ORIGINAL COMMUNICATION

A Comprehensive Analysis of Muscle

Recruitment Patterns During Shoulder Flexion:
An Electromyographic Study
DUANGJAI WATTANAPRAKORNKUL,
1
MARK HALAKI,
2
CRAIG BOETTCHER,
1
IAN CATHERS,
2
AND KAREN A. GINN
1
*
1
Discipline of Biomedical Science, School of Medical Sciences, Sydney Medical School,
The University of Sydney, Sydney, Australia
2
Discipline of Exercise and Sport Science, Faculty of Health Sciences,
The University of Sydney, Sydney, Australia
Although exion is a common component of the routine clinical assessment of
the shoulder the muscle recruitment patterns during this movement are not
clearly understood making valid interpretation of potential muscle dysfunction
problematic. The purpose of this study was to comprehensively examine
shoulder muscle activity during exion in order to compare the activity levels
and recruitment patterns of shoulder exor, scapular lateral rotator and rota-
tor cuff muscles. Electromyographic (EMG) data were recorded from 12
shoulder muscles sites in 15 volunteers. Flexion was performed in standing in
the sagittal plane at no load, 20%, and 60% of each subjects maximum load.
EMG data were normalized to maximum values obtained during maximum vol-
untary contractions. Results indicated that anterior deltoid, pectoralis major,
supraspinatus, infraspinatus, serratus anterior, upper, and lower trapezius
were activated at similar moderate levels. However, subscapularis was acti-
vated at low levels and signicantly lower than supraspinatus and infraspina-
tus. Similar activity patterns across time were demonstrated in the muscles
that produce exion torque, laterally rotate the scapula, as well as supraspina-
tus and infraspinatus, and did not change as exion load increased. The onset
of activity in supraspinatus and anterior deltoid occurred at the same time and
prior to movement of the limb at all loads with infraspinatus activity also
occurring prior to movement onset at the medium and high load conditions
only. Posterior rotator cuff muscles appear to be counterbalancing anterior
translational forces produced during exion and it would appear that supraspi-
natus is one of the muscles that consistently initiates exion. Clin. Anat.
24:619626, 2011. V VC
2011 Wiley-Liss, Inc.
Key words: shoulder; electromyography; exion; rotator cuff; motor control
INTRODUCTION
Flexion is a common functional movement at the
shoulder, which is routinely assessed as part of the
normal clinical assessment of patients with shoulder
dysfunction. However, the normal recruitment pat-
terns of shoulder muscles activated during exion is
not clearly understood. Therefore, a valid interpreta-
tion of potential shoulder muscle dysfunction associ-
ated with difculty (pain, restricted movement, etc.)
during shoulder exion is problematic.
*Correspondence to: Karen Ginn, Discipline of Biomedical
Science, School of Medical Sciences, Sydney Medical School, The
University of Sydney, Cumberland Campus C42, East St, Lid-
combe, NSW1825, Australia. E-mail: karen.ginn@sydney.edu.au
Contract grant sponsor: 2008 Endeavour International Post-
graduate Research Scholarship (EIPRS).
Received 21 September 2010; Revised 24 November 2010;
Accepted 29 November 2010
Published online 12 January 2011 in Wiley Online Library
(wileyonlinelibrary.com). DOI 10.1002/ca.21123
V VC
2011 Wiley-Liss, Inc.
Clinical Anatomy 24:619626 (2011)
Although co-ordinated movement of the humerus
and scapula and control of the humeral head on the
glenoid fossa is required to achieve full exion range
of motion, only four previous electromyographic (EMG)
studies have simultaneously investigated activity in
muscles from each of the axiohumeral, axioscapular,
and rotator cuff muscle groups during exion (Inman
et al., 1944; Ito, 1980; McCann et al., 1993; Myers
et al., 2005). In addition, comparisons with previous
EMG studies investigating shoulder exion are difcult
because of differences in the manner in which the EMG
data were reported, e.g., raw data (Inman et al.,
1944; Ito, 1980) versus normalized data (Kronberg
et al., 1990; Townsend et al., 1991; Moseley et al.,
1992; McCann et al., 1993; Ekstrom et al., 2003;
Myers et al., 2005), unclear normalization procedures
(Kronberg et al., 1990; Townsend et al., 1991; Mose-
ley et al., 1992; McCann et al., 1993), the use of
unknown relative load levels (Kronberg et al., 1990;
Townsend et al., 1991; McCann et al., 1993) and the
lack of statistical analysis to determine signicantly dif-
ferent levels of muscle activity (Kronberg et al., 1990;
Townsend et al., 1991; Moseley et al., 1992; McCann
et al., 1993; Ekstrom et al., 2003; Myers et al., 2005).
High anterior deltoid activity during shoulder exion
is consistently reported (Kronberg et al., 1990; Town-
send et al., 1991; Myers et al., 2005). However, de-
spite pectoralis major (clavicular portion) being listed
as a major shoulder exor in anatomy textbooks (Pal-
astanga et al., 2006; Susan, 2008), one (Kronberg
et al., 1990) of the two (Kronberg et al., 1990; Town-
send et al., 1991) EMG studies that investigated pec-
toralis major activity during a shoulder exion task,
by recording over the sternoclavicular portion,
reported low activity levels in this muscle which were
approximately equal to the levels recorded during
shoulder extension tasks. Greater inconsistency exists
in the relative activity levels of the rotator cuff
muscles reported during shoulder exion. While one
previous EMG study appears to indicate higher levels
of infraspinatus compared with subscapularis activity
during shoulder exion (Kronberg et al., 1990)
another reported similar levels of activity in these two
rotator cuff muscles (Townsend et al., 1991). In addi-
tion, similar activity levels have been reported in
supraspinatus and infraspinatus during exion tasks in
some studies (Townsend et al., 1991; Myers et al.,
2005) while another study indicated higher activity in
infraspinatus than supraspinatus (Kronberg et al.,
1990). Finally, lower trapezius activity levels appear
to be similar to serratus anterior levels in some stud-
ies (Moseley et al., 1992; Myers et al., 2005) but
higher in another (Ekstrom et al., 2003).
To provide clinicians with the information neces-
sary to validly interpret shoulder physical assess-
ment ndings and to inform the prescription of
shoulder exercises, a study to clarify the activity lev-
els and recruitment patterns of normal shoulder
muscles during full range shoulder exion is needed.
Because shoulder muscles are required to move the
humerus and scapula and to provide shoulder joint
stability during normal full range shoulder exion, a
comprehensive study would require investigation of
shoulder exor, axioscapular, and rotator cuff
muscles. The addition of load is a common way in
which to progress an exercise program to improve
shoulder muscle function. However, to date, no
study has systematically investigated the effect of
increasing exion load on shoulder muscle recruit-
ment. Finally, no information is available that relates
shoulder muscle activity to the onset of movement
into shoulder exion.
Therefore, the aim of this study was to compre-
hensively examine shoulder muscle recruitment pat-
terns during dynamic exion performed in the sagit-
tal plane in order to:
compare the activity levels of shoulder muscles
that ex the shoulder, laterally rotate the scap-
ula and provide dynamic stability to the
shoulder joint;
compare the recruitment patterns of these same
shoulder muscles;
determine the relative timing between onset of
muscle activity and movement into shoulder
exion;
determine the effect of increasing load on
shoulder muscle recruitment patterns.
MATERIALS AND METHODS
Subjects
The dominant shoulder of 15 subjects (12 males
and 3 females) (mean age, 28.4 years; range, 19
47 years) with normal shoulder function was exam-
ined. Normal shoulder function was dened as hav-
ing no shoulder symptoms for at least the previous
two years and never having sought treatment for
shoulder pain. In addition, a screening physical ex-
amination was performed on all subjects by an expe-
rienced physical therapist to ensure they demon-
strated full pain-free normal range of shoulder
motion, had normal scapulohumeral rhythm
(assessed visually) and did not have pain during
maximal isometric internal and external shoulder
strength tests of their dominant shoulder. The Uni-
versity of Sydney Human Research Ethics Committee
approved this research and all subjects gave
informed written consent prior to participation.
Instrumentation
Twelve shoulder muscle sites were investigated
using a combination of surface and indwelling elec-
trodes. Surface electrodes were used to record activ-
ity in upper trapezius, deltoid (anterior, middle, and
posterior sections), and pectoralis major (sternocla-
vicular portion). The subjects skin was prepared
with alcohol and an abrasive gel to reduce skin im-
pedance. Two surface electrodes (Red Dot, 2258,
3M, Sydney, Australia) spaced 2-cm apart were
placed perpendicular to the orientation of the muscle
bers (Kelly et al., 1996; McLean et al., 2003;
Boettcher et al., 2008). Interelectrode resistances
were measured and ensured to be below 6 kO.
With the subjects lying prone, indwelling electro-
des were inserted into supraspinatus, infraspinatus,
middle, and lower trapezius, serratus anterior, latis-
simus dorsi (Geiringer, 1994), and subscapularis
(Kadaba et al., 1992). Intramuscular electrodes
620 Wattanaprakornkul et al.
were used for muscles that either underlie more su-
percial muscles (supraspinatus, subscapularis), are
thin and overlie other muscles (lower trapezius, latis-
simus dorsi), or for muscles that shift with respect to
the overlying soft tissue during shoulder movement
(infraspinatus, serratus anterior). A large ground
electrode (Universal Electrosurgical Pad: Split,
9160F, 3M, Sydney, Australia) was placed on the
scapula of the nondominant side, extending along
the spine and acromion. Correct indwelling electrode
placement was conrmed by visual comparison of
EMG signals during the performance of standardized
submaximal tests expected to produce a large
amount of activity in each of these muscles and com-
pared with tests expected to generate low activity in
these muscles or activate surrounding muscles into
which the electrode may have been incorrectly
placed.
EMG signals were hardware ltered (HP 10 Hz,
LP 1 kHz) and amplied (Iso-DAM 8 ampliers,
World Precision Instruments, Sarasota, FL; gain
100). The data were acquired on a PC with a 16 bit
analog to digital converter (1401, Cambridge Elec-
tronics Design, Cambridge, UK) at a sample rate of
3571 Hz using Spike2 software (version 4.00, Cam-
bridge Electronics Design, Cambridge, UK) for later
off-line analysis (Matlab version 7, The Math Works,
Natick, MA).
Examination Procedures
On the day prior to the testing, subjects were
instructed as to the correct performance of the ex-
ion movement in the sagittal plane, starting with the
arm by the side moving to full range of shoulder ex-
ion and then back to the arm by the side (Fig. 1) at a
standardized speed. The exion task was performed
with the shoulder in neutral rotation, the elbow in
extension and the forearm in mid-rotation (thumb
pointing up) and completed in six seconds3 sec to
complete each of the concentric and eccentric
phases. The maximum load (1 RM) each subject
could lift during this exion movement was deter-
mined.
On the testing day, following the conrmation of
correct electrode placement, subjects performed
three repetitions, with a minimum 30-sec break
between repetitions, of the Shoulder Normalization
Tests which are four maximum voluntary isometric
contractions (MVC) that have been reported to have
a 95% chance of eliciting maximum activity in all the
muscles tested in this study (Boettcher et al., 2008).
Resting EMG levels were then established. Two repe-
titions of the exion movement were performed at
no load, 20 and 60% of maximum load, in random
order to avoid any systematic inuences related to
learning effects and/or fatigue. A dual axis goniome-
ter (Norangle II, Noraxon Inc, USA) was used to syn-
chronise the EMG signals with the movement.
Signal and Statistical Analyses
All EMG signals were processed using MATLAB
(Version 14, Mathworks, USA): high pass ltered (10
Hz, 8th order Butterworth), rectied, and low pass
ltered (3 Hz, 8th order Butterworth). The envelopes
of EMG activity extracted in this way were normal-
ized to the maximum EMG value measured during
the MVCs, referred to as %MVC, and then time nor-
malized (0100% of total movement duration) and
averaged across the two repetitions.
The level of muscle activity was graded based on
the %MVC of the normalized EMG activity. Muscle
activity <10% MVC was considered minimal, 10
25% MVC low, 2650% MVC moderate, and >50%
MVC high.
A two factor repeated measures ANOVA (Statistica
Version 7.1 Statsoft, USA) was performed to com-
pare the EMG levels across loads (0, 20, and 60% of
maximum load) and the 12 shoulder muscle sites
examined. Tukeys post hoc analysis was performed
when signicant ANOVA results were obtained. An
alpha level p 0.05 was set as the level of signi-
cance.
To evaluate the consistency of patterns of mus-
cle activation, Pearsons correlation analyses were
conducted between pairs of averaged time normal-
ised EMG signals. The patterns were related across
loads for each muscle and between muscles at
each load.
The onset of activity in each muscle at each load
was found from the ltered and rectied EMG sig-
nals prior to time normalizing by rst selecting a
0.5-sec window, prior to movement onset as base-
line. The average and standard deviation of the ac-
tivity was calculated for each trial and each muscle
at each load. The point at which the activity levels
in the muscles reached 2 standard deviations from
the baseline mean was selected as the time of
the initiation of the activity in that muscle. The
mean timing and 95% condence intervals were
Fig. 1. Full range shoulder exion performed in the
sagittal plane.
621 Muscle Recruitment During Shoulder Flexion
calculated for each muscle at each load. A negative
starting time indicates the given muscle was acti-
vated prior to movement. A two factor repeated
measures ANOVA was performed to compare the
onset of activity across loads (0, 20, and 60% of
maximum load) and the 12 shoulder muscle sites
examined. Tukeys post hoc analysis was performed
when signicant ANOVA results were obtained. A
single sample T test was used to compare the onset
of activity in each muscle at each load to zero
(movement onset time). An alpha level p 0.05
was set as the level of signicance.
RESULTS
Activation Levels
The normalized average EMG (mean 6 SD) of
the twelve shoulder muscle sites examined at no
load, 20%, and 60% of maximum load are illus-
trated in Figure 2. In general, there was a system-
atic increase in muscle activity as load increased.
Most of the muscles (pectoralis major, anterior and
middle deltoid, supraspinatus, infraspinatus, ser-
ratus anterior, upper, and lower trapezius) were
activated at approximately 40% MVC on average
during the exion task performed at 60% maxi-
mum load. Posterior deltoid, subscapularis, and
middle trapezius were activated at low levels
(*20%MVC) during the highest load exion task
examined with latissimus dorsi minimally activated
(<10% MVC).
The results of the two factor repeated measures
ANOVA indicated that the activity levels increased
systematically with increasing load (F
2,28
157.6, P
< 0.01) and that there were different activity levels
across muscles (F
11,154
13.0, P < 0.01). There was
also a signicant interaction between load and mus-
cle (F
22,308
5.2, P < 0.01).
The results of the Tukey post-hoc analyses indi-
cated the following.
Differences in Muscle Activation Levels
Between Muscles Across All Loads.
all parts of deltoid were activated at similar lev-
els to pectoralis major (P 0.09) and all were
activated at signicantly higher levels than latis-
simus dorsi (P < 0.05). Anterior deltoid was
activated at signicantly higher levels than pos-
terior deltoid (P < 0.05).
supraspinatus and infraspinatus (RC muscles
that attach to the posterior/dorsal surface of the
scapulaposterior RC muscles) were acti-
vated at similar levels (P 0.35), and both
were activated at signicantly higher levels than
subscapularis (RC muscle attaching to the ante-
rior/ventral surface of the scapulaanterior RC
muscle) (P < 0.05).
upper and lower trapezius and serratus anterior
were activated at similar levels (P 0.73), and
all were activated at signicantly higher levels
than middle trapezius (P < 0.05).
Differences in Muscle Activation Levels of
Each Muscle Between Loads. Muscle activity
increased with increasing load (P < 0.05) for all
muscles except for latissimus dorsi which was the
Fig. 2. The average (6SD) EMG levels (%MVC) of the 12 muscle sites exam-
ined during full range exion performed in the sagittal plane with no load, 20 and
60% of maximum load.
622 Wattanaprakornkul et al.
only muscle activated at minimal levels at all loads
(P 1.00).
Patterns of Muscle Activation
The time normalised and averaged muscle activity
of all muscles examined during exion at no load, 20
and 60% maximum load are indicated in Figure 3. The
correlation coefcients that relate the patterns of activ-
ity in each muscle across the three loads are indicated.
Most of the shoulder muscles examined (pectoralis
major, all parts of deltoid, supraspinatus, infraspina-
tus, serratus anterior, upper, and lower trapezius)
demonstrated a consistent activation pattern (r
0.72, P < 0.05), subscapularis and middle trapezius
activation patterns were less consistent (r 0.51, P <
0.05) and latissimus dorsi activation patterns were
inconsistent (0.61 r 0.10) across the three loads.
Relating the patterns of activity between muscles,
the following was found:
supraspinatus and infraspinatus had similar pat-
terns at all loads (r 0.55, P < 0.05);
serratus anterior and all parts of trapezius and
deltoid had similar patterns at all loads (r
0.69, P < 0.05);
supraspinatus, infraspinatus, and pectoralis
major had similar patterns (r 0.78, P < 0.05)
at high load only.
Onset of Muscle Activity
The average (95% condence limits) time of onset
of activity for each muscle at each load is illustrated
in Figure 4. There was a signicant difference in the
onset of activity between muscles (F
11,154
12.35, P
< 0.01) with no differences between loads (F
2,284

2.71, P 0.08). The onset of activity in supraspina-
tus and anterior deltoid occurred at the same time (P
> 0.99) and prior to movement of the limb (t
14

2.67, P < 0.05) at all loads. With the addition of an
external load (20 and 60% maximum load) the onset
of activity in infraspinatus occurred at the same time
as supraspinatus and anterior deltoid (P > 0.05) and
prior to movement of the limb (t
14
3.64, P < 0.05).
DISCUSSION
The results of this comprehensive evaluation of
muscle activity during full range shoulder exion
performed in the sagittal plane indicate that the av-
erage activity levels in pectoralis major were similar
to those in all parts of deltoid at low, medium, and
high loads. As pectoralis major (Palastanga et al.,
2006) and anterior deltoid (Kronberg et al., 1990;
Townsend et al., 1991; McCann et al., 1993; Myers
et al., 2005) are known to produce exion torque,
activity in these muscles was expected during the
Fig. 3. Time normalized EMG levels (%MVC) aver-
aged across all subjects of the 12 muscle sites exam-
ined during full range exion performed in the sagittal
plane with no load, 20 and 60% of maximum load. Bot-
tom right graph shows goniometer output normalized to
maximum voltage.
623 Muscle Recruitment During Shoulder Flexion
dynamic exion task examined. Conversely, because
latissimus dorsi produces shoulder extension, the
minimal activity recorded in this muscle during ex-
ion at all loads was also expected (Palastanga et al.,
2006).
However, a more detailed consideration of the
plane of the exion movement examined is required
to explain the activity recorded in middle and poste-
rior deltoid. In this study, exion was performed in
the sagittal plane. Contraction of pectoralis major
will not only ex but also adduct the shoulder caus-
ing movement toward the midline and, therefore,
out of the sagittal plane (Palastanga et al., 2006). In
order to maintain movement in the sagittal plane
deltoid activity is required to produce abduction to
prevent pectoralis major from adducting the
shoulder, i.e., deltoid must act as a synergist for
pectoralis major. This could explain the similar level
of activity recorded in pectoralis major, and middle
and posterior deltoid, and further explain the signi-
cantly higher activity levels recorded in anterior del-
toid than posterior deltoid in this study. As anterior
deltoid can produce both shoulder exion and abduc-
tion it is likely to not only be producing exion tor-
que, but also contributing to canceling out unwanted
shoulder adduction.
Even though pectoralis major and anterior deltoid
were activated at similar average levels during the
dynamic exion task examined, the pattern of acti-
vation for each of these muscles was different. Ante-
rior deltoid was more active than pectoralis major in
inner range exion i.e. from 908 to full range exion.
In this range, activity in the sternal head of pectora-
lis major will result in shoulder extension i.e., will
oppose the desired movement of exion in the sagit-
tal plane. The motor strategy demonstrated in this
study to recruit anterior deltoid at higher levels than
pectoralis major in inner range exion, presumably
to produce the exion torque, is therefore logical.
Contrary to the common belief that all the RC
muscles are activated at similar levels to provide
dynamic shoulder joint stability (Magarey and Jones,
2003), the results of this study indicate that the RC
muscles were activated at signicantly different lev-
els during dynamic exion in the sagittal plane. Both
posterior RC muscles (supraspinatus and infraspina-
tus) were activated at signicantly higher levels than
the anterior RC (subscapularis) at all load levels
examined in the current study. In addition, supraspi-
natus and infraspinatus were activated at similar av-
erage levels (*50% MVC) and at similar levels to
pectoralis major and deltoid. The higher activity lev-
els in the posterior RC muscles compared with the
anterior RC muscle support the ndings of a previous
EMG study conducted by Kronberg et al. (1990).
Although they only reported the mean normalized
EMG activity levels of shoulder muscles without a
measure of the variance in the data and without sta-
tistical analysis to conrm signicant differences in
shoulder muscle activity, infraspinatus appears to be
activated at higher levels than subscapularis during
the dynamic exion task performed with a standard
load examined in this earlier study.
Because the RC muscles are not being activated
at similar levels, it is unlikely that they are providing
shoulder joint dynamic stability during exion via the
mechanism of concavity compression (Wilk et al.,
1997; Ludewig and Borstead, 2005). One possible
explanation of the signicantly higher posterior RC
activity during exion in the sagittal plane is that
supraspinatus and infraspinatus are functioning to
prevent humeral head translation in a manner simi-
lar to the role of the inferior RC muscles during
shoulder abduction (Inman et al., 1944; Blasier
et al., 1992; Sharkey et al., 1994) i.e., to prevent
deltoid from superiorly translating the humeral head.
Evidence indicates that pectoralis major can displace
the humeral head anteriorly (Sinha et al., 1999;
McMahon et al., 2002). In a cadaver model McMahon
and Lee (2002) demonstrated an increase in anteri-
orly directed force on the humeral head when pec-
toralis major loading was emulated. In addition, a
case of irreducible anterior shoulder dislocation
caused and maintained by tonic spasm of the pector-
alis muscles has been reported (Sinha et al., 1999).
The signicantly higher activity in the posterior RC
muscles demonstrated during the exion task exam-
ined in the current study suggests that the mecha-
nism whereby these muscles are providing dynamic
shoulder stability is by counterbalancing potential
Fig. 4. The average (695% condence intervals)
time of onset of activity in the 12 muscle sites examined
from the onset of the exion movement (0 sec) per-
formed with no load and at 20 and 60% of maximum
load.
624 Wattanaprakornkul et al.
anterior translation of the humeral head by the ex-
ion torque producing muscles.
The similar activation pattern demonstrated by
supraspinatus, infraspinatus, and pectoralis major
provide further evidence in support of the above
suggestion that the posterior RC muscles are func-
tioning to stabilize the shoulder joint by preventing
pectoralis major from anteriorly translating the hum-
eral head. As pectoralis major activity increases,
higher anterior translation forces will be applied to
the humeral head. This will require simultaneous
increases in posterior RC activity to provide a poste-
rior force of equal magnitude to prevent the humeral
head from gliding forward on the glenoid fossa, thus
maintaining appropriate articular surface alignment
at the shoulder joint during movement into exion.
In this study supraspinatus, infraspinatus and an-
terior deltoid were the only three muscles consis-
tently recruited prior to movement onset. The
recruitment of supraspinatus and infraspinatus prior
to movement onset supports the argument proposed
above that the posterior RC muscles are responsible
for providing dynamic shoulder joint stability during
exion. However, contraction of the posterior RC
muscles will result in shoulder external rotation
(Boettcher et al., 2010). To prevent this external
rotation, so the exion task could be performed in
shoulder mid-range rotation as examined in this
study, an internal rotation torque would be required.
As anterior deltoid can internally rotate the shoulder
(Palastanga et al., 2006), its recruitment prior to
movement, as demonstrated in this study could be
to perform this role.
This study found that upper and lower trapezius
and serratus anterior demonstrated similar levels of
activity regardless of load, as has also been reported
in previous EMG studies investigating dynamic
shoulder exion (Moseley et al., 1992; Ekstrom
et al., 2003; Myers et al., 2005). The current study
has also demonstrated that there were no signicant
differences in the activity levels of these axioscapular
muscles and the muscles producing exion torque
(pectoralis major and anterior deltoid) at all loads
examined. In addition, the recruitment patterns of
anterior deltoid, serratus anterior, and upper and
lower trapezius were similar throughout exion dur-
ing all loads examined, reecting the coordination
between the scapula and humerus required to
achieve full range shoulder exion movement. As
upper and lower trapezius and serratus anterior form
the force couple that produces lateral rotation of the
scapula which accompanies exion at the shoulder
joint, these results were expected. Activity in all the
muscles that contribute to full range shoulder exion
movement (muscles exing the humerus and
muscles laterally rotating the scapula) would be
expected to increase similarly as load increases.
In this study, serratus anterior was the only axios-
capular muscles to be recruited prior to the onset of
exion movement, although this was only statistically
signicant at the 20% maximum load level. As one of
the traditionally accepted roles of serratus anterior is
to stabilize the scapula against the posterior thoracic
cage (Moseley et al., 1992), perhaps the early recruit-
ment of serratus anterior during dynamic exion is
reecting a scapular stabilizing role. As discussed ear-
lier, anterior deltoid, supraspinatus, and infraspinatus
are also recruited prior to the onset of exion move-
ment. As all these muscles attach from the scapula
to the humerus, activity in these muscles not only
has the potential to move the humerus, but also to
move the scapula if axioscapular muscles do not con-
tract to stabilize the scapula against movement. It
would seem that the earlier contraction of serratus
anterior prior to the onset of movement is in response
to the contraction of posterior RC and anterior deltoid
in order to prevent these muscles from destabilizing
the scapula so that normal coordinated scapulohum-
eral movement can occur during exion.
During the dynamic exion task examined in the
current study, the recruitment pattern across time of
all muscles activated above minimum levels was simi-
lar during low, medium, and high loads with a sys-
tematic increase in activity level in all active muscles
as load increased. This result indicates that the
shoulder muscle recruitment pattern to produce ex-
ion including exion torque producing, RC and axios-
capular muscles, is established at low load levels and
does not vary as load increases. These results indi-
cate, therefore, that the normal strategy to accom-
modate increasing exion load is to increase activity
in all the shoulder muscles which are recruited at low
levels, and thus, do not support the Law of Minimal
Muscle Action proposed by MacConaill and Basma-
jian (1977, p 190). This law proposes that muscle
recruitment is energetically efcient. Muscles that
specically performed the required movement are
recruited under low load conditions with additional
less specic muscles recruited as load increases. The
similar recruitment patterns in shoulder muscles
regardless of load, demonstrated in the current study,
support increasing evidence that normal shoulder
motor patterns do not vary with load. Similar
increases in activity with increasing load in all
shoulder muscles recruited at low load levels have
been reported during shoulder rotation (Dark et al.,
2007; Boettcher et al., 2010), abduction (Alpert
et al., 2000) and adduction (Reed et al., 2010).
ACKNOWLEDGMENTS
The protocol for this study was approved by the
Human Research Ethics Committee at the University
of Sydney (Reference no.:07-2007/10122).
REFERENCES
Alpert SW, Pink MM, Jobe FW, McMahon PJ, Mathiyakom W. 2000.
Electromyographic analysis of deltoid and rotator cuff function
under varying loads and speeds. J Shoulder Elbow Surg 9:47
58.
Blasier RB, Guldberg RE, Rothman ED. 1992. Anterior shoulder sta-
bility: Contributions of rotator cuff forces and the capsular liga-
ments in a cadaver model. J Shoulder Elbow Surg 1:140150.
Boettcher CE, Cathers I, Ginn KA. 2010. The role of shoulder
muscles is task specic. J Sci Med Sport 13:651656.
Boettcher CE, Ginn KA, Cathers I. 2008. Standard maximum iso-
metric voluntary contraction tests for normalizing shoulder mus-
cle EMG. J Orthop Res 26:15911597.
625 Muscle Recruitment During Shoulder Flexion
Dark A, Ginn KA, Halaki M. 2007. Shoulder muscle recruitment pat-
terns during commonly used rotator cuff exercises: An electro-
myographic study. Phys Ther 87:10391046.
Ekstrom RA, Donatelli RA, Soderberg GL. 2003. Surface electromyo-
graphic analysis of exercises for the trapezius and serratus ante-
rior muscles. J Orthop Sports Phys Ther 33:247258.
Geiringer SR. 1994. Anatomic Localization for Needle Electromyog-
raphy. Sydney: Hanley and Belfus, Inc.
Inman VT, Saunders JB, Abbott LC. 1944. Observations on the func-
tion of the shoulder joint. J Bone Joint Surg Am 26:130.
Ito N. 1980. Electromyographic study of shoulder joint. J Jpn Orthop
Ass 54:15291540.
Kadaba MP, Cole A, Wootten ME, McCann P, Reid M, Mulford G, April
E, Bigliani L. 1992. Intramuscular wire electromyography of the
subscapularis. J Orthop Res 10:394397.
Kelly BT, Kadrmas WR, Speer KP. 1996. The manual muscle exami-
nation for rotator cuff strength. An electromyographic investiga-
tion. Am J Sports Med 24:581588.
Kronberg M, Nemeth G, Brostrom LA. 1990. Muscle activity and
coordination in the normal shoulder. An electromyographic
study. Clin Orthop Relat Res 257:7685.
Ludewig P, Borstead J. 2005. Upper Extremity Joint Complexes: The
Shoulder Complex. 4th Ed. Philadelphia, PA: F.A. Davis Co.
Magarey ME, Jones MA. 2003. Dynamic evaluation and early man-
agement of altered motor control around the shoulder complex.
Man Ther 8:195206.
McCann PD, Wootten ME, Kadaba MP, Bigliani LU. 1993. A kinematic
and electromyographic study of shoulder rehabilitation exer-
cises. Clin Orthop Relat Res 288:179188.
McLean L, Chislett M, Keith M, Murphy M, Walton P. 2003. The effect
of head position, electrode site, movement and smoothing win-
dow in the determination of a reliable maximum voluntary acti-
vation of the upper trapezius muscle. J Electromyogr Kinesiol
13:169180.
McMahon PJ, Eberly VC, Yang BY, Lee TQ. 2002. Effects of anteroin-
ferior capsulolabral incision and resection on glenohumeral joint
reaction force. J Rehabil Res Dev 39:535542.
Moseley JB Jr., Jobe FW, Pink M, Perry J, Tibone J. 1992. EMG anal-
ysis of the scapular muscles during a shoulder rehabilitation pro-
gram. Am J Sports Med 20:128134.
Myers JB, Pasquale MR, Laudner KG, Sell TC, Bradley JP, Lephart
SM. 2005. On-the-eld resistance-tubing exercises for throwers:
An electromyographic analysis. J Athl Train 40:1522.
Palastanga N, Field D, Soames R. 2006. Anatomy and Human Move-
ment: Structure and Function. 5th Ed. Edinburgh: Butterworth-
Heinemann Elsevier.
Reed D, Halaki M, Ginn K. 2010. The rotator cuff muscles are acti-
vated at low levels during shoulder adduction: An experimental
study. J Physiother 56:259264.
Sharkey NA, Marder RA, Hanson PB. 1994. The entire rotator
cuff contributes to elevation of the arm. J Orthop Res 12:699
708.
Sinha A, Higginson DW, Vickers A. 1999. Use of botulinum A toxin
in irreducible shoulder dislocation caused by spasm of pectoralis
major. J Shoulder Elbow Surg 8:7576.
Susan S. 2008. Grays Anatomy: The Anatomical Basis of Clinical
Practice. Edinburgh: Churchill Livingstone.
Townsend H, Jobe FW, Pink M, Perry J. 1991. Electromyographic
analysis of the glenohumeral muscles during a baseball rehabili-
tation program. Am J Sports Med 19:264272.
Wilk KE, Arrigo CA, Andrews JR. 1997. Current concepts: The stabi-
lizing structures of the glenohumeral joint. J Orthop Sports Phys
Ther 25:364379.
626 Wattanaprakornkul et al.