J Appl Physiol 96: 15221529, 2004.
First published December 5, 2003; 10.1152/japplphysiol.00687.2003.
HIGHLIGHTED TOPIC Neural Control of Movement
Intramuscular pressure and EMG relate during static contractions
but dissociate with movement and fatigue
Gisela Sjgaard,1 Bente R. Jensen,2 Alan R. Hargens,3 and Karen Sgaard1
1
Department of Physiology, National Institute of Occupational Health, DK-2100 Copenhagen; and
2
Institute of Exercise and Sport Sciences, University of Copenhagen, DK-2100 Copenhagen, Denmark;
and 3Department of Orthopaedics, University of California, San Diego, California 92103
Submitted 3 July 2003; accepted in final form 28 November 2003
Sjgaard, Gisela, Bente R. Jensen, Alan R. Hargens, and Karen
Sgaard. Intramuscular pressure and EMG relate during static con-
tractions but dissociate with movement and fatigue. J Appl Physiol 96:
15221529, 2004. First published December 5, 2003; 10.1152/
japplphysiol.00687.2003.Intramuscular pressure (IMP) and elec-
tromyography (EMG) mirror muscle force in the nonfatigued
muscle during static contractions. The present study explores
whether the constant IMP-EMG relationship with increased force
may be extended to dynamic contractions and to fatigued muscle.
IMP and EMG were recorded from shoulder muscles in three
sessions: 1) brief static arm abductions at angles from 0 to 90,
with and without 1 kg in the hands; 2) dynamic arm abductions at
angular velocities from 9 to 90/s, with and without 1 kg in the
hands; and 3) prolonged static arm abduction at 30 for 30 min
followed by recovery. IMP and EMG increased in parallel with
increasing shoulder torque during brief static tasks. During dy-
namic contractions, peak IMP and EMG increased to values higher
than those during static contractions, and EMG, but not IMP,
increased significantly with speed of abduction. In the nonfatigued
supraspinatus muscle, a linear relationship was found between IMP
and EMG; in contrast, during fatigue and recovery, significant
timewise changes of the IMP-to-EMG ratio occurred. The results
indicate that IMP should be included along with EMG when
mechanical load sharing between muscles is evaluated during
dynamic and fatiguing contractions.
shoulder; static muscle contraction; dynamic muscle contraction;
prolonged muscle contraction
SHOULDER FUNCTION INVOLVES complex interaction among sev-
eral muscles, even in standardized arm positions and during
rather simple movements (24, 25). Electromyography (EMG)
patterns in general, as well as shoulder modeling based on
EMG or optimization principles (36), provide useful informa-
tion on shoulder function in nonfatigued muscles during static
contractions. However, our knowledge is limited during dy-
namic shoulder function and prolonged fatiguing contractions.
Prior investigations suggest that the intramuscular pressure
(IMP) as well as EMG can be used as quantitative indexes for
the contraction force of individual muscle groups during static
contractions of nonfatigued muscles (8, 23, 30, 33). Linear
relationships have been reported between these variables and
muscle force or joint torque, and IMP has been suggested to be
a better index of muscle contraction force than EMG (1). Most
Address for reprint requests and other correspondence: G. Sjgaard, Dept. of
Physiology, National Institute of Occupational Health, Lers Parkalle 105,
DK-2100 Copenhagen , Denmark (E-mail: gs@ami.dk).
1522 8750-7587/04 $5.00 Copyright 2004 the American Physiological Society
recently, a study using an animal muscle model allowing for
direct muscle force measurements simultaneous with IMP
recordings during passive and active isometric force develop-
ment confirmed a positive relationship between these variables
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017
(7); although some caution should be taken when transferring
findings from dissected muscle to the in vivo condition. For
shoulder muscles, these relationships have been studied, in
particular, during a large number of static shoulder positions
(1316). However, to what extent these variables result in
coherent or complementary information during dynamic or
fatiguing contractions of these muscles remains to be eluci-
dated.
During dynamic contractions, some studies are in support of
IMP as well as EMG to serve as indexes of force. For example,
IMP in the human tibialis and soleus muscles was reported to
be dependent only on ankle torque but independent of contrac-
tion mode (static vs. isokinetic) and velocity of contraction
during both concentric and eccentric exercise (2, 30). In con-
trast, in the rabbit tibialis muscle during isotonic contractions,
the highest values of IMP were found at the lowest force levels
corresponding to the highest shortening velocity, and IMP
appeared to be related to the speed of shortening (8). Accord-
ing to physics, a larger force must be developed during dy-
namic contractions because the velocity is increasing during
each contraction, i.e., an acceleration occurs that is larger the
higher the velocity attained; therefore, correspondingly larger
values for EMG and IMP were to be expected. Furthermore,
during dynamic contractions of the human vastus lateralis
muscle, the IMP-to-torque ratio was lower during eccentric
than concentric contractions (6). In the human shoulder mus-
cles, only very slow concentric movement of 1/s has been
studied that may be considered to be a quasi-static contraction
(27). Similarly, for sustained, prolonged fatiguing contractions,
EMG and IMP have only been investigated systematically in a
few studies, and the results are contradictory. It is generally
accepted that prolonged fatiguing contractions at constant sub-
maximal static force level induce a significant increase in the
EMG amplitude, and a close correlation between increasing
EMG and increasing IMP was reported in vastus lateralis
muscle during prolonged contractions at 35% maximum mus-
cle strength (29). Also, these authors found a constant IMP
during a prolonged fatiguing contraction where the EMG
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked advertisement
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
http://www.jap.org
 IMP-TO-EMG RATIO DEPENDS ON CONTRACTION MODE
amplitude was kept constant, despite a decreasing external
force development. However, an increased IMP during pro-
longed constant-force development could not be confirmed by
others (31, 33). Moreover, differential responses were reported
in IMP and EMG during low- vs. high-level static contractions
to fatigue in terms of IMP increasing during 25% MVC, but
not during 70% MVC, despite increases in EMG amplitude
throughout both contraction levels (5). One possible explana-
tion of these different findings is that the physiological mech-
anisms for increases in EMG and IMP have no common route.
Therefore, differences in contraction force level may account
for different responses, such that the increase in IMP during
muscle contraction level is abolished at high (short-term, due
to exhaustion) as well as at very low contraction forces.
Additionally, local physiological and anatomic factors may
affect the relation between EMG amplitude and IMP during
prolonged contractions, and results may, therefore, depend on
the anatomic region and/or the species investigated. Due to the
complex anatomy of the shoulder, and corresponding complex
shoulder function, it is not known if a change in the EMG
amplitude is indicative of a change in muscle force or devel-
opment of muscle fatigue. IMP is influenced by fascial com-
pliance and the actual fluid content of the muscle. Specifically,
a certain degree of muscle edema affects IMP, dependent on
the local anatomy. Thus, in the quadriceps muscle, no measur-
able effect of an increased fluid content of 10% was seen on
the IMP; whereas, in other muscles, e.g., in the lower leg
compartment, swelling may raise IMP due to a low compliance
of the muscle compartment (9, 11, 33). The anatomy of the
supraspinatus muscle region is, to a large degree, comparable
to that of the lower leg compartments (16). Furthermore, for
supraspinatus muscle, a significant thickening of the muscle,
indicating muscle edema, was found in a previous study during
submaximal static contractions (17), but it remains unknown
whether there is an effect of increased fluid content on the IMP
during prolonged contractions of supraspinatus muscle. Inter-
estingly, during repeated maximal isokinetic contractions as
well as submaximal isometric contractions of vastus lateralis
muscle, an increase in IMP during the relaxation periods was
reported and may be causally related to an increase in muscle
fluid content (4, 5). Based on the above, we presumed IMP in
the supraspinatus muscle to increase with time, in line with the
EMG during sustained static contractions. During dynamic as
well as fatiguing shoulder muscle activity, the load sharing
between the different fractions of the shoulder girdle anatomy
may change. A classic subdivision of shoulder muscle function
is into stabilizers and movers (for references, see Ref. 21).
However, such strict subdivision is not always in accordance
with actual in vivo activities (25, 26, 34). Also, via biofeed-
back, subjects can voluntarily change the load sharing (28).
Our knowledge regarding load sharing among shoulder mus-
cles is still limited, and, by measuring EMG in combination
with IMP, possible changes in load sharing may be elucidated
during dynamic as well as fatiguing contractions.
The purpose of the present study was to elucidate whether
the constant IMP-EMG relationship seen with increased shoul-
der abduction torque may be extended to dynamic contractions
and separately to fatigued muscle. More specifically, the hy-
potheses examined for supraspinatus muscle were as follows:
1) during dynamic contractions, both IMP and EMG ampli-
tudes increase with increasing velocity, and 2) both IMP and
J Appl Physiol VOL
1523
EMG increase with time during a constant static contraction
sustained for a prolonged period of time. Additionally, with
inclusion of EMG from two more shoulder muscles, the aim
was to study recovery as well as shoulder muscle load sharing
between rotator cuff muscles (e.g., supraspinatus muscle) and
other shoulder girdle stabilizers (e.g., trapezius muscle de-
scending part) relative to shoulder joint prime movers (e.g.,
deltoideus muscle medial part), during brief static, dynamic,
and sustained static contractions. The underlying hypothesis
regarding load sharing was that stabilizers were relatively more
active during static and prime movers during dynamic contrac-
tions.
METHODS
Six healthy women with a mean (range) age of 31 (2537) yr, body
weight of 66 (6076) kg, and height of 1.75 (1.561.80) m volun-
teered in this study after giving their informed, written consent. None
of the subjects had a medical history of shoulder or neck problems.
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017
The study was approved by the local ethical committee.
Experimental Setup and Protocol
The subjects were sitting in an adjustable experimental chair, with
their back vertical, thighs horizontal, and lower legs vertical. An IMP
transducer-tipped catheter as well as intramuscular wire electrodes for
EMG recording were inserted into the right side supraspinatus muscle
under the guidance of ultrasound imaging B scan. Images were taken
to also measure distances from skin surface as well as muscle
thickness. Furthermore, surface EMG electrodes were mounted above
deltoideus and trapezius muscles, and, for heart rate (HR) and blood
pressure (BP) recordings, a small cuff was placed around the third
finger of the left hand. The subjects then, in the sitting position,
performed arm abductions in the frontal plane with the elbows straight
and palms facing the floor. All contractions were performed bilaterally
to attain symmetric loading of the spine. After maximal voluntary
contractions (MVC), three sessions of contractions were performed:
static arm abductions at different abduction angles, dynamic arm
abductions at different velocities, and, finally, prolonged static arm
abduction. At least 10 min of rest were allowed between the sessions.
MVC. Arm abduction muscle strength was measured at the position
of bilateral 30 arm abduction and taken as the highest value of three
to five attempts. Resting periods of 1- to 2-min duration were allowed
between the MVC contractions.
Brief static arm abductions. Bilateral static arm abductions at 30,
60, and 90 (0 vertical arm and 90 horizontal arm) of abduction
were performed for 10 s each, with and without an external weight of
1 kg in each hand. Additionally, the arms were passively positioned in
the three abduction positions supported by the experimenter. The
contractions and passive positions were performed in random order.
Dynamic arm abductions. Bilateral dynamic arm abductions
through the range from 0 to 90, followed by arm adductions from 90
to 0, were performed at mean angular velocities of 9, 18, 45, and
90/s. The 0 position and 90 position were sustained for a 10-s
period between each of the dynamic contractions. The contractions
were performed with and without a 1-kg weight in the hands. The
dynamic contractions were guided by a metronome and performed in
random order.
Prolonged static arm abduction. Prolonged bilateral 30 arm ab-
duction was maintained for 30 min and followed by a recovery period
of 20 min. During the recovery period, two 30 arm abduction tests (1
min) were performed after 10 and 20 min, respectively. During the
prolonged and the test abductions, the subjects maintained position by
slightly pushing their hands against a position-adjustable force bar.
The external abduction force applied was 0.4 N at the level of the
hands (corresponding to 0.3% MVC). During the prolonged con-
96 APRIL 2004 www.jap.org
1524 IMP-TO-EMG RATIO DEPENDS ON CONTRACTION MODE
traction and subsequent recovery period, HR, BP, and rating of
perceived exertion (RPE) were recorded.
Force and Torque Recordings
MVC was measured by using a strain-gauge-based, three-dimen-
sional force transducer (AMTI-MC3-6-1000) mounted with a handle
that the hand gripped onto. A bilateral setup was used, but only force
exerted by the right hand was measured. Force data were sampled at
128 Hz, and root-mean-square (RMS) values were calculated for
0.26-s time periods. The highest 1-s value (average of 4 time periods)
was considered to represent MVC. During the prolonged abduction
session, left and right abduction force were measured with strain-
gauge-based transducers built into two force bars that were placed to
touch the dorsal side of the hands, when the subjects had attained the
correct 30 abduction position. The orientation of left and right
abduction forces was in the frontal plane and in the direction perpen-
dicular to the arm. To help the subjects maintain the position, a small,
constant force output of 0.4 N was requested, and visual feedback was
given in front of the subjects, from both left and right transducers. To
convert hand force into shoulder torque, the distance from the center
of caput humeri to the handle (midhand) and to the force bar position
on the hand where the external load was applied, respectively, was
measured. The relative load in the abducted positions was calcu-
lated based on simple moment calculation using force and moment
arm (20).
Intramuscular Recording Sites
Three local anatomic dimensions in the shoulder region were
measured at a lateral and medial aspect by using ultrasound imaging
(Diasonics, VST Master Series, GE-Medical, Copenhagen, Denmark)
to guide the insertions of wire for EMG and Millar transducer-tipped
catheter for IMP recordings (for details see Ref. 17). These recordings
were limited to the supraspinatus muscle due to equipment restriction.
The distance from skin surface to the supraspinatus muscle, thickness
of the supraspinatus muscle, and thickness of the trapezius muscle
were measured. For example, in one subject, a negative pressure was
measured initially. Ultrasound imaging of the shoulder showed that
the catheter was located next to the bottom of fossa supraspinata. The
catheter was then moved to the central part of the muscle to obtain the
actual muscle pressure. Finally, the distance from skin surface to the
tip of the EMG wire and the IMP catheter tip was measured at the end
of the experiment when these were withdrawn from the muscle.
IMP
Local anesthesia (Zylocain, 20 mg/ml) was given to the skin and
subcutis above the right supraspinatus muscle at the site for IMP
recording. A transducer-tipped pressure catheter (Millar Micro-Tip
Houston, TX) was then inserted in parallel to the fiber orientation into
the central part of supraspinatus muscle through a Teflon catheter
(Venflon, 14G/2 mm OD). After insertion, the Teflon catheter was
withdrawn, leaving the IMP catheter in the muscle. The IMP trans-
ducer-tipped catheter was zero adjusted (in isotonic saline) and cali-
brated with an electrical input and with a water column (isotonic
saline) in each experiment. No systematic difference was found
between the two methods. IMP was sampled at 1,024 Hz, and RMS
values for 0.26-s time periods were calculated. Maximum value of
IMP was calculated during MVC as the highest 1-s value (mean of 4
time periods). During the brief static contractions, the mean IMP
values were calculated over 6 s (first and last 2 s were omitted of the
10-s contractions), and peak values (0.1 s) of IMP were calculated
during the dynamic concentric contractions, together with IMP values
over 6 s (as for the brief static contractions) during the 0 and 90
position sustained immediately before and after the dynamic concen-
tric contractions. During the prolonged abduction, 1-min average
values were calculated every minute, and in the recovery period 1-min
values were calculated during the two test periods.
J Appl Physiol VOL 96 APRIL 2004
EMG
Intramuscular EMG was recorded from bipolar Teflon-coated steel
wire electrodes with a diameter of 0.05 mm inserted into the central
part of the right supraspinatus muscle through a needle. Surface EMG
from the middle part of the right deltoideus muscle and the descending
part of the right trapezius muscle were recorded by using bipolar
surface electrodes (Medicotest, N-10-A) placed on the skin above the
muscles, with a distance of 2 cm between the recording areas. Possible
cross talk was demonstrated to be negligible by having the subject
perform a few practical tests while recording the EMG channels. A
commercially available EMG recording system was used (IC-600,
Medinik, Orbyhus, Sweden). Skin surface was carefully cleaned with
alcohol and shaved to decrease skin resistance. The quality of the raw
EMG signals was controlled continuously on oscilloscopes and sam-
pled online at 1,024 Hz. The same calculation periods were applied
for the three channels of EMG as for the IMP, and all EMG data were
presented as RMS amplitude in percent maximum EMG (EMGmax).
HR, BP, and RPE
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017
Fingertip HR (in beats/min) and BP (in terms of mean arterial BP,
mmHg) were measured continuously (Finapres, Ohmeda 2300) and
displayed every minute during and after the prolonged bilateral static
abduction. The values presented in the results are as measured at the
fingertip, and, for correction to heart level, a mean of 34 mmHg is to
be subtracted. RPE was measured according to a 10-graded Borg scale
(3) every 2 min during the prolonged abduction as well as in the
recovery period during the brief abductions.
Statistics
All data are presented as means SE, unless otherwise indicated.
Repeated-measures ANOVA was performed with the dependent vari-
ables, IMP and the three EMG recordings, and the independent
variables, position and load during the brief static contractions and
speed and load during the dynamic contractions. In case of signifi-
cance, multiple comparison was performed with the Tukeys test. If
significant interaction was found, ANOVA was performed for each of
the dependent variables separately (SigmaStat). Changes over time
were analyzed by the Friedman test. P 0.05 was considered
statistically significant.
RESULTS
Muscle Strength, Muscle Thickness, and Recording Sites
for IMP and EMG
The maximal abduction force measured at the hands was
54 7.5 N, and the moment arm was 0.74 0.02 m, resulting
in maximal abduction torque in the shoulder joint of 38.9 3.0
Nm, including the torque requested to counteract the weight of
the arms in the 30 abducted position. During this, an IMP
value of 324 38 mmHg was attained in supraspinatus
muscle, and the EMGmax values were 1,164 249 V for
supraspinatus muscle, 628 91 V for deltoideus muscle, and
947 103 V for trapezius muscle. The 30 abducted arm
position corresponded to 13.8 1.1% MVC with no hand load
and 14.1 1.1% MVC with 0.4 N (reaction force during the
prolonged contraction). The thickness of trapezius muscle was
6.3 0.3 and 6.1 0.3 mm for the lateral and medial aspect,
and the corresponding values for supraspinatus muscle were
20.9 1.7 and 11.2 1.6 mm, respectively. The distance from
the skin surface to the superficial fascia of supraspinatus
muscle was 14.4 0.9 and 13.8 1.2 mm at the lateral and
medial aspect, respectively. The corresponding distance to the
tip of the pressure transducer was 40 2 mm, and the insertion
www.jap.org
 IMP-TO-EMG RATIO DEPENDS ON CONTRACTION MODE 1525
Table 1. IMP and EMG values of shoulder muscles during brief static arm abductions
Abduction

IMP and EMG Muscle Load 0 30 60 90

IMP, mmHg passive 20.93.8 16.13.6 29.17.3 32.67.3

Supraspinatus no load 13.43.3 55.76.8 87.212.7 80.69.1
1-kg load 13.33.1 72.113.7 117.48.8 120.021.4
EMG, %EMGmax passive 1.10.3 1.80.6 2.30.7 3.81.7
Supraspinatus no load 1.30.6 11.82.7 14.92.6 15.33.6
1-kg load 2.20.7 15.82.4 20.82.8 22.04.5
EMG, %EMGmax passive 2.51.1 4.32.3 5.82.8 7.32.9
Deltoideus no load 2.51.5 8.82.0 14.72.2 21.42.3
1-kg load 2.91.4 13.71.7 22.32.3 28.72.1
EMG, %EMGmax passive 5.21.5 6.72.2 6.72.5 7.62.5
Trapezius no load 3.71.3 10.32.4 17.44.2 23.25.3
1-kg load 6.11.7 20.06.0 29.66.7 39.47.9
Values are means SE. IMP, intramuscular pressure; EMG, electromyography; EMGmax, maximum EMG.

angle was 60 relative to skin surface, resulting in a vertical natus muscle decreased with shortening velocity and was signif-
depth of 33 mm. Accounting for the location of the sensor

Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017

icantly lower at maximum compared with minimum velocity at no
placed on the side of the catheter 5 mm above the tip resulted hand load. Also, the peak IMP/EMG was lower during the
in a pressure recording depth of 28 mm. The distance from the dynamic than the static contractions.
skin surface to the tip of the wire EMG electrode was 26 1
mm and thus very close to that of the IMP recordings, with
both being well below the superficial fascia of supraspinatus
muscle.
Brief Static Arm Abductions
IMP in supraspinatus muscle and EMG in supraspinatus,
deltoideus, and trapezius muscles increased significantly with
increasing angle of abduction from 0 to 90 and with increas-
ing external shoulder torque from passive to 0 kg and 1-kg load
in the hands (Table 1). IMP tended to increase during passive
arm abduction, but this was not significant. The highest IMP
value of 120 mmHg was attained in the 90 abducted arm
position with 1-kg load in the hand, which corresponded to
33% of the IMP during MVC. Correspondingly, supraspinatus
muscle EMG amounted to 22% EMGmax in the 90 abducted
arm position with 1 kg.
Dynamic Arm Abductions
The peak IMP and EMG values from the supraspinatus
muscle during the concentric dynamic contraction phases, with
and without 1-kg hand load, are presented in Fig. 1, together
with the corresponding data during the 90 arm abduction
sustained immediately after. The values at 0 arm abduction
preceding the dynamic contractions (not shown in Fig. 1) and
at 90 arm abduction after the dynamic task were not signifi-
cantly different from the corresponding values reported above
during the brief static contractions. However, during the dy-
namic concentric contractions, IMP and EMG peak values
were significantly higher than the values at the subsequent
position at 90 arm abduction, and the values with 1-kg load were
higher than those without hand load. Interestingly, EMG peak
values for all muscles increased significantly with increased short-
ening velocity, but this was not the case for the IMP peak values.
No significant differences were found in the EMG ratio between
muscles (i.e supraspinatus/deltoideus, supraspinatus/trapezius, Fig. 1. Intramuscular pressure (IMP) and electromyography (EMG) ampli-
tudes at different load and arm angular velocities (from 0 to 90 abduction,
and deltoideus/trapezius) when dynamic contractions at different A 9/s, B 18/s, C 45/s, and D 90/s) during concentric contractions
velocities or dynamic vs. static contraction were compared. In (open bars) and subsequent static contractions sustained in the 90 abduction
contrast, the peak IMP-to-EMG ratio (IMP/EMG) for supraspi- position (hatched bars). Values are means SE. EMGmax, maximum EMG.

J Appl Physiol VOL 96 APRIL 2004 www.jap.org

1526 IMP-TO-EMG RATIO DEPENDS ON CONTRACTION MODE

Prolonged Static Arm Abduction

The average value of IMP initially in the prolonged abduc-
tion (after 1 min) amounted to 53 8 mmHg and thus was
similar to the 56 7 mmHg in the above-reported brief static
30 arm abduction session (Table 1). During the prolonged
contraction, all subjects maintained the abducted position with
only minor adjustments, according to the experimenter and the
force feedback from the hands. Despite this, IMP decreased in
four subjects but remained at the same level in two subjects,
with the overall mean value being only 30 9 mmHg at 29
min of contraction. The mean values over time of these IMP
data are depicted together with the corresponding data for the
three EMG recordings that showed significant increases for all
muscles during the prolonged abduction from initial values
(Fig. 2). These initial EMG values corresponded to those
reported above during the brief static 30 arm abduction
session (Table 1). From the first to the last minute of abduction,
the EMG values, on average, increased to 200, 220, and 217%

Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017

of the initial level for supraspinatus, deltoideus, and trapezius
muscles, respectively.
The resting level of IMP measured 1 and 2 min after the
prolonged abduction was similar to the resting level measured
before the prolonged abduction that corresponded to those in
the brief static as well as dynamic contractions. The levels of
the IMP during the test contractions after 10 min, 53 12
mmHg, and 20 min, 53 11 mmHg, of recovery were similar
to the initial (1 min) level during the prolonged abduction. For
all muscles, the EMG values decreased to resting level imme-
diately after the abduction. Of particular note, during the test
abductions after 10 and 20 min, EMG values for supraspinatus
and deltoideus muscles still remained significantly above the
initial abduction level, whereas that for trapezius muscle had
recovered within 10 min.
In the nonfatigued muscle, a linear relationship was seen
between IMP and the relative EMG of supraspinatus muscle
(%EMGmax) in the 30 abducted position, with and without
1-kg hand load. However, a clear progressive deviation be-
tween the IMP and EMG of supraspinatus muscle relative to
the nonfatigued situation was found during the 30-min abduc-
tion period (Fig. 3). This was due to the steady EMG increase
from the initial 10% EMGmax toward around twice this level
(Fig. 2).
Resting value of fingertip BP was 124.5 13.1 mmHg, and
this increased significantly during the prolonged contraction,
from 6 mmHg above resting value at 1 min to 32 mmHg above
the resting value at 29 min. Furthermore, HR increased from a
resting value of 70 6.3 to 74 5.5 beats/min after 1 min and
to 84 6.1 beats/min at 29 min. RPE during the prolonged
abduction increased from an initial value (1 min) of 0.5 0.3
to 7.8 0.5 at the end of the contraction (29 min). During the
test contractions after 10 min of recovery, RPE was 2.4 0.9,
i.e., elevated relative to initial (1 min) value, and remained at
this level also after 20 min of recovery, i.e., 2.8 0.9.
DISCUSSION

The main finding of the present study was the dissociation

between IMP and EMG amplitude with dynamic contractions Fig. 2. IMP and EMG during the prolonged abduction (Abd) and the subse-
and as fatigue develops, in contrast to the well-known linear quent recovery period. Values are means SE.
relationship between IMP and EMG amplitude during acute
static contractions with increased muscle force.
J Appl Physiol VOL 96 APRIL 2004 www.jap.org
 IMP-TO-EMG RATIO DEPENDS ON CONTRACTION MODE
Association Between IMP and EMG Amplitude During Static
But Not Dynamic Contractions
The present study confirmed prior investigations on the
effect of abduction angle and significance of external load on
the IMP in supraspinatus muscles during brief static arm
abductions (1315, 27). However, as in earlier studies on
female subjects (16, 20), the absolute level of the IMP during
submaximal as well as maximal contractions was lower than
the level obtained in the above studies on male subjects. The
somewhat high resting values were likely to be due to our
setup, where some soft elastic bandage was used to compress
some of the equipment to the skin.
The results from the dynamic contractions showed an in-
creasing peak value of the IMP that was higher with increased
load in the hand, but no significant increase was revealed with
increased angular velocity. This was puzzling but could be due
to large shape changes of the supraspinatus muscle during a
90 arm abduction. At the highest angular velocities, the
instantaneous distribution of IMP may then be inhomogeneous
with the time history of the IMP being dependent on the spatial
location of the measuring site. In particular, IMP is a less
accurate index of muscle force in the shortened muscle position
(7), which is attained earlier the faster the contraction is. Thus
a higher peak IMP in supraspinatus muscle during the fastest
contraction may have been offset due to this occurring at a
shorter muscle length than at a slower abduction velocity. This
may account for the somewhat lower peak IMP at the highest
contraction velocity (Fig. 1) and, thereby, a lack of significance
between peak IMP and contraction velocity. From a mechan-
ical point of view, an increase in muscle force was expected
with increased movement velocity due to the need for extra
force to increase the acceleration of the arm. Such correspond-
ing increase was seen regarding the EMG for all three muscles,
and it remains unclear whether EMG or IMP reflects muscle
force best during dynamic contractions. By monitoring ankle
joint torque and position with an isokinetic dynamometer
during maximal effort plantar-flexion and dorsiflexion of the
foot, IMP and EMG activity in the tibialis anterior and soleus
muscles was studied (35). In this study, a dissociation was seen
between IMP and ankle joint torque during eccentric muscular
J Appl Physiol VOL
1527
Fig. 3. Relationship between IMP and EMG of supraspinatus
muscle during brief static contraction [diagonal line drawn
through mean (SE) values at 0, 30, and 60 abduction;
means SE are presented in Table 1] and during the fatiguing
contraction (open circles and dotted line, mean values) and
recovery (solid circles, mean values). Data for the circles
correspond to the mean values as given in Fig. 2.
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017
activity compared with static muscular activity, because IMP
increased to a lesser extent during eccentric than during iso-
metric contractions. In other words, the IMP vs. torque rela-
tionship is linear, but the slope is less for eccentric compared
with isometric and concentric contractions. However, EMG vs.
torque is generally lower during eccentric contractions com-
pared with isometric contractions, indicating that the IMP-
EMG relationship may be similar during various contraction
modes. In another study of leg muscles, IMP of soleus and
tibialis anterior was analyzed during gait (2). In this investi-
gation as in ours, IMP was continuously recorded by using
Millar pressure catheters. A linear relationship existed between
soleus IMP and ankle joint torque as monitored by a dyna-
mometer during treadmill walking and running. Thus phasic
elevations of IMP during exercise are probably generated due
to muscle force development. Tibialis anterior IMP showed a
biphasic response, with the largest peak (90 mmHg during
walking and 151 mmHg during running) occurring shortly after
heel strike, in concordance with the function of this muscle
during gait. IMP magnitude increased with gait speed in both
muscles, thus further underlining its relationship to force.
Finally, at these submaximal contractions, the force vs. torque
relationship was not different between contraction modes.
Dissociation of IMP and EMG Amplitude During Prolonged
Static Contraction
As outlined in the Introduction, there were several reasons to
expect IMP to increase in supraspinatus muscle during pro-
longed contractions. A previous study of prolonged shoulder
abductions on muscle fluid balance indicated development of
muscle edema in supraspinatus muscle that is expected to
increase resting IMP (17). Individual differences may occur,
depending on the individual muscle compartment compliance
relative to the degree of muscle edema, but, in any case, some
increase in IMP is expected if the same force is sustained. This
is in contrast to the decrease in IMP seen in four of the
subjects. It may be considered then that the different patterns
may not reflect individual differences, but rather differences
due to slightly different locations of the IMP recording site
among subjects. Interestingly, in an earlier study, no such
96 APRIL 2004 www.jap.org
1528 IMP-TO-EMG RATIO DEPENDS ON CONTRACTION MODE
individual or local decreases in IMP were seen in the vastus
lateralis muscle, although this muscle is considered to be
located in a more compliant surrounding than supraspinatus
muscle, and, despite an even lower contraction level of only
5%, MVC of the knee extensors was sustained (33). Further-
more, we have measured IMP throughout a wide range of
recording depths, indicating that IMP in supraspinatus muscle
was rather homogeneous (16). One reason for the present
finding may then be changed load sharing between abductor
muscles, as discussed below.
The obvious physiological advantage of a stable or even
decreased IMP over time is the preservation of muscle blood
flow. According to a simple model, blood flow is proportional
to the pressure head that is considered to correspond to the
difference between arterial BP and IMP (32). Increases in IMP
increase BP (10), but, in the present study, BP did increase by
25 mmHg, despite unchanged, or even decreased, IMP. This
response, in combination with a probable decrease of vascular
resistance (metabolically induced), likely gradually increased
blood flow relative to the initial level during the abduction.
Such increased blood flow may then postpone or prevent local
fatigue development. Thus, in the present study, IMP may not
have been a critical factor of significance for muscle blood
flow, although more advanced models on the interaction be-
tween IMP and perfusion, including spatial heterogeneity of
IMP, suggest that venous pressure may exceed IMP obstruct-
ing local perfusion (37). Furthermore, despite maintained
blood flow, IMP might affect performance during prolonged
contraction of supraspinatus muscle of 10% MVC, because
IMP exceeded 30 mmHg, a minimum pressure that has been
shown to elicit necrosis if maintained for 8 h (12).
Load Sharing Between Shoulder Muscles During Abduction
The EMG of nonfatigued muscles is related to force devel-
opment. However, in fatigued muscles, the EMG amplitude
may increase without a corresponding increase in force output,
even at low-contraction levels corresponding to 30 shoulder
abduction (10% MVC) (22, 29). The lack of increased IMP
in general and the occurrence of decreased IMP for some
subjects suggest a decreased supraspinatus muscle force during
the prolonged abduction (18). Measurement of the IMP during
the brief contractions showed that the initial level of the IMP
was reestablished within 10 min of recovery after the pro-
longed contraction. A recent study on motor control at motor
unit level showed that muscles compensate for the develop-
ment of fatigue by motor unit recruitment and frequency
modulation (19). In particular, trapezius muscle showed only
an increase in amplitude, but not a decrease in mean power
frequency, as well as a fast recovery of the amplitude in
concert with the present study. Thus trapezius muscle may
have been less fatigued than the other shoulder muscles, and its
increased amplitude has reflected an increased force develop-
ment rather than fatigue development. This may indicate an
increased stabilizing function of the trapezius muscle as fatigue
developed in other muscles and, thereby, a change in the load
sharing between shoulder muscles. Taken together, our present
results suggest that the relative force distribution among shoul-
der muscles during the prolonged abduction probably changed
over time. This means that the force distribution principles by
which the shoulder function is optimized may change over
J Appl Physiol VOL
time. This concept should be considered in future improve-
ments of shoulder function models.
In conclusion, IMP and EMG amplitudes increased in pro-
portion in supraspinatus muscle with acute increased shoulder
joint abduction torque, but during dynamic contraction a dis-
sociation of this relationship occurred because only the EMG
increase was velocity dependent. Furthermore, during an
10% MVC static contraction sustained for 30 min, EMG
increased to about twice its amplitude, whereas IMP remained
constant or even decreased in some subjects. With recovery, a
change occurred in the IMP-EMG relationship during the test
contractions due to the slow recovery of the EMG amplitude.
Finally, EMG amplitudes indicated the load sharing between
different functional shoulder muscles not to be different during
brief static and dynamic contractions, but, in combination with
IMP measures, a change in load sharing among muscles was
indicated during sustained fatiguing static contraction.
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017
REFERENCES
1. Aratow M, Ballard RE, Crenshaw AG, Styf J, Watenpaugh DE,
Kahan NJ, and Hargens AR. Intramuscular pressure and electromyogra-
phy as indexes of force during isokinetic exercise. J Appl Physiol 74:
26342640, 1993.
2. Ballard RE, Watenpaugh DE, Breit GA, Murthy G, Holley DC, and
Hargens AR. Leg intramuscular pressures during locomotion in humans.
J Appl Physiol 84: 19761981, 1998.
3. Borg G. Perceived exertion as an indicator of somatic stress. Scand J
Rehabil Med 2: 9298, 1970.
4. Crenshaw AG, Gerdle B, Heiden M, Karlsson S, and Friden J.
Intramuscular pressure and electromyographic responses of the vastus
lateralis during repeated maximal isokinetic knee extensions. Acta Physiol
Scand 170: 119126, 2000.
5. Crenshaw AG, Karlsson S, Gerdle B, and Friden J. Differential
responses in intramuscular pressure and EMG fatigue indicators during
low vs. high level isometric contractions to fatigue. Acta Physiol Scand
160: 353361, 1997.
6. Crenshaw AG, Karlsson S, Styf J, Backlund T, and Friden J. Knee
extension torque and intramuscular pressure of the vastus lateralis muscle
during eccentric and concentric activities. Eur J Appl Physiol 70: 1319,
1995.
7. Davis J, Kaufman KR, and Lieber RL. Correlation between active and
passive isometric force and intramuscular pressure in the isolated rabbit
tibialis anterior muscle. J Biomech 36: 505512, 2003.
8. Degens H, Salmons S, and Jarvis JC. Intramuscular pressure, force and
blood flow in rabbit tibialis anterior muscles during single and repetitive
contractions. Eur J Appl Physiol 78: 1319, 1998.
9. Friden J, Sfakianos PN, and Hargens AR. Muscle soreness and intra-
muscular fluid pressure: comparison between eccentric and concentric
load. J Appl Physiol 61: 21752179, 1986.
10. Gallagher KM, Fadel PJ, Smith SA, Norton KH, Querry RG, Oliven-
cia-Yurvati A, and Raven PB. Increases in intramuscular pressure raise
arterial blood pressure during dynamic exercise. J Appl Physiol 91:
23512358, 2001.
11. Hargens AR, Mubarak SJ, Owen CA, Garetto LP, and Akeson WH.
Interstitial fluid pressure in muscle and compartment syndromes in man.
Microvasc Res 14: 110, 1977.
12. Hargens AR, Schmidt DA, Evans KL, Gonsalves MR, Cologne JB,
Garfin SR, Mubarak SJ, Hagan PL, and Akeson WH. Quantitation of
skeletal-muscle necrosis in a model compartment syndrome. Bone Joint
Surg (Am) 63A: 631636, 1981.
13. Jarvholm U, Palmerud G, Herberts P, Hogfors C, and Kadefors R.
Intramuscular pressure and electromyography in the supraspinatus muscle
at shoulder abduction. Clin Orthop 245: 102109, 1989.
14. Jarvholm U, Palmerud G, Karlsson D, Herberts P, and Kadefors R.
Intramuscular pressure and electromyography in four shoulder muscles.
J Orthop Res 9: 609619, 1991.
15. Jarvholm U, Palmerud G, Styf J, Herberts P, and Kadefors R.
Intramuscular pressure in the supraspinatus muscle. J Orthop Res 6:
230238, 1988.
96 APRIL 2004 www.jap.org
 IMP-TO-EMG RATIO DEPENDS ON CONTRACTION MODE
16. Jensen BR, Jrgensen K, Huijing PA, and Sjgaard G. Soft tissue
architecture and intramuscular pressure in the shoulder region. Eur J
Morphol 33: 205220, 1995.
17. Jensen BR, Jrgensen K, and Sjgaard G. The effect of prolonged
isometric contractions on muscle fluid balance. Eur J Appl Physiol 69:
439444, 1994.
18. Jensen BR, Laursen B, and Sjgaard G. Aspects of shoulder function in
relation to exposure demands and fatiguea mini review. Clin Biomech
(Bristol, Avon) 15, Suppl 1: S17S20, 2000.
19. Jensen BR, Pilegaard M, and Sjgaard G. Motor unit recruitment and
rate coding in response to fatiguing shoulder abductions and subsequent
recovery. Eur J Appl Physiol 83: 190199, 2000.
20. Jensen BR, Sjgaard G, Bornmyr S, Arborelius M, and Jrgensen K.
Intramuscular laser-Doppler flowmetry in the supraspinatus muscle during
isometric contractions. Eur J Appl Physiol 71: 373378, 1995.
21. Jensen C, Laursen B, and Sjgaard G. Shoulder and neck. In: Biome-
chanics in Ergonomics, edited by Kumar S. London: Taylor & Francis,
1999, p. 201220.
22. Jrgensen K, Fallentin N, Krogh-Lund C, and Jensen BR. Electro-
myography and fatigue during prolonged, low-level static contractions.
Eur J Appl Physiol 57: 316321, 1988.
23. Korner L, Parker P, Almstrom C, Anderson GBJ, Herberts P, Kade-
fors R, Palmerud G, and Zetterberg C. Relation of intramuscular
pressure to the force output and myoelectric signal of skeletal muscle.
J Orthop Res 2: 289296, 1984.
24. Laursen B, Jensen BR, Nemeth G, and Sjgaard G. A model predicting
individual shoulder muscle forces based on relationship between EMG and
3D external forces in static position. J Biomech 31: 731739, 1998.
25. Laursen B, Jensen BR, and Sjgaard G. Effect of speed and precision
demands on human shoulder muscle electromyography during a repetitive
task. Eur J Appl Physiol 78: 544548, 1998.
26. Laursen B, Sgaard K, and Sjgaard G. Biomechanical model predict-
ing electromyographic activity in three shoulder muscles from 3D kine-
J Appl Physiol VOL 96 APRIL 2004
1529
matics and external forces during cleaning work. Clin Biomech (Bristol,
Avon) 18: 287295, 2003.
27. Palmerud G, Forsman M, Sporrong H, Herberts P, and Kadefors R.
Intramuscular pressure of the infra- and supraspinatus muscles in relation
to hand load and arm posture. Eur J Appl Physiol 83: 223230, 2000.
28. Palmerud G, Kadefors R, Sporrong H, Jarvholm U, Herberts P,
Hogfors C, and Peterson B. Voluntary redistribution of muscle activity
in human shoulder muscles. Ergonomics 38: 806815, 1995.
29. Sadomoto T, Bonde-Petersen F, and Suzuki Y. Skeletal muscle tension,
flow, pressure, and EMG during sustained isometric contractions in
humans. Eur J Appl Physiol 51: 395408, 1983.
30. Sejersted OM and Hargens AR. Intramuscular pressures for monitoring
different tasks and muscle conditions. In: Motor Control VII, edited by
Stuart DG. Tucson, AZ: Motor Control, 1995.
31. Sejersted OM, Hargens AR, Kardel KR, Blom P, Jensen , and
Hermansen L. Intramuscular fluid pressure during isometric contraction
of human skeletal muscle. J Appl Physiol 56: 287295, 1984.
32. Sjgaard G. Muscle fatigue. Med Sports Sci 26: 98 109, 1987.
33. Sjgaard G, Kiens B, Jrgensen K, and Saltin B. Intramuscular pres-
sure, EMG and blood flow during low-level prolonged static contraction in
man. Acta Physiol Scand 128: 475484, 1986.
34. Sgaard K, Laursen B, Jensen BR, and Sjgaard G. Dynamic loads on
the upper extremities during two different floor cleaning methods. Clin
Downloaded from http://jap.physiology.org/ by 10.220.33.5 on March 14, 2017
Biomech (Bristol, Avon) 16: 866879, 2001.
35. Styf J, Ballard R, Aratow M, Crenshaw A, Watenpaugh D, and
Hargens AR. Intramuscular pressure and torque during isometric, con-
centric and eccentric muscular activity. Scand J Med Sci Sports 5:
291296, 1995.
36. Van der Helm FCT. A finite element musculoskeletal model of the
shoulder mechanism. J Biomech 27: 551569, 1994.
37. Van Donkelaar CC, Huyghe JM, Vankan WJ, and Drost MR. Spatial
interaction between tissue pressure and skeletal muscle perfusion during
contraction. J Biomech 34: 631637, 2001.
www.jap.org