J. Phys. Ther. Sci.

Original Article 26: 1871–1874, 2014

Rehabilitation Exercises to Induce Balanced

Scapular Muscle Activity in an Anti-gravity Posture

Tomonobu Ishigaki, PT1, 2), Masanori Yamanaka, PhD3)*, Motoki Hirokawa, PT1),
K eita Tai, PT4), Yuya Ezawa, PT1, 2), Mina Samukawa, PhD3), Harukazu Tohyama, MD3),
Makoto Sugawara, MD2)
1) Graduate School of Health Sciences, Hokkaido University, Japan
2) Matsuda Orthopaedic Memorial Hospital, Japan
3) Faculty of Health Sciences, Hokkaido University: West 5, North 12, Kita-ku, Sapporo, Hokkaido

060-0812, Japan
4) Higashi-Naebo Hospital, Japan

Abstract. [Purpose] The purpose of this study was to compare the intramuscular balance ratios of the upper
trapezius muscle (UT) and the lower trapezius muscle (LT), and the intermuscular balance ratios of the UT and the
serratus anterior muscle (SA) among prone extension (ProExt), prone horizontal abduction with external rotation
(ProHAbd), forward flexion in the side-lying position (SideFlex), side-lying external rotation (SideEr), shoulder
flexion with glenohumeral horizontal abduction load (FlexBand), and shoulder flexion with glenohumeral hori-
zontal adduction load (FlexBall) in the standing posture. [Methods] The electromyographic (EMG) activities of
the UT, LT and SA were measured during the tasks. The percentage of maximum voluntary isometric contraction
(%MVIC) was calculated for each muscle, and the UT/LT ratios and the UT/SA ratios were compared among the
tasks. [Results] The UT/LT ratio with the FlexBand was not significantly different from those of the four exercises
in the side-lying and prone postures. The UT/SA ratio with the FlexBall demonstrated appropriate balanced activity.
[Conclusion] In an anti-gravity posture, we recommend the FlexBand and the FlexBall for inducing balanced UT/
LT and UT/SA ratios, respectively.
Key words: Scapular muscle exercise, Balanced scapular muscle activity, Electromyography
(This article was submitted Apr. 16, 2014, and was accepted May 29, 2014)

INTRODUCTION that four exercises were able to facilitate LT activity with

minimal activation of the UT, from the point of view of the
Recently, scapula movement has begun to attract atten- scapular muscle balance (Fig. 1a–d)13). A previous study re-
tion. Previous studies have reported relationships between ported that the scapular muscle rehabilitation exercises per-
the resting position and the dynamic motion (dyskinesis) formed by athletes with impingement syndrome improved
of the scapula and shoulder pain and dysfunction1–5). De- their pain and function14). All of the exercises are carried
creased activation of the lower trapezius (LT) muscle and out in side-lying or prone postures, even though upper ex-
the serratus anterior (SA) muscle together with simultane- tremity activities of daily living frequently occur in anti-
ous excessive activation of the upper trapezius (UT) muscle gravity postures such as sitting or standing. To the best of
have been reported to be related to scapula dyskinesis6). our knowledge, few studies have examined scapular muscle
Some previous studies have reported that scapular muscle exercises in sitting or standing postures which aim to elicit
exercises positively affect not only shoulder pain but also balanced scapular muscle activity. A high level of maxi-
shoulder function7, 8). Restoration of balanced scapular mum voluntary isometric contraction in the LT muscle was
muscle activity is one of the aims of scapular muscle ex- reported for exercises including scapular retraction mo-
ercises, and of decreased activation of the LT and the SA tion15). SA muscle activity has been induced not only with
and increased activation of the UT indicate the importance shoulder horizontal adduction16), but also with shoulder
of the intramuscular balance ratio (UT/LT) and intermus- horizontal abduction in an anti-gravity posture17). Taking
cular balance ratio (UT/SA)6, 9–12). Cools et al. concluded these previous studies into account, we hypothesized that if
glenohumeral flexion with horizontal abduction directional
load or with horizontal adduction directional load, both of
*Corresponding author. Masanori Yamanaka (E-mail: which induce scapular retractor activity, were peformed,
yamanaka@hs.hokudai.ac.jp) balanced scapular intramuscular and also intermuscular ac-
©2014 The Society of Physical Therapy Science. Published by IPEC Inc. tivation would be induced in an anti-gravity posture. Thus,
This is an open-access article distributed under the terms of the Cre- the purpose of the current study was to compare the intra-
ative Commons Attribution Non-Commercial No Derivatives (by-nc- muscular and intermuscular scapular muscle ratios (UT/LT,
nd) License <http://creativecommons.org/licenses/by-nc-nd/3.0/>. UT/SA) among commonly used scapular muscle exercises
1872 J. Phys. Ther. Sci. Vol. 26, No. 12, 2014

studies13, 16). During the MVIC tasks, subjects performed

and held each posture for 5 seconds against manual resis-
tance and each trial was repeated three times. After signal
filtering with a sixth order Butterworth 6 Hz low-pass filter,
EMG value of the middle one-second window of the 5 sec-
onds was averaged for each trial. The mean of the trials was
calculated and used as the normalization value.
Each participant performed the four exercises, which
were selected as exercises representative of those used for
promoting balanced scapular muscle activity. They were
prone extension (ProExt), prone horizontal abduction with
Fig. 1. Scapular muscle exercises external rotation (ProHAbd), forward flexion in the side-
a: prone extension (ProExt); b: prone horizontal abduc- lying position (SideFlex), and side-lying with external ro-
tion (ProHAbd); c: forward flexion in the side-lying posi- tation (SideEr) (Fig. 1a–d). In addition, shoulder flexion
tion (SideFlex); d: side-lying external rotation (SideEr); e: with glenohumeral horizontal abduction load (FlexBand)
shoulder flexion with glenohumeral horizontal abduction
and horizontal adduction load (FlexBall) in standing pos-
load (FlexBand); f: shoulder flexion with glenohumeral
ture (Fig. 1e, f) were also performed. The order of the tasks
horizontal adduction load (FlexBall)
was randomized to avoid the effects of muscle fatigue. As
in the previous study by Cools et al.13), all exercises, except
for those performed in the side-lying posture, were per-
carried out in the side-lying and prone postures and flexion formed bilaterally. Before data collection, the participants
with glenohumeral horizontal adduction load or horizontal performed the exercises without resistance to familiarize
abduction load in an anti-gravity posture, such as sitting or themselves with their execution. Each exercise was divided
standing. into concentric, isometric and eccentric phases. Each phase
was performed for 5 seconds regulated by a metronome
SUBJECTS AND METHODS beat. The subjects held 1 kg dumbbells in both hands, and
completed 3 trials of each task. Between trials, the partici-
Sixteen healthy males volunteered for this study (mean pants had 30 seconds rest, and between exercises, they took
age; 21.9 years ± 0.83 years). Exclusion criteria included a 1 minute rest to avoid muscle fatigue. During the exercis-
shoulder pain and an orthopedic and/or a neurological his- es, when necessary, performance correction was verbally
tory of the shoulder. All participants read and signed an in- given to the subjects by the examiner (M. H).
formed consent form prior to their inclusion in this study. All raw EMG signals were transferred to a Windows
This study was approved by the institutional review board computer through a USB analog/digital (A/D) converter
of the Faculty of Health Sciences, Hokkaido University (ID: at 1,000 Hz and a 16 bit A/D board. They were full-wave
13-76). rectified and filtered with a 6th order Butterworth 6 Hz low-
Surface electromyographic (EMG) activities of the pass filter. For each phase, the data were averaged across the
UT, LT and SA muscle were measured. Before placing the middle 3 seconds of the 5 seconds. Then, the data of each
surface electrodes, the skin was cleaned with alcohol on trial were averaged within the same exercise. The results
a cotton swab in order to get a good electrode-skin con- were normalized to the MVIC data. The mean EMG data,
tact. When necessary, the skin was shaved to reduce skin expressed as a %MVIC, were used to assess the activities of
artifacts. Bipolar surface electrodes (Blue Sensor P-00-S, the UT, LT and SA muscle in each exercise.
Ambu, Denmark) were placed over the UT and the LT In order to assess intermuscle and intramuscle balance
muscles and the lower region of the SA muscle with a 2 cm of the scapular muscle, the UT/LT ratio and the UT/SA ra-
inter-electrode distance. Placement of the electrodes for tio were calculated. One-way repeated analysis of variance
the UT and the LT followed SENIAM18), and the electrode (ANOVA), and the Tukey HSD test as a post hoc test, were
placement for the SA followed the descriptions of previous conducted for the comparison of data across the exercises
studies13, 16). The reference electrode was placed over the in each phase. The statistical analysis was performed using
seventh cervical vertebra. All of the electrodes were con- the IBM SPSS Statistics 18 software program (IBM, Chi-
nected to a MyoSystem 1,200 electromyographic receiver cago, IL, USA). The α level for the analysis of variance was
(Noraxon USA Inc., Scottsdale, AZ, USA). Unit specifica- chosen as 0.05. In addition, following the study of Cools et
tions include a differential input impedance of greater than al.13), the exercises were divided into 3 subgroups based on
10 MΩ, a gain of 1000, and a common mode rejection ratio the ratios of 100 to 80% (moderate), 80 to 60% (good), and
of greater than 100 dB at 60 Hz. The sampling rate was set <60% (excellent), and categorized as follows: exercises in
at 1,000 Hz. Baseline noise was filtered with a band-pass which the ratio was smaller than 60% in each phase (cat-
filter of 10–500 Hz. egory 1); exercises in which the ratio was smaller than 80%
Before performing the series of the scapular muscle in each phase, with at least 1 phase having a ratio between
activation tasks, the EMG signal quality was verified for 60 to 80% (category 2); exercises with a ratio much smaller
each muscle during the performance of the maximal volun- than 100%, with at least 1 phase between the 60 to 80% lim-
tary isometric contraction (MVIC) tests specified for each its (category 3); and exercises with at least 1 of the 3 phases
muscle of interest as described by SENIAM and previous
 1873

Table 1. The UT/LT ratios of all the exercises

ProExt ProHAbd SideFlex SideEr FlexLoop FlexBall

Concentric 55.2±50.8% 42.4±36.7% 82.8±494.7% 38.9±52.6% 112.0±58.2% 299.5±259.5% *
Isometric 85.3±75.2% 59.5±29.3% 20.8±17.9% 32.4±41.8% 93.8±59.9% 539.5±502.9% *
Eccentric 74.8±64.1% 66.6±30.3% 28.9±50.7% 39.9±50.69% 118.7±76.9% 353.0±272.2% *
*: Significantly greater than the other exercises

Table 2. The UT/SA ratios of all the exercises

ProExt ProHAbd SideFlex SideEr FlexLoop FlexBall

Concentric 265.8±313.1% 714.3±596.0% * 244.1±644.9% 431.9±660.9% 81.8±55.6% 46.1±32.7% ***
Isometric 380.8±534.5% 1,335.7±1,038.9% ** 93.6±103.1% 358.7±532.9% 67.4±41.7% 56.1±29.5%
Eccentric 338.3±436.4% 1,100.2±939.1% ** 134.4±121.3% 372.6±567.3% 103.7±76.4% 49.6±30.2%
*: Significant differences between all exercises except the FlexBall exercise
**: Significant differences between all exercises
***: Significant difference from the prone horizontal abduction exercise

much higher than 100% (category 4). in an anti-gravity posture were significantly lower than or
similar to the exercises in the side-lying and prone postures,
RESULTS we concluded that the exercises in the anti-gravity posture
had induced balanced scapular muscle activity.
The results of one-way repeated ANOVA for the UT/ Our study revealed that the UT/LT ratio of the FlexBand
LT ratio are summarized in Table 1. As the result of the exercise was not significantly different from those of the
ANOVA, there were significant differences in each phase four exercises performed in the side-lying and prone pos-
(p<0.001). There were no significant differences between tures in each phase. Because horizontal abduction load elic-
FlexBand exercise and the other exercises (p>0.05 among its lower trapezius muscle activity in the FlexBand exer-
all exercises), whereas the FlexBall exercise demonstrated a cise, the FlexBand exercise induced balanced intra-scapular
significantly greater UT/LT ratio than those of the all other muscle activity. However, the FlexBall exercise elicited a
exercises in the concentric, isometric and eccentric phases significantly greater UT/LT ratio than the other exercises
(p<0.05, respectively). in each phase. Based on these results, we recommend that
As the result of the ANOVA, Significant differences the FlexBand exercise be used to enhance balanced muscle
were found in the UT/SA ratio (Table 2) in each phase activity between the UT and LT muscles.
(p<0.001). The ProHAbd exercise demonstrated a sig- A previous study by Cools et al. reported that the ProExt,
nificantly greater UT/SA ratio than the other exercises in ProHAbd, SideFlex, and SideEr exercises were not suitable
each phase (p<0.001). The SideEr exercise demonstrated a for inducing balanced UT/SA activity13). Our present results
significantly greater UT/SA ratio than the other exercises indicate that the ProHAbd exercise elicited a significantly
in all phases (p<0.05), except the FlexBall exercise in the greater UT/SA ratio than the other exercises in all the phas-
concentric phase. The UT/SA ratio of the FlexBall exercise es, while the FlexBall exercise demonstrated a significantly
was significantly smaller than those of the ProHAbd and the smaller UT/SA ratio than the ProHAbd exercise in the con-
SideEr exercises in the concentric phase (p<0.05). centric phase. However, under the classification criteria, the
According to the described above classification, the UT/ FlexBall exercise belonged to category one. Therefore, we
SA ratio of the FlexBall exercise belonged to category one, propose that the FlexBall exercise might induce balanced
whereas the UT/LT ratio of the FlexBand and the FlexBall UT/SA activity in an anti-gravity posture.
exercises belonged to category four. This study has some limitations. First, during the Flex-
Ball exercise, the participants held a ball and dumbbells,
DISCUSSION and this might have changed muscle activity. Second, UT/
SA ratio varied widely among subjects. We should consider
The purpose of the present study was to compare the these limitations when conducting the exercises with pa-
UT/LT ratio and the UT/SA ratio among commonly used tients.
scapular muscle exercises which are carried out in side-
lying and prone postures and flexion with glenohumeral REFERENCES
horizontal abduction or horizontal adduction load in an
anti-gravity posture. To the best of our knowledge, this is 1) McClure PW, Michener LA, Karduna AR: Shoulder function and 3-dimen-
sional scapular kinematics in people with and without shoulder impinge-
the first study to have investigated a strategy that induces ment syndrome. Phys Ther, 2006, 86: 1075–1090. [Medline]
balanced scapular muscle activity in anti-gravity postures 2) Merolla G, De Santis E, Campi F, et al.: Supraspinatus and infraspinatus
such as sitting and standing. In our study, when the intra- weakness in overhead athletes with scapular dyskinesis: strength assess-
ment before and after restoration of scapular musculature balance. Muscu-
muscular and intermuscular activity ratios of the exercises
1874 J. Phys. Ther. Sci. Vol. 26, No. 12, 2014

loskelet Surg, 2010, 94: 119–125. [Medline] [CrossRef] Surg, 2008, 17: 48S–53S. [Medline] [CrossRef]
3) Ludewig PM, Reynolds JF: The association of scapular kinematics and 11) Cools AM, Witvrouw EE, Declercq GA, et al.: Scapular muscle recruit-
glenohumeral joint pathologies. J Orthop Sports Phys Ther, 2009, 39: 90– ment patterns: trapezius muscle latency with and without impingement
104. [Medline] [CrossRef] symptoms. Am J Sports Med, 2003, 31: 542–549. [Medline]
4) Struyf F, Nijs J, Baeyens JP, et al.: Scapular positioning and movement 12) Ludewig PM, Cook TM: Alterations in shoulder kinematics and associated
in unimpaired shoulders, shoulder impingement syndrome, and glenohu- muscle activity in people with symptoms of shoulder impingement. Phys
meral instability. Scand J Med Sci Sports, 2011, 21: 352–358. [Medline] Ther, 2000, 80: 276–291. [Medline]
[CrossRef] 13) Cools AM, Dewitte V, Lanszweert F, et al.: Rehabilitation of scapular
5) Mihata T, Jun BJ, Bui CN, et al.: Effect of scapular orientation on shoulder muscle balance: which exercises to prescribe? Am J Sports Med, 2007, 35:
internal impingement in a cadaveric model of the cocking phase of throw- 1744–1751. [Medline] [CrossRef]
ing. J Bone Joint Surg Am, 2012, 94: 1576–1583. [Medline] [CrossRef] 14) De Mey K, Danneels L, Cagnie B, et al.: Scapular muscle rehabilitation
6) Cools AM, Declercq GA, Cambier DC, et al.: Trapezius activity and intra- exercises in overhead athletes with impingement symptoms: effect of a
muscular balance during isokinetic exercise in overhead athletes with im- 6-week training program on muscle recruitment and functional outcome.
pingement symptoms. Scand J Med Sci Sports, 2007, 17: 25–33. [Medline] Am J Sports Med, 2012, 40: 1906–1915. [Medline] [CrossRef]
7) Park SI, Choi YK, Lee JH, et al.: Effects of shoulder stabilization exercise 15) Kibler WB, Sciascia AD, Uhl TL, et al.: Electromyographic analysis of
on pain and functional recovery of shoulder impingement syndrome pa- specific exercises for scapular control in early phases of shoulder rehabili-
tients. J Phys Ther Sci, 2013, 25: 1359–1362. [Medline] [CrossRef] tation. Am J Sports Med, 2008, 36: 1789–1798. [Medline] [CrossRef]
8) Choi SH, Lee BH: Clinical usefulness of shoulder stability exercises for 16) Decker MJ, Hintermeister RA, Faber KJ, et al.: Serratus anterior muscle
middle-aged women. J Phys Ther Sci, 2013, 25: 1243–1246. [Medline] activity during selected rehabilitation exercises. Am J Sports Med, 1999,
[CrossRef] 27: 784–791. [Medline]
9) Wadsworth DJ, Bullock-Saxton JE: Recruitment patterns of the scapular 17) Park KM, Cynn HS, Yi CH, et al.: Effect of isometric horizontal abduction
rotator muscles in freestyle swimmers with subacromial impingement. Int on pectoralis major and serratus anterior EMG activity during three exer-
J Sports Med, 1997, 18: 618–624. [Medline] [CrossRef] cises in subjects with scapular winging. J Electromyogr Kinesiol, 2013, 23:
10) Moraes GF, Faria CD, Teixeira-Salmela LF: Scapular muscle recruitment 462–468. [Medline] [CrossRef]
patterns and isokinetic strength ratios of the shoulder rotator muscles in 18) SENIAM: Recommendations for sensor locations on individual muscles:
individuals with and without impingement syndrome. J Shoulder Elbow http://www.seniam.org/ (Accessed Feb 5, 2013)