Revision of the Shoulder Normalization Tests Is Required to Include
Rhomboid Major and Teres Major
K.A. Ginn,1 M. Halaki,2 I. Cathers2
1
Discipline of Biomedical Science, Sydney Medical School, The University of Sydney, Lidcombe, NSW 1825, Australia, 2Discipline of Exercise
and Sport Science, Faculty of Health Sciences, The University of Sydney, Lidcombe, NSW, Australia
Received 8 November 2010; accepted 30 May 2011
Published online 27 June 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jor.21488
ABSTRACT: The four ‘‘Shoulder Normalization Tests’’ were found previously to be a parsimonious set of isometric tests that produce
maximal voluntary isometric contractions (MVIC) in the supraspinatus, infraspinatus, subscapularis, trapezius, serratus anterior,
deltoid, latissimus dorsi, and pectoralis major [Boettcher et al. (2008). J Orthop Res 26:1591–1597]. However, these tests have not been
validated for rhomboid major and teres major. In the current study, these Shoulder Normalization Tests were evaluated and compared
to three other tests that could possibly elicit maximum activity in rhomboid major and teres major: abduction/extension in 908
abduction; adduction at 908 abduction; and extension in 308 abduction. No statistical difference was found in the mean activation of
rhomboid major and teres major in these additional MVIC tests compared to the Shoulder Normalization Tests. However, the extension
MVIC test produced maxima for at least 50% of subjects in rhomboid major, teres major, and latissimus dorsi. We concluded that the
original Shoulder Normalization Tests should be expanded to include the extension MVIC test. The EMG normalization reference value
for any of the above muscles would be the maximum EMG level generated across these Revised Shoulder Normalization Tests. ß 2011
Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 29:1846–1849, 2011.
Keywords: EMG; normalization; teres major; rhomboid major; shoulder muscles
Extending the work of previous studies,1,2 Boettcher
et al.3 identified four maximal voluntary isometric
contractions (MVIC) that maximally activated the
supraspinatus, infraspinatus, subscapularis, trapezius,
serratus anterior, deltoid, latissimus dorsi, and pectoralis major. They termed these isometric tests the
‘‘Shoulder Normalization Tests’’ and suggested their
adoption as a standard for generating MVIC for
normalization in shoulder EMG research to facilitate
reliable comparison among studies and thus, to
progress understanding of shoulder muscle function.
Rhomboid major and teres major are two large
shoulder muscles involved in positioning the scapula
and moving the humerus, respectively, which were not
examined by Boettcher et al.3 A standard normalization procedure for shoulder EMG that also incorporates these two muscles would enable a more
comprehensive understanding of the contribution of
muscle dysfunction to shoulder joint pathology. Previous EMG research compared rhomboid major activity
levels during eight isometric manual muscle tests
including three positions specifically aimed at eliciting
maxima in rhomboid muscles: Kendall positionresisted shoulder adduction and elevation in prone
lying;4 Kendall-Alternative position-resisted shoulder
horizontal extension at 908 abduction in prone lying;4
and the Hislop–Montgomery position-resisted shoulder
extension/adduction in prone lying with hand on
sacrum.5 Results indicated that the isometric test
position recommended for posterior deltoid (resisted
shoulder abduction/extension in 908 of shoulder
Correspondence to: K.A. Ginn (T: þ61-2-9351-9352; F: þ61-29351-9520; E-mail: karen.ginn@sydney.edu.au)
ß 2011 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.
1846
JOURNAL OF ORTHOPAEDIC RESEARCH DECEMBER 2011
abduction) elicited higher activity in rhomboid major
than any of the three manual test positions
recommended for rhomboid muscles, although this difference was only significantly higher than the Hislop–
Montgomery test position.6 There do not seem to be
any EMG studies comparing teres major activity
during isometric manual muscle tests.
Therefore, our aim was to compare the activity
levels generated in rhomboid major and teres major
during the four Shoulder Normalization Tests with
those generated during resisted shoulder abduction/
extension in 908 of shoulder abduction, previously
recommended by Smith et al.6 as the manual muscle
test to be used to produce maximal rhomboid activity
for normalization purposes during kinesiologic studies,
and in tests likely to generate maximum activity in
teres major. The manual muscle test position for teres
major recommended by Kendall et al.4 is resisted
shoulder extension and adduction with the humerus in
an internally rotated position. As the Shoulder
Normalization Tests include a maximal isometric
internal rotation test, we included maximal isometric
shoulder extension and adduction tests to include all
shoulder movements likely to generate maximum teres
major activity. Because scapular downward rotation,
produced by rhomboid major, accompanies adduction,
the inclusion of a separate adduction manual test had
the added advantage of potentially generating maximum activity in rhomboid major. Our objective was to
determine if the Shoulder Normalization Tests could
be used to validly determine maximal activity in
rhomboid major and teres major for the purposes of
normalization of shoulder EMG, or if this recommended standard set of shoulder normalization
procedures required modification to include these additional shoulder muscles.
REVISED SHOULDER NORMALIZATION TESTS
METHODS
Subjects
Fourteen subjects (6 female, 8 male, 18–49 years old (mean
22.5 years)), who had not experienced dominant shoulder
pain in the past 2 years and had never sought treatment for
dominant shoulder pain, participated. Subjects demonstrated
normal shoulder range of motion and scapulohumeral
rhythm and be pain-free on isometric internal and external
rotation strength testing. Subjects were fully informed of the
protocol and provided signed consent prior to participation.
The study was approved by the University’s Human
Research Ethics Committee.
Instrumentation
EMG data were collected simultaneously from 11 shoulder
muscle sites using a combination of surface and indwelling
electrodes: rhomboid major, teres major, supraspinatus,
infraspinatus, subscapularis, upper and lower trapezius,
serratus anterior, deltoid, pectoralis major, and latissimus
dorsi. Surface electrodes were used to record activity in
upper trapezius, middle deltoid, and pectoralis major. Before
surface electrode application and with the participant seated,
the skin was prepared with alcohol and an abrasive gel to
reduce skin impedance. Over each muscle site, 2 surface
electrodes (Red Dot, 2258, 3M, Sydney, Australia) were
placed 2 cm apart parallel to the orientation of the muscle
fibers. Inter-electrode resistances were always ensured to be
<10 kV. Intramuscular electrodes, manufactured in our
laboratory using the technique described by Basmajian and
De Luca,7 were inserted into the remaining muscles.8,9
Intramuscular electrodes were used for muscles that either
underlie more superficial muscles (rhomboid major, supraspinatus, subscapularis), are thin and overlie other muscles
(lower trapezius, latissimus dorsi), or for muscles that shift
with respect to the overlying soft tissue during shoulder
movement (teres major, infraspinatus, serratus anterior).
Correct indwelling electrode placement was confirmed by
visual inspection of the EMG signals on a monitor during the
performance of standardized submaximal tests expected to
produce a large amount of activity in each muscle and
compared with tests expected to generate low activity or
activate surrounding muscles into which the electrode may
have been incorrectly placed.3 Because of the difficulty in
distinguishing between rhomboid major and lower trapezius
using this method, intramuscular electrodes were inserted
into these muscles using ultrasonic guidance (Mindray, DP9900). A large ground electrode (Universal Electrosurgical
Pad:Split, 9160F, 3M, Sydney, Australia) was placed over the
contralateral acromion and scapular spine.
EMG signals were amplified and filtered (Iso-DAM 8
amplifiers, World Precision Instruments, Sarasota, FL;
gain ¼ 100, bandpass between 10 and 1 kHz) before transferring to a PC with a 16 bit analog to digital converter (1401,
Cambridge Electronics Design, Cambridge, UK) at a sampling rate of 2,564 Hz using Spike2 software (version 4.00,
Cambridge Electronics Design).
MVIC Tests
Seven MVIC tests were performed in a seated position. These
included shoulder abduction/extension in 908 abduction
(‘‘abduction/extension’’) identified by Smith et al.6 to generate
the greatest activity in rhomboid major, and the 4 Shoulder
Normalization Tests as described by Boettcher et al.3: abduction at 908 with internal rotation; internal rotation in 908
1847
abduction; flexion at 1258 with scapula resistance; and
horizontal adduction at 908 flexion. In addition, shoulder
adduction at 908 abduction (‘‘adduction’’) and shoulder extension at 308 abduction (‘‘extension’’) were included so that all
the known actions of teres major were tested. Each test
lasted 5 s: 1 s to reach maximum, 3 s sustained maximum;
and gradual release over the final 1 s. Each test was performed three times in random order across tests. There was
a minimum rest interval of 30 s between repetitions and a
minimum 1 min rest interval between each new test position.
During testing, subjects could view the raw EMG signal on
the computer screen, were given verbal encouragement, and
were closely monitored to ensure that they did not attempt
compensatory movements of the scapula or trunk.
Signal and Statistical Analyses
EMG signals were high pass filtered (10 Hz, dual pass 4th
order Butterworth), rectified, and low pass filtered (4 Hz,
dual pass 4th order Butterworth) using (Matlab version 7,
The Math Works, Natick, MA). The maximum EMG (MVIC)
was taken as the maximum of this rectified and filtered
signal. The MVIC from abduction/extension, adduction, and
extension tests were compared to the MVIC obtained using
the set of 4 Shoulder Normalization Tests using a single
factor, 4 level repeated measures ANOVA (Statistica, version
7.1, StatSoft, Tulsa, OK) for each muscle. Since no one test
maximally activated a specific muscle in all subjects, the
statistical confidence of the Shoulder Normalization Tests to
produce maximal activation required use of the complete set
of 4 tests.3 Thus, the 4 Shoulder Normalization Tests were
analyzed collectively. Significant results were accepted at
p < 0.05. Tukey’s HSD post hoc test was used when
significant differences were found.
RESULTS
The mean (standard error) MVIC EMG levels for
rhomboid major and teres major for all MVIC tests are
shown in Figure 1. For rhomboid major, no significant
differences were found in the levels for any of the tests
(F3,39 ¼ 1.28, p ¼ 0.29). For teres major, significant
differences were found among tests (F3,39 ¼ 5.02,
p < 0.05). Post hoc tests indicated that the EMG levels
were significantly higher during the extension MVIC
Figure 1. Mean (standard error) MVIC EMG levels obtained
for rhomboid major and teres major during the abduction/extension, adduction, extension MVIC tests, and the Shoulder Normalisation Tests.3 Indicates significant (p < 0.05) difference
between extension and abduction/extension in teres major only.
JOURNAL OF ORTHOPAEDIC RESEARCH DECEMBER 2011
1848
GINN ET AL.
test compared to EMG levels obtained using the shoulder adduction/extension in 908 abduction MVIC test
(p < 0.05), with no significant differences (p 0.05) in
EMG levels across all other MVIC tests.
The number and percentage of subjects that
produced a maximum EMG level for each MVIC test
compared to the 4 Shoulder Normalization Tests are
shown in Table 1. Although no difference was found
between EMG levels among the extension test and the
4 Shoulder Normalization Tests in both rhomboid
major and teres major, the extension test produced a
maximum EMG in both muscles for 50% of subjects,
while the 4 Shoulder Normalization Tests produced a
maximum for <30% of subjects.
To investigate the effects of adding the extension
test to the Shoulder Normalization Tests, the number
and percentage of subjects that produced a maximum
during the extension MVIC test in the nine muscle
sites originally tested by Boettcher et al.3 are shown in
Table 2. Of these muscles, only latissimus dorsi generated maximum EMG levels in a large percentage of
subjects (71%) during the extension MVIC test.
DISCUSSION
Our results indicate that, of the MVIC tests examined,
the shoulder extension test is most likely to achieve
maximal activation of both rhomboid major and teres
major. For rhomboid major, this test generated high
mean EMG activity that was not significantly different
from that generated during the standard set of
shoulder EMG normalization procedures (Shoulder
Normalization Tests) recommended by Boettcher
et al.3 or the abduction/extension and adduction MVIC
tests. For teres major, the extension MVIC test generated significantly higher EMG activity than the abduction/extension test, but not different from that
generated by the Shoulder Normalization Tests.
However, a greater percentage of subjects achieved
maximum rhomboid major and teres major activity
during the extension MVIC test (57% and 50%, respectively) than during the Shoulder Normalization Tests
(14% and 28%, respectively) or the other two MVIC
tests (maximum 14% in both muscles). Because many
tests generated similar high mean EMG activity in
Table 1. The Number and (Percent) of Subjects That
Produced Maximum EMG in Rhomboid Major and Teres
Major during Each of the MVIC Tests and the Shoulder
Normalization Tests3
MVIC Test
Abduction/extension
Adduction
Extension
Shoulder Normalization
Tests
Rhomboid
Major
Teres
Major
2 (14%)
2 (14%)
8 (57%)
2 (14%)
2 (14%)
1 (7%)
7 (50%)
4 (28%)
JOURNAL OF ORTHOPAEDIC RESEARCH DECEMBER 2011
Table 2. The Number and (Percent) of Subjects That
Produced Maximum EMG during the Extension MVIC
Test in the Nine Muscle Sites Previously Tested by
Boettcher et al.3
Muscle
Supraspinatus
Infraspinatus
Subscapularis
Upper trapezius
Lower trapezius
Serratus anterior
Deltoid
Pectoralis major
Latissimus dorsi
Number of
Subjects (%)
0 (0%)
1 (7%)
1 (7%)
0 (0%)
0 (0%)
0 (0%)
0 (0%)
0 (0%)
10 (71%)
many of the muscles they examined, Boettcher et al.3
used the two criteria of tests that produced significantly highest mean EMG activity and highest percentage
of subjects achieving maximum activity, to determine
a parsimonious, standard set of maximum tests for use
in normalizing shoulder EMG data. Based on these
same criteria, our study would suggest that shoulder
extension performed at 308 abduction be added to the
original Shoulder Normalization Tests to generate valid appropriate normalization reference values (Fig. 2).
As acknowledged by Boettcher et al.,3 a standard
set of maximum isometric shoulder normalization tests
may not generate absolute maximum activity in all
subjects; this is also true of the shoulder extension
MVIC test recommended above for use in normalizing
activity in teres major and rhomboid major. There
may be other tests not examined in this study that
may generate absolute maximum activity in rhomboid
major and teres major in a larger percentage of subjects than the recommended shoulder extension MVIC
test. However, we would agree with Boettcher et al.3
that the advantage of adopting a standard set of tests
as soon as possible to make comparisons between
future shoulder EMG studies more reliable outweighs
these limitations.
Our results suggest that the shoulder extension at
308 abduction MVIC test needs to be added to the 4
Shoulder Normalization Tests3 to have a high likelihood of achieving maximal activation in rhomboid
major and teres major. The extension MVIC test also
has the benefit of achieving similarly high levels of
latissimus dorsi activity as the Shoulder Normalization Tests but with 70% of the cohort examined in
this study achieving maximum latissimus dorsi activity during this test. Boettcher et al.3 also found that
latissimus dorsi activity was not significantly different
from its maximum level of activation during a similar
isometric extension test as we examined. Therefore,
because the extension MVIC test is necessary to
reliably achieve maximum activity levels in rhomboid
major and teres major, and because it is likely to
achieve absolute maximum activity levels in latissimus
REVISED SHOULDER NORMALIZATION TESTS
1849
individuals. Therefore, similar to the recommendations
of Boettcher et al.3 the normalization reference level
for each of these muscles should be taken as the
maximum level of activation generated across all five
Revised Shoulder Normalization Tests.
In conclusion, our results indicate that the original
standard shoulder normalization procedures3 should
be expanded to include the isometric extension test
performed at 308 abduction. These Revised Shoulder
Normalization Tests thus include: abduction at 908
with internal rotation; internal rotation in 908 abduction flexion at 1258 with scapula resistance; horizontal
adduction at 908 flexion; and extension at 308 abduction. This set of tests: will be able to reliably generate
maximal activity in more shoulder muscles than the
original Shoulder Normalization Tests; is more likely
to generate maximal activity in latissimus dorsi than
the original Shoulder Normalization Tests; and is still
few in number to minimize subject fatigue. The recommended Revised Shoulder Normalization Tests builds
on the rigorous examination of shoulder normalization
tests by Boettcher et al.3 to provide a standard set of
normalization procedures for use in EMG studies.
REFERENCES
Figure 2. The extension MVIC test: subject seated with the
arm at 308 abduction, elbow fully extended, and thumb toward
the body; arm extended as resistance applied over the distal
forearm.
dorsi in a higher percentage of subjects than the
Shoulder Normalization Tests, we would recommend
that the extension MVIC test be added. This set of five
Revised Shoulder Normalization Tests still satisfies
the criterion of being small in number to reduce the
likelihood of fatigue during subject preparation, while
enabling standardized, reliable normalization of EMG
data recorded from more shoulder muscles than the
original Shoulder Normalization Tests. As demonstrated previously3,10 and supported by our results, maximum activity in many shoulder muscles may be
generated from many isometric tests in different
1. Ekstrom R, Soderberg G, Donatelli R. 2005. Normalization
procedures using maximum voluntary isometric contractions
for serratus anterior and trapezius muscles during surface
EMG analysis. J Electromyogr Kinesiol 15:418–428.
2. Kelly B, Kadrmas W, Kirkendall D, et al. 1996. Optimal
normalisation tests for shoulder muscle activation: An
electromyographic study. J Orthop Res 14:647–653.
3. Boettcher C, Ginn K, Cathers I. 2008. Standard maximum
isometric voluntary contraction tests for normalizing shoulder muscle EMG. J Orthop Res 26:1591–1597.
4. Kendall F, McCreary E, Provance P, et al. 2005. Muscles:
Testing and function with posture and pain. 327–328 pp.
5. Hislop H, Montgomery J. 2002. Muscle testing and function,
7th ed. Philadelphia: WB Saunders. 467 p.
6. Smith J, Padgett D, Kaufman K, et al. 2004. Rhomboid muscle electromyography activity during 3 different manual
muscle tests. Arch Phys Med Rehab 85:987–992.
7. Basmajian J, De Luca C. 1985. Muscles alive: Their
functions revealed by electromyography, 5th ed. Baltimore:
Williams & Wilkins Co. 561 p.
8. Geiringer S. 1994. Anatomic localization for needle electromyography. Philadelphia: Mosby; 153 p.
9. Kadaba M, Cole A, Wooten M, et al. 1992. Intramuscular
wire electromyography of the subscapularis. J Orthop Res
10:394–397.
10. Chopp J, Fischer S, Dickerson C. 2010. On the feasibility of
obtaining multiple muscular maximal voluntary excitation
levels from test exertions: A shoulder example. J Electromyogr Kinesiol 20:896–902.
JOURNAL OF ORTHOPAEDIC RESEARCH DECEMBER 2011