:
one participant did not experience pain during PainRF following the
hypertonic saline injection, and two participants felt mild dizziness
following their second pain injection (one each in PainRF and
PainVM). Thus data were available for analysis from 18 and 19
participants for the PainRF and PainVM, respectively. As our hypotheses do not imply a comparison between PainRF and PainVM trials,
and because including both PainRF and PainVM trials would reduce the
Pain intensity and location. In some participants, pain
intensity was reported below 2/10 before completion of the
three pain contractions, and subsequently the pain trial was
ceased. For PainRF, trials were ceased after two contractions
in nine participants, and one contraction in one participant.
For PainVM, trials were ceased after two contractions in
eight participants, and after one contraction in two participants. The mean pain intensity reported during the pain
trials (where pain was reported ⱖ2/10) was 4.1 ⫾ 1.2 for
PainRF and 4.2 ⫾ 1.1 for PainVM. Pain was reported around
the site of the hypertonic saline injection, except for two
participants who reported pain in an area that also extended
over the patella during PainVM (Fig. 2A).
Knee extension force. The target force level was accurately
matched between conditions [standard error of measurement
(SE) between baseline and pain conditions was 2.2% and 1.5%
for PainRF and PainVM, respectively]. There was no significant
difference between conditions for PainRF (20.7 ⫾ 1.2 vs.
20.8 ⫾ 1.3% of MVC for baseline and pain, respectively; P ⫽
0.25; d ⫽ 0.27) or PainVM (20.7 ⫾ 1.3 vs. 20.7 ⫾ 1.5% of
MVC; P ⫽ 0.89; d ⫽ 0.03).
Muscle shear elastic modulus. A representative example of
the shear elastic modulus map is depicted in Fig. 3. During the
baseline contractions the mean shear elastic modulus was
25.5 ⫾ 12.1 kPa in RF, and this decreased significantly to 21.0 ⫾
12.5 kPa (i.e., ⫺17 ⫾ 33%; P ⫽ 0.021; d ⫽ ⫺0.60; Fig. 4A)
when nociceptive stimulation was induced in this muscle. The
decrease in shear elastic modulus indicates that the force
produced by RF during homonymous muscle pain decreased
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
sample size as a result of casewise deletion of missing values (by the
3 participants discussed above), data were analyzed separately. Thus,
for both PainRF and PainVM trials, shear elastic modulus and knee
extension force (in % of MVC) were compared between conditions
(baseline and pain) using paired t-tests.
Experiment II. One participant did not maintain VL EMG consistently within the target range (error ⬎10%), and their data were
discarded, leaving data for analysis from 12 participants. As for
experiment I, data were analyzed separately for each pain location
(PainRF and PainVM). Isometric knee extension force (in % of MVC)
and VL EMG amplitude (feedback) were compared between conditions (baseline, pain, washout) with a repeated-measures ANOVA. To
assess the putative difference in muscle coordination strategy between
the force-matched contractions and VL EMG-matched contractions (6
contractions of the first baseline condition), normalized EMG amplitude was compared between muscles (RF and VM, within factor) and
contractions (force-matched vs. VL EMG-matched, within factor)
with a repeated-measures ANOVA. To determine the adaptations
during pain, both shear elastic modulus and normalized EMG amplitude were compared between muscles (RF and VM; within factor) and
conditions (baseline, pain, washout; within factor) using a repeatedmeasures ANOVA. Post hoc analyses were performed using the
Fisher LSD test. Finally, the relationship between changes in muscle
shear elastic modulus during pain (in kPa, if any) and changes in RF
EMG amplitude (in %EMGmax) or changes in knee extension force (in
% of MVC) was tested using Pearson’s correlation coefficient.
Experiment III. Finally, potential changes in EMG amplitude when
participants attempted to voluntarily change RF or VM EMG were
tested using a repeated-measures ANOVA (muscles and conditions as
within factor) for each trial (VM and RF trial), separately. Post hoc
analyses were performed using the Fisher LSD test.
1136
Between-Muscle Differences in Pain Adaptation
•
Hug F et al.
Fig. 3. Representative example of the shear
elastic modulus maps obtained in RF (top
panels) and VM (bottom panels) during experiment II. The map of shear elastic modulus is superposed onto a B-mode image,
with the color scale depicting graduation of
shear elastic modulus. To obtain a representative value, the shear elastic modulus was
averaged over the region of interest indicated by a white rectangle. When pain was
induced in VM (PainVM), there was no
change in shear elastic modulus. In contrast,
RF pain (PainRF) was associated with a reduction in RF shear elastic modulus.
Experiment II
Pain intensity and location. In some participants, pain intensity was reported below 2/10 before completion of the six
Pain contractions, and the trial was ceased. For PainRF, trials
were ceased after four contractions (i.e., 2 SSI measurements
per muscle) in four participants. For PainVM, trials were ceased
after four contractions in four participants, and after five
contractions in one participant. The mean pain intensity reported during the Pain trials (where pain was reported ⱖ2/10)
was 3.8 ⫾ 1.1 for PainRF and 4.4 ⫾ 1.3 for PainVM. Pain was
perceived around the site of the saline injection, except for one
participant who reported pain in an area that extended over the
patella during PainVM (Fig. 2B).
Knee extension force. In contrast to experiment I, participants were asked to maintain VL EMG activity during experiment II. Thus knee extension force could change. There was a
significant main effect of condition during PainRF (P ⫽
0.0004). Knee extension force was less when pain was induced
by saline injection into RF (15.6 ⫾ 5.6% of MVC) than the
preceding baseline trial (18.8 ⫾ 5.9% of MVC; P ⫽ 0.0001,
d ⫽ ⫺1.17). Force remained lower during washout (16.8 ⫾
5.7% of MVC) compared with the baseline (P ⫽ 0.008; d ⫽
Fig. 4. Muscle shear elastic modulus during baseline and pain (experiment I).
Although pain induced a significant decrease in shear elastic modulus in RF
during homonymous muscle pain (A), no changes in VM were observed when
pain was induced in VM (B). Mean and SD are shown. *P ⬍ 0.05 for
comparison with baseline (d ⫽ ⫺0.60; moderate effect).
⫺0.78). No change in force was observed between baseline,
pain, and washout conditions when pain was induced in VM,
i.e., PainVM (main effect of condition: P ⫽ 0.29) (Fig. 5).
Shear elastic modulus. Although the condition ⫻ muscle
interaction was not significant (P ⫽ 0.42) during PainRF, there
was a significant main effect of muscle (P ⫽ 0.0002) and
condition (P ⫽ 0.004) (Fig. 6A). When pain was induced in RF
the shear elastic modulus (pooled between RF and VM) was
lower during pain (26.3 ⫾ 13.0 kPa) than during baseline
(30.1 ⫾ 12.2 kPa; P ⫽ 0.006; d ⫽ ⫺0.86) and washout (30.7 ⫾
13.1 kPa; P ⫽ 0.002; d ⫽ ⫺1.05). There was no difference
between washout and baseline (P ⫽ 0.66; d ⫽ 0.13). A
significant correlation was found between the changes in RF
shear elastic modulus during pain and the changes in knee
extension force (r ⫽ 0.60, P ⫽ 0.036). Although a significant
main effect of muscle was found, during PainVM (P ⫽ 0.001)
there was neither a main effect of condition (P ⫽ 0.40) nor an
interaction between condition ⫻ muscle (P ⫽ 0.80) (Fig. 6B).
Electromyography. EMG was not recorded from VM and
RF for participant 1 in experiment II. Thus EMG data for these
two muscles were available for 11 participants. During the first
three force-matched contractions performed at 20% of MVC,
EMG amplitude was 13.1 ⫾ 3.7, 13.1 ⫾ 4.9, and 11.9 ⫾ 5.2%
of EMGmax for VL, VM, and RF, respectively. Muscle coordination strategy did not change between the force-matched
Fig. 5. Change in knee extension force during pain and washout conditions
(experiment II). Experiment II involved contractions maintained at target VL
EMG amplitude and thus allowing the knee extension force to be modified.
Although pain induced a significant decrease in knee extension force after
saline injection into RF (A), no systematic changes were observed when saline
was injected into VM (B). *P ⬍ 0.05 for comparison with baseline [d ⫽ 1 .1
(large effect) and 0.73 (moderate effect) for pain and washout, respectively].
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
from that in the baseline condition. In contrast, when nociceptive stimulation was induced in VM, there was no significant
change in shear elastic modulus (32.3 ⫾ 12.5 vs. 31.0 ⫾ 13.3
kPa for the baseline and pain conditions, respectively; P ⫽
0.23; d ⫽ ⫺0.19; Fig. 4B).
Between-Muscle Differences in Pain Adaptation
•
Hug F et al.
1137
Fig. 6. Muscle shear elastic modulus during
VL EMG-matched trials (experiment II).
When pain was induced in RF, the shear
elastic modulus was lower during pain compared with both baseline and washout (with no
different behavior between muscles as the
muscle ⫻ condition interaction was not significant) (A). In contrast, pain did not affect muscle shear elastic modulus during PainVM (B).
shear elastic modulus (r ⫽ 0.59; P ⫽ 0.05). When pain was
induced in VM (PainVM), no changes were observed for RF
or VM EMG amplitude (effect muscle: P ⫽ 0.18; effect
condition: P ⫽ 0.55; and muscle ⫻ condition interaction:
P ⫽ 0.41) (Fig. 7B).
Experiment III
When the participants were asked to voluntary reduce RF
EMG while maintaining the knee extension force (⫾ 5%),
there was a significant muscle ⫻ condition interaction (P ⫽
0.018). RF decreased from 11.3 ⫾ 4.9 to 8.5 ⫾ 4.5% of
EMGmax (P ⫽ 0.011; d ⫽ ⫺1.1 Fig. 8). EMG amplitude of
neither VM (P ⫽ 0.18; d ⫽ 0.37) nor VL (P ⫽ 0.53, d ⫽ 0.19)
changed significantly. During the VM trials there was neither a
main effect of condition (P ⫽ 0.18), muscle (P ⫽ 0.87), or a
muscle ⫻ condition interaction (P ⫽ 0.08). This indicates that
the participants failed to voluntarily decrease VM EMG (d ⫽
⫺0.38). EMG of VL (d ⫽ ⫺0.12) and RF (d ⫽ 0.51) were also
unchanged during this task.
DISCUSSION
This series of experiments has three main observations.
First, RF stress significantly reduces when pain is induced in
this muscle during isometric knee extension (experiment I).
Second, when feedback of VL EMG activity was provided
Fig. 7. EMG amplitude during VL EMGmatched trials (experiment II). When pain was
induced in RF, RF EMG amplitude was lower
during pain compared with both baseline and
washout. However, VM EMG amplitude was
not affected (A). When pain was induced in VM,
RF and VM EMG amplitude was not systematically modified (B). *P ⬍ 0.05 for comparison
with baseline (d ⫽ ⫺0.93; large effect); #P ⬍
0.05 for comparison to washout (d ⫽ ⫺0.85;
large effect).
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
and VL EMG-matched contractions. When VM and RF EMG
amplitudes were compared between these force-matched contractions at 20% of MVC (those used to determine VL EMG
level for the feedback) and the first baseline condition (when
VL RMS EMG level was used as feedback), there was no
systematic difference for either PainRF (main effect muscle:
P ⫽ 0.64; main effect condition: P ⫽ 0.59; interaction condition ⫻ muscle: P ⫽ 0.81) or PainVM (main effect muscle: P ⫽
0.32; main effect condition: P ⫽ 0.56; interaction condition ⫻
muscle: P ⫽ 0.16).
The target VL EMG amplitude was accurately matched
between conditions (SE between baseline and pain ⫽ 0.4% and
1.8% for PainRF and PainVM, respectively). VL EMG did not
differ between conditions for either PainRF (main effect condition: P ⫽ 0.29) or PainVM (main effect condition: P ⫽ 0.44).
When pain was induced in RF (PainRF) there was no effect
of muscle (P ⫽ 0.18) on EMG amplitude. However, there was
a significant condition ⫻ muscle interaction (P ⫽ 0.013). EMG
amplitude of RF was less (8.2 ⫾ 3.3% of EMGmax) during pain
than during both baseline (12.5 ⫾ 4.5% of EMGmax; P ⬍
0.0001; d ⫽ ⫺0.93) and washout (11.3 ⫾ 4.3% of EMGmax;
P ⫽ 0.0003; d ⫽ ⫺0.85) (Fig. 7A). There was no difference in
RF EMG between baseline and washout (P ⫽ 0.10), and no
change was observed for VM EMG between any condition
(all post hocs: P ⬎ 0.15). The changes observed in RF EMG
during pain were positively correlated to the changes in RF
1138
Between-Muscle Differences in Pain Adaptation
Fig. 8. Change in EMG amplitude during the RF and VM trials of experiment
III (voluntary changes). Participants were instructed to decrease RF (RF trials;
A) or VM (VM trials; B) EMG amplitude while maintaining the knee extension
force. *P ⬍ 0.05 for comparison with baseline (d ⫽ ⫺1.1; large effect).
Reduced Stress Within RF During Homonymous Muscle
Nociceptive Stimulation
When experimental pain is induced by noxious stimulation
of RF, stress within this muscle decreased. This systematic
decrease in RF muscle stress was observed when either knee
extension force (experiment I) or VL EMG activity (experiment II) was matched between the contractions without and
with pain. Reduced RF EMG in experiment II indicates the
decrease in muscle stress during pain is explained (at least
partly) by reduced neural drive to RF rather than subtle
changes in pelvis position that could have reduced passive
stress (and thus the overall muscle stress) without change in
muscle activity.
The dissociation of neural drive between RF and the vastii
muscles observed here concurs with previous observations of
clear dissociation between EMG activity of these muscles
during knee extension performed at low [2.5 to 5% of MVC
(14)] and moderate contraction intensities [⬃40% of MVC
(21)]. The ability to dissociate the drive of RF from the vastii
enables reduction of stress in the painful RF while maintaining
force (experiment I) or neural drive to VL (experiment II).
Experiment III confirms the potential to also dissociate activity
voluntarily while maintaining knee extension force. Taken
together, these data indicate the CNS can dissociate the activity
of RF and vastii. A disassociation seems logical given the
different functional roles (knee extensor alone vs. knee exten-
Hug F et al.
sor and hip flexor) and anatomy (mono- vs. biarticular) of the
RF and vastii. Monoarticular muscles are mainly responsible
for the generation of positive torque, whereas the biarticular
muscles have a unique role in controlling the distribution of net
moments at the joint, and thus the direction of external force
(25, 26).
Performance of isometric knee extension with VL EMG
feedback did not theoretically constrain RF activity. This is
supported by the reduction in RF activation (EMG amplitude)
during pain. Considering the total energy cost of the contraction, it may be energetically more optimal to keep this reduced
RF activity during washout, but this was not observed (no
difference between baseline and washout). This implies that
some advantage was derived from greater activation of RF.
One possibility is that there was a greater neural “cost” to
maintain activity of VL with lesser RF contribution. For
instance, this may have required suppression of a preferred
muscle synergy. This result concurs with previous observations
that patterns of muscle coordination are organized on the basis
of habit rather than optimization of energy or other costs (8).
Alternatively, it is possible that RF activity is needed to control
the distribution of net moments at the joint (25, 26).
The consistent decrease in both RF EMG amplitude and
stress during pain supports the prediction of pain theories that
the motor adaptations in the presence of pain aim to unload
irritated tissue (10, 16). However, this result requires consideration in relation to the conflicting observation for VM.
Absence of Reduced VM Stress During Homonymous Muscle
Nociceptive Stimulation
In contrast to the observations for RF, our data show that
neither VM stress (experiments I and II) nor VM EMG amplitude (experiment II) reduce during pain when either force or
VL EMG are matched to trials without pain. This result is
consistent with that observed for VL during force-matched
contractions (at 10% of MVC) during pain (24).
In experiment II, we considered the potential to dissociate
drive to VM and VL may be limited by the task objective (to
maintain force). Therefore, we provided a different task goal,
i.e., VL EMG amplitude rather than knee extension force. As
this task does not constrain VM activity we argue that this
provided a goal that can be maintained theoretically without
the need for VM force. However, even in this context there was
no consistent decrease in VM EMG or shear elastic modulus.
This observation implies that either the CNS cannot selectively
modulate the activity of VM and VL, or the cost of this strategy
outweighs the benefit of reduced stress during pain. Experiment III provides evidence that the participants were also
unable to voluntarily reduce the activation of VM while maintaining force, at least with the limited training used here. This
concurs with studies that suggest an inability to selectively
increase activation of VM over VL by alteration of lower limb
position (see review by 22). Taken together, these observations
provide strong evidence of an undifferentiated descending
drive from the CNS to VM and VL in healthy individuals. This
is consistent with the observation of highly synchronized
discharge of motor units in VM and VL, which implies strong
common drive to both muscles (18). Although true for healthy
individuals, synchronization of motor unit discharge between
the obliquus fibers of VM and VL is substantially less in people
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
(rather than knee extension force) and matched during pain,
neither EMG nor stress was systematically altered in VM or RF
during nociceptive stimulation of VM. In contrast, RF stress
and EMG amplitude decreased during nociceptive stimulation
of RF (experiment II). Third, experiment III provides preliminary evidence that participants can voluntary reduce RF activity when maintaining external knee extension force. However,
the same could not be achieved for VM, at least during this
short testing period with limited training (3 ⫻ 20 s contractions). These results highlight between-muscle differences in
adaptation to experimental pain that might be explained by
their function (biarticular vs. monoarticular) and/or the neurophysiological constraints associated to the muscles activation.
Taken together these results provide evidence that pain adaptation may need to be considered with direct reference to
constraints related to the muscle under investigation and the
goal of the experimental task.
•
Between-Muscle Differences in Pain Adaptation
•
Hug F et al.
1139
with chronic patellofemoral pain (19), and greater differentiation of drive may be possible in this group. This is in line with
observations of some independence of drive to VM and VL
as indicated by delayed onset of VM EMG activity relative
to VL during stair ascent in patients with chronic patellofemoral pain (7).
Although there was significant effect of Pain on shear elastic
modulus during PainRF during experiment II, the condition ⫻
muscle interaction was not significant, which suggests that VM
and RF exhibited a similar behavior (decreased stress). However, this interaction was significant for the EMG data; VM
EMG was unchanged and RF EMG amplitude decreased.
Consistent changes in both parameters for RF implies the CNS
decreased activation of this muscle to decrease RF stress
(painful muscle). Changes in VM stress, without concomitant
changes in VM EMG, might be explained by the nonnegligible
intermuscular force transmission (23), such that decreased
force produced by RF during pain induced a lower force
transmission between RF and VM. If so, this observation
provides additional evidence that EMG and shear elastic modulus do not provide interchangeable information and that both
measurements are necessary to fully understand the outcome of
motor adaptation during pain. This is further confirmed by the
moderate coefficient of correlation found between changes in
RF shear elastic modulus and RF EMG activity.
pain, we are confident that fatigue, if any, did not interfere in
the conclusion of this study. Finally, it is important to note that
fatigue does not alter the relationship between shear elastic
modulus and stress (2).
Methodological Considerations
DISCLOSURES
These results demonstrate that between-muscle differences
in neuromechanical adaptation to experimental pain are likely
explained by neurophysiological constraints and/or different
functional roles. Consequently, pain adaptation may require
consideration with reference to constraints related to the muscle system under investigation and the goal of the experimental
task.
ACKNOWLEDGMENTS
We thank M4 Healthcare (Australia) for loan of the ultrasound scanner
(experiment I).
GRANTS
The purchased ultrasound (experiments II and III) was partially supported
by a UQ MEI and National Health and Medical Research Council (NHMRC)
Equipment Grant and the Rebecca L. Cooper Medical Research Foundation.
The study was funded by a Program Grant (ID631717) from the NHMRC of
Australia. P. W. Hodges holds a Senior Principal Research Fellowship
(APP1002190) and K. Tucker a Clinical Research Career Development Fellowship (ID1009410) from the NHMRC.
No conflicts of interest, financial or otherwise, are declared by the author(s).
AUTHOR CONTRIBUTIONS
Author contributions: F.H., P.W.H., and K.J.T. conception and design of
research; F.H., W.v.d.H., and K.J.T. performed experiments; F.H. and
W.v.d.H. analyzed data; F.H., P.W.H., and K.J.T. interpreted results of experiments; F.H. prepared figures; F.H. and K.J.T. drafted manuscript; F.H.,
P.W.H., W.v.d.H., and K.J.T. edited and revised manuscript; F.H., P.W.H.,
W.v.d.H., and K.J.T. approved final version of manuscript.
REFERENCES
1. Bercoff J, Tanter M, Fink M. Supersonic shear imaging: a new technique for soft tissue elasticity mapping. IEEE Trans Ultrason Ferroelectr
Freq Control 51: 396 –409, 2004.
2. Bouillard K, Hug F, Guevel A, Nordez A. Shear elastic modulus can be
used to estimate an index of individual muscle force during a submaximal
isometric fatiguing contraction. J Appl Physiol 113: 1353–1361, 2012.
3. Bouillard K, Jubeau M, Nordez A, Hug F. Effect of vastus lateralis
fatigue on load sharing between quadriceps femoris muscles during
isometric knee extensions. J Neurophysiol 111: 768 –776, 2014.
4. Bouillard K, Nordez A, Hug F. Estimation of individual muscle force
using elastography. PLos One 6: e29261, 2011.
5. Catheline S, Gennisson JL, Delon G, Fink M, Sinkus R, Abouelkaram
S, Culioli J. Measuring of viscoelastic properties of homogeneous soft
solid using transient elastography: an inverse problem approach. J Acoust
Soc Am 116: 3734 –3741, 2004.
6. Cohen JH. Statistical Power Analysis for the Behavioral Sciences. Earlbaum, 1988.
7. Cowan SM, Bennell KL, Hodges PW, Crossley KM, McConnell J.
Delayed onset of electromyographic activity of vastus medialis obliquus
relative to vastus lateralis in subjects with patellofemoral pain syndrome.
Arch Phys Med Rehabil 82: 183–189, 2001.
8. de Rugy A, Loeb GE, Carroll TJ. Muscle coordination is habitual rather
than optimal. J Neurosci 32: 7384 –7391, 2012.
9. Hart HF, Ackland DC, Pandy MG, Crossley KM. Quadriceps volumes
are reduced in people with patellofemoral joint osteoarthritis. Osteoarthritis Cartilage 20: 863–868, 2012.
10. Hodges PW, Tucker K. Moving differently in pain: a new theory to
explain the adaptation to pain. Pain 152: S90 –S98, 2011.
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
This experiment requires consideration of several methodological aspects. First, because EMG does not provide accurate
information regarding muscle stress (4), this experiment used
an elastography technique to test the hypothesis that muscle
stress is altered during pain. As injection of hypertonic saline
does not alter the shear elastic modulus measured at rest or
during contraction (24), the relationship between muscle stress
and shear elastic modulus is unlikely to be altered. Consequently, the changes in shear elastic modulus can be interpreted as changes in muscle stress, even after the injection of
saline. Second, the present study focused on the acute adaptation to pain and it is important to consider that these adaptations are likely to differ from this after a longer exposure to
pain. Previous work provides some evidence for the effectiveness of biofeedback training of muscle coordination, including
the ability to alter the ratio between VM and VL EMG
amplitude (20). Third, adaptation to pain was studied during
isometric tasks performed at ⬃20% of MVC. As redistribution
of muscle activity among the quadriceps muscles depends on
contraction intensity during fatiguing tasks (14), it is relevant
to explore the present observations in tasks that require higher/
lower effort and/or dynamic tasks. Finally, in the absence of
direct measurements, we cannot confirm the absence of fatigue
that could interfere with interpretation of the data. However,
care was taken to limit fatigue. To this end, we studied short
contractions (12-s) at a relatively low intensity (20% of MVC)
with 30 s of recovery between each. The contractions following the pain conditions began 10 min after complete cessation
of pain to provide an extra recovery period for the participants.
The order of the pain conditions was randomized ensuring that
fatigue, if any, did not compromise the between-muscle comparison of pain adaptations. Because the changes observed in
experiment II (EMG, shear elastic modulus) recovered after
Conclusions
1140
Between-Muscle Differences in Pain Adaptation
Hug F et al.
20. Ng GY, Zhang AQ, Li CK. Biofeedback exercise improved the EMG
activity ratio of the medial and lateral vasti muscles in subjects with
patellofemoral pain syndrome. J Electromyogr Kinesiol 18: 128 –133,
2008.
21. Place N, Matkowski B, Martin A, Lepers R. Synergists activation
pattern of the quadriceps muscle differs when performing sustained
isometric contractions with different EMG biofeedback. Exp Brain Res
174: 595–603, 2006.
22. Smith TO, Bowyer D, Dixon J, Stephenson R, Chester R, Donell ST.
Can vastus medialis oblique be preferentially activated? A systematic
review of electromyographic studies. Physiother Theory Pract 25: 69 –98,
2009.
23. Tian M, Herbert RD, Hoang P, Gandevia SC, Bilston LE. Myofascial
force transmission between the human soleus and gastrocnemius muscles
during passive knee motion. J Appl Physiol 113: 517–523, 2012.
24. Tucker K, Hodges PW, Van den Hoorn W, Nordez A, Hug F. Does
stress within a muscle change in response to an acute noxious stimulus?
PLos One 9: e91899, 2014.
25. van Ingen Schenau GJ, Boots PJ, de Groot G, Snackers RJ, van
Woensel WW. The constrained control of force and position in multi-joint
movements. Neuroscience 46: 197–207, 1992.
26. van Ingen Schenau GJ, Dorssers WM, Welter TG, Beelen A, de Groot
G, Jacobs R. The control of mono-articular muscles in multijoint leg
extensions in man. J Physiol 484: 247–254, 1995.
J Appl Physiol • doi:10.1152/japplphysiol.00561.2014 • www.jappl.org
Downloaded from http://jap.physiology.org/ by 10.220.32.246 on June 15, 2017
11. Hug F, Hodges PW, Tucker K. Task dependency of motor adaptations to
an acute noxious stimulation. J Neurophysiol 111: 2298 –2306, 2014.
12. Koo TK, Guo J, Cohen JH, Parker JP. Relationship between shear
elastic modulus and passive muscle force: an ex-vivo study. J Biomech 46:
2053–2059, 2013.
13. Kooistra RD, de Ruiter CJ, de Haan A. Muscle activation and blood
flow do not explain the muscle length-dependent variation in quadriceps
isometric endurance. J Appl Physiol 98: 810 –816, 2005.
14. Kouzaki M, Shinohara M. The frequency of alternate muscle activity is
associated with the attenuation in muscle fatigue. J Appl Physiol 101:
715–720, 2006.
15. Lacourpaille L, Hug F, Bouillard K, Hogrel JY, Nordez A. Supersonic
shear imaging provides a reliable measurement of resting muscle shear
elastic modulus. Physiol Measurement 33: N19 –N28, 2012.
16. Lund JP, Donga R, Widmer CG, Stohler CS. The pain-adaptation
model: a discussion of the relationship between chronic musculoskeletal
pain and motor activity. Can J Physiol Pharmacol 69: 683–694, 1991.
17. Maisetti O, Hug F, Bouillard K, Nordez A. Characterization of passive
elastic properties of the human medial gastrocnemius muscle belly using
supersonic shear imaging. J Biomech 45: 978 –984, 2012.
18. Mellor R, Hodges P. Motor unit synchronization between medial and
lateral vasti muscles. Clin Neurophysiol 116: 1585–1595, 2005.
19. Mellor R, Hodges PW. Motor unit synchronization is reduced in anterior
knee pain. J Pain 6: 550 –558, 2005.
•