Debra S Anderson, Martha F Jackson, Debra S Kropf 1984; 64:24-28.

Electromyographic Analysis of Selected Muscles
during Sitting Push-ups: Effects of Position and Sex
Debra S Anderson, Martha F Jackson, Debra S Kropf
and Gary L Soderberg
PHYS THER. 1984; 64:24-28.
The online version of this article, along with updated information and services, can
be found online at:
This article, along with others on similar topics, appears
in the following collection(s):
Therapeutic Exercise
To submit an e-Letter on this article, click here or click on
"Submit a response" in the right-hand menu under
"Responses" in the online version of this article.
E-mail alerts
Sign up here to receive free e-mail alerts
Downloaded from by guest on August 29, 2014
Electromyographic Analysis of Selected Muscles
during Sitting Push-ups
Effects of Position and Sex
The purpose of this study was to determine if differences in EMG activity of the
latissimus dorsi, pectoralis major and triceps brachii occurred between muscles
and between the 16 male and 16 female subjects performing push-ups from three
different sitting positions. Surface electrodes and associated instrumentation
recorded a linear envelope during seated push-ups performed 1) in a wheelchair,
2) in a long-sit position with elbows at 90 degrees, and 3) in a long-sit position
with maximum elbow flexion and shoulder abduction. Results showed that women
produced greater mean EMG activity than men in all muscles at all positions.
Altering the exercise position did not have a consistent effect on level of activity
recorded from either sex. The study concludes that use of these exercises should
be based on a knowledge of the differences in muscle activity in exercise
positions for men and women before treatment objectives can be effectively
Key Words: Elbow, Electromyography, Men, Muscle contraction, Shoulder, Women.
Effective use of therapeutic exercise is
of importance and of interest to physical
therapists. The selection of specific exercise techniques may significantly influence achievement of the treatment
goal. One commonly used exercise, the
sitting push-up, has been observed to be
widely used before crutch ambulation
and transfer activities of the spinal cordinjured patient. Strength training purportedly accomplished by this exercise
may also benefit patients with lower extremity amputations, total joint replacements, ligamentous repairs, and fractures.
The standard sitting push-up is used
for all patients, even though men and
women have known differences in
Ms. Anderson, Ms. Jackson, and Ms Kropf were
students in their final year of the Certificate Program
in Physical Therapy at The University of Iowa, Iowa
City, IA 52242 when this study was done.
Ms. Anderson is currently a physical therapist,
Schoitz Medical Center, Waterloo, IA 50702 (USA).
Ms. Jackson is currently a physical therapist,
Walter Boswell Memorial Hospital, Sun City, AZ
Ms. Kropf is currently a physical therapist, Blessing Hospital, Quincy, IL 62301.
Dr. Soderberg is Associate Professor and Associate Director of Physical Therapy Education at
the University of Iowa, Iowa City, IA 52242.
This article was submitted March 10, 1983; was
with the authors for revision five weeks; and was
accepted July 21, 1983.
strength.1 (pl23) Virtually no information
is available on how strength differences
might influence performance during exercise tasks. Other variables, such as sitting height and upper-extremity ranges
of motion, may also create different
muscular demands.
Although several studies have investigated the function of scapulohumeral
muscles during single plane or diagonal
motions, researchers do not understand
how the selection of position influences
the degree of muscle activity.2-6 In spite
of different biomechanical requirements, virtually no work has assessed
the effect of different exercise positions
for men and women. Based on known
anatomical and functional characteristics, however, glenohumeral adductor
and elbow extensor muscles appear to
play a primary role in the performance
of sitting push-ups.7 Because no published studies have reported use of EMG
to assess the performances of several
muscles during the seated push-up, this
study was designed to determine if differences in EMG from the latissimus
dorsi (LD), pectoralis major (PM), and
triceps brachii (TB) muscles occurred
between male and female subjects performing push-ups from three different
sitting positions.
Sixteen healthy men and 16 healthy
women signed a consent form before
voluntarily participating in this study.
Men ranged in height from 1.7 m to 2.0
m with an average of 1.8 m (5.6 ft to
6.6 ft, average 5.9 ft). Their weight
ranged from 63.6 kg to 106.8 kg, averaging 72.7 kg (140.2 lb to 235.5 lb,
average 160.3 lb). Women averaged 59.0
kg in weight and 1.67 m in height (130
lb, 5.5 ft). Their respective ranges were
50.0 kg to 68.2 kg (110.2 lb to 150.4 lb)
and 1.6 m to 1.8 m (5.3 ft to 5.9 ft). Age
of both men and women ranged from
21 to 31 years.
Instrumentation included three bipolar surface electrode assemblies with related signal conditioners. The electrode
assemblies,* which measured 33 mm x
17 mm x 10 mm (1.3 in x 0.7 in x 0.4
in), contained circuitry for preampliiication with a gain of 35. Each assembly
*Rehabilitation Engineering Center, Moss Rehabilitation Hospital, 12th St & Tabor Rd, Philadelphia, PA 19141.
Downloaded from by guest on August 29, 2014
Fig. 1. Subject positioned to perform, from left to right, the standard-wheelchair, mid-position and elevated-sitting push-ups.
held two silver-silver chloride electrodes
that were 8 mm (0.3 in) in diameter
with a 20 mm (0.8 in) distance between
centers. After lightly scrubbing the sub­
jects' skin with alcohol-soaked gauze,
each electrode assembly was fitted with
double-sided foam adhesive tape. Holes
in the tape over the electrode sites were
filled with conductive gel before the tape
was attached to the skin. High input
impedance of the EMG system and the
onsite preamplifier made measurement
of skin resistance impractical and un­
All subjects reported right-hand dom­
inance, and all recordings were taken
from the right side of the body. In all
instances, the electrode assembly was
aligned in the center of and parallel to
the direction of the muscle fibers. The
TB muscle electrode was located 30 per­
cent of the distance from the olecranon
process to the posterior aspect of the
acromion process. The electrode for the
PM muscle was located 50 percent of
the distance from the anterior-most bor­
der of the axilla to the center of the
sternum. For all subjects, the LD muscle
electrode was applied 25 percent of the
distance from the inferior angle of the
scapula to the superior aspect of the iliac
crest. A common ground electrode was
positioned over the muscle-free portion
of the distal ulna, and electrode leads
were plugged into main amplifiers.
The combined preamplifier and main
amplifier system provided a gain from
100 to 10,000 with a bandwidth of 7 Hz
to 6 kHz. Input resistance was less than
15 pF in parallel with 2 mΩ. After am­
plification, the EMG signals were fullwave rectified and subjected to low-pass
filtering with a cut-off frequency of 8
Hz. The result was a linear envelope.
The EMG signals were cabled to an
Interdata 7/16 digital computer,† which
† Interdata, 2 Crescent Place, Oceanport, NJ
sampled each data channel at a rate of
10 samples per second. The analog to
digital converter voltage range was 0 to
plus 1.28 V with resolution of .05 per­
cent of full scale. Signals were simulta­
neously monitored for offsets and arti­
facts on a four-channel oscilloscope. A
separate main amplifier channel was
used to provide raw EMG for purposes
of making periodic checks of the quality
of the signals coming from each pream­
After we applied the electrodes, we
seated the subjects in a standard wheel­
chair. Each was told to perform maxi­
mum contractions during isometric test­
ing. The EMG recordings were made
during three trials of isometric contrac­
tions of each muscle. During these rec­
ordings, subject instruction and data
sampling were computer controlled.
After a teletype keystroke by the inves­
tigator, the computer executed these
following functions: 1) displayed
"READY" on the display screen for two
seconds, 2) displayed "SET" for two sec­
onds (during which the subject prepared
for movement), 3) displayed "GO" for
three seconds while simultaneously
sampling the three EMG channels, 4)
displayed "REST" until a new trial be­
gan, and 5) calculated and printed the
average value for each EMG signal for
the last two seconds of the contraction
period. Occasionally in the early trials,
amplifier gains for each muscle had to
be adjusted so that EMG signals would
be maintained within the 1.28 V limi­
tation of the computer's analog-to-digi­
tal converter.
Because we normalized EMG values,
each subject's maximum two-second av­
eraged EMG was selected from the high­
est value produced during three trials of
maximum isometric effort of each mus­
cle. Test positions were based on pilot
work and adapted so that each could be
performed by a person sitting in a wheel­
chair. Contraction of PM muscle was
achieved by crossing the arms in front
of the body and pressing laterally against
the opposite wheelchair arm. The TB
muscle contraction was obtained by
flexing the shoulder approximately 45
degrees and connecting a strap from the
distal forearm to the highest portion of
the right upright of the wheelchair. For
this test, the elbow was 90 degrees from
the completely extended position. Con­
traction of LD muscle occurred by ex­
tending the humerus against the poste­
rior upright of the wheelchair. In an
attempt to achieve maximal contrac­
tion, the investigator verbally encour­
aged the subjects during the trials. The
sequence of all isometric exercises was
After the investigator collected iso­
metric data from the subjects, they per­
formed three push-ups from each of
three positions. Order of the positions
was randomized. Instruction was given
and practice allowed so that the subject
could demonstrate a constant velocity
and still complete elbow extension in
the required three seconds. Figure 1
shows that the standard (S) push-up was
performed from a seated position in a
conventional wheelchair. For the mid
(M) position, the subject sat with legs
extended (long sit) on the floor with
push-up blocks positioned to cause the
elbows to assume 90 degrees of flexion.
Forearms were perpendicular to the
floor. For the elevated (E) position, the
subject, again in a long-sit position, per­
formed a push-up from the highest point
possible on the push-up blocks. In po­
sitions M and E, the push-up blocks
were aligned in a coronal plane with the
subject's trunk. In all three positions,
the subjects were instructed not to use
their legs to assist with the push-up.
Data Analysis
The data set used in the analysis were
the normalized values, in percentages,
Volume 64 / Number 1, January 1984
Downloaded from by guest on August 29, 2014
Means and Standard Errors for All Muscles Expressed as a Percentage of Isometric
Maximum Values (N = 16 men, 16 women)
Triceps brachii
(SE = 1.81)a
Pectoralis major
(SE = 1.83)a
Latissimus dorsi
(SE = 2.34)a
Pooled across groups and positions.
to the maximum for each muscle for the
three trials from each of the three exercise positions. Means and standard deviations were computed on EMG data
according to sex, muscle, and position.
Differences in between-sex and withinmuscle/between-push-up positions were
tested with a two-way analysis of variance (ANOVA). Bonferroni adjusted t
tests were used for all post hoc analyses.8
Descriptive and ANOVA summary
data for all muscles for each position are
shown in Table 1 and Table 2, respectively. Analysis of variance for the TB
muscle revealed no significant interaction between sex and position. The main
effects for position, subject, and sex were
significant. The t tests, used to compare
differences between positions, revealed
that push-ups performed in the S and E
positions did not produce significantly
different EMG activity in the TB muscle. Activity evoked in the S and E positions were, however, significantly
greater than EMG evoked in the M position. Data are represented graphically
in Figure 2.
Analyses of variance for PM and LD
muscles showed significant interaction
for sex by exercise position (Tab. 2). In
all cases, activity levels were higher for
the women. Comparisons by post hoc t
test revealed significant differences in
the men for PM muscles between positions S and E and between positions M
and E (E had higher levels). Women
showed significant differences between
all positions; M evoked the least activity
and E the greatest. For LD muscles in
the men, the S position produced significantly less activity than either the M or
E positions. LD muscle activity in the
men was significantly different for all
position comparisons; S was the lowest
and E the highest in EMG activity. Data
for the PM and LD muscles are shown
in Figures 3 and 4, respectively.
In interpreting data from this study,
we took care to use a well-established
method for analyzing EMG data. Normalization to each individual's maximum value is a commonly accepted
procedure and indicates the general level
of activity evoked. The method used in
this study is the best method available
to compare various activities even
though isotonic movements confound
the linear and nonlinear relationships
established for EMG and isometric tension.9
The differences between the sexes
were clearly demonstrated when levels
of activity were assessed by muscle. Specific comparisons could not be made
because of the confounding effect of
position, but in general, the values were
20 to 30 percent higher for women (Tab.
1). The fact that strength in women,
when corrected for body size, is only 80
percent of the strength of men probably
accounts for the higher levels of tension
that result in greater EMG values.1
Therapists should note that only for
TB muscles in women was the maximum test EMG exceeded in positions S
and E. Because EMG is a strong indicator of level of tension produced, the
relatively low values for men in Table 1
indicate that significant resistive loads
would need to be added to the trunk
before the levels of muscular activity
would be consistent with increased muscle strength in other muscles and in
other positions and consistent with the
effort produced by women.
Although we did not assess specific
physiological variables, the data may
demonstrate between-sex differences in
mechanical efficiency of work, that is,
increased EMG yields decreased work
efficiency.1(p99) The between-muscle differences may reflect compensation or
redistribution of the relative contribution each muscle makes to the different
tasks required by the different positions.
In other words, the position may bias
the particular muscle to be more or less
Analysis of Variance Summary: Normalized EMGa for Three Different Muscles According
to Sex and Push-up Positionb
Source of Variation
Triceps brachii
Subject (sex)d
Sex x position
Pectoralis Major
Subject (sex)d
Sex x position
Latissimus Dorsi
Subject (sex)d
Sex x position
Test data were expressed as a percentage of EMG evoked during maximal isometric
Three push-up positions: standard wheelchair position (S); long sit with elbows flexed 90°
(M); long sit with maximum elbow and shoulder abduction (E).
Subject (sex) used as error term.
Includes all male and female subjects.
Downloaded from by guest on August 29, 2014
active. Because women have less muscle
mass than men, greater tension in the
available fibers would be required to
perform the push-ups.10
Depending on treatment objectives,
more optimal or difficult exercise positions may be selected. Examination of
the "total" percent EMG produced
shows that for women, position E clearly
demanded the greatest activity. Men,
however, did not appear to have marked
differences across positions. The exercise positions used in this study were
based on clinical applicability and the
mechanical requirements necessary to
perform the exercises. The differences
demonstrated between positions were to
be expected. None of the muscles, however, demonstrated identical sequences
of increase in activity level. For example, the PM muscle increased in activity
from position S to M to E for women.
For the PM muscle in men, however,
the order was position M followed by
positions S and E. The data did not
explain why these differences between
muscle activity in sexes occurred, but
therapeutic intervention and accomplishment of treatment goals may depend on selecting the most effective exercise for each sex.
Height as a possible factor did not
explain the difference in muscle activity
between sexes because sitting height negated much of the difference in the subject groups. In addition, for the M position, the push-up blocks were adjusted
so that the elbows were at 90 degrees of
flexion. Therefore, height was taken into
account. In the E position, the subjects
performed push-ups from an extreme
elevation and if women had a greater
range of motion, greater EMG may have
been produced. Observation of subjects'
performances, however, would not support this argument for higher EMG levels in position E. Body weight may also
have been an explanation but can hardly
be considered because of the greater
weights of male subjects. According to
Astrand and Rodahl, height and weight
are more important considerations during growth.1(p123)
The results of this study are interesting when assessing the data according to
position and sex. Although significant
interactions occurred for the analysis of
the PM and LD muscles, in every comparison between men and women of
exercise by position, the women produced greater percentages of EMG (Figs.
2-4). This finding is remarkable, considering that electrode locations were
standardized across subjects in terms of
Fig. 2. Triceps brachii muscle activity for female subjects (F) and male subjects (M). S =
standard wheelchair position; M = long sit with elbows flexed 90 degrees; E = long sit with
maximum elbow flexion and shoulder abduction; EMG = percentage of activity evoked during
maximum voluntary contraction.a Significant differences between female subjects and male
subjects, p < .01; b Significant differences between positions, p < .01, for female subjects (F)
and male subjects (M).
Fig. 3. Pectoralis major muscle activity. All position notations are same as for Figure 2.
Significant differences across position are represented for women by a solid line and for men
by a dotted line.
percentage of limb length. Furthermore,
every attempt was made to control body
alignment and the associated mechanics
during the course of the exercise. Minor
variations in the S and E positions were
found, however, because of sitting
height and shoulder range of motion.
Greater glenohumeral abduction and
elbow flexion occurred from position S
to M to E (Fig. 1). The angular changes,
approximately 20 to 30 degrees between
positions, produced some variation in
muscle length but apparently had no
consistent effect on the results. Although
torque requirements were different for
the positions, this effect also demonstrated no consistent pattern either for
men or women (Tab. 1). Because both
Volume 64 / Number 1 , January 1984
Downloaded from by guest on August 29, 2014
not be ordered for level of difficulty
from standard wheelchair push-ups to a
position requiring greater glenohumeral
abduction and elbow flexion. Placing
subjects in a position requiring a combination of maximum shoulder abduction and elbow extension for training
does not necessarily evoke maximum
levels of EMG for the TB, LD, and PM
muscles. Therefore, therapists should be
cognizant of the effect of position and
the patient's sex, as demonstrated in this
study, before selecting the exercise most
likely to yield the best treatment outcome. Future studies are needed to compare the levels of activity produced during therapeutic exercises with functional
activities such as transfers and crutch
Fig. 4. Latissimus dorsi muscle activity. All position notations are same as for Figure 2.
Significant differences across position are represented for women by a solid line and for men
by a dotted line.
sexes completed the exercises from the
same positions, the muscle length and
mechanical requirements cannot explain why the women always produced
greater percentages of EMG.
The muscles studied are those most
responsible for generating simultaneous
shoulder and elbow extension. This
function for these muscles has been well
established in studies by Jonsson et al,5
Broome and Basmajian,11 Shevlin et al,3
Scheving and Pauly,2 and Reeder.4
Shevlin et al3 noted differences in sternal
versus clavicular fibers in the PM muscle, and Reeder4 and Ito et al12 noted
differences in levels of activity between
the lower and upper LD muscle fibers.
Caution must be applied when interpreting the former studies, however, because neither the analysis by Reeder4
nor by Ito et al12 appear to make acrosschannel comparisons that could be regarded as legitimate. The current study
accepted the functions of the muscles
assessed and made no attempt to assess
level of activity on a muscle segment
basis. Others have previously established
that individual muscles are capable of
functioning as separate segments.13. 14
Beam has assessed several muscles in
the shoulder but used no methods that
required the performance of functional
activities.15 The most relevant work
seems to have been performed during
the EMG study by Ito et al on the latissimus dorsi muscle.12 The study compared 26 activities, including long-sit
push-ups and transfer activities, performed by healthy subjects and nine
subjects with T l to L1 spinal cord injuries. As might be expected from the
level of injury, the pattern and level of
activity evoked from the patients was
similar to the healthy subjects. The standard long-sit push-ups or the transfer
activities produced only light to moderate activity. By comparison, the present
study has reported EMG values that
ranged from 50 to 85 percent of maximum when the exercise was done in the
S or E position. We recommend applying additional resistance if treatment
goals are to include muscle strengthening.
Several clinical implications result
from this study. Data clearly show that
for the muscles and exercises assessed,
varied amounts of tension, as measured
by EMG, were required. The level of
exercise difficulty was considered to be
greater when greater EMG was recorded. If therapists are using these exercises for accomplishing increased
muscle strength, greater loads need to
be added to the trunk; men, particularly,
require greater loads on the trunk to
produce maximum muscle activity.
Data also point out the effect of specificity of exercise. Table 1 shows that the
exercise positions used in this study can-
1. Astrand P, Rodahl K: Textbook of Work Physiology. New York, NY, McGraw-Hill, Inc, 1977
2. Scheving LE, Pauly JE: An electromyographic
study of some muscles acting on the upper
extremity of man. Anat Rec 135:239-246,
3. Shevlin M G, Lehmann JF, Lucci JA: Electromyographic study of the function of some muscles crossing the gleno-humeral joint. Arch
Phys Med Rehabil 50:264-270,1969
4. Reeder G: Electromyographic study of the latissimus dorsi muscle. Phys Ther 43:165-172,
5. Jonsson B, Olofsson BM, Steffnox L: Function
of the teres major, latissimus dorsi and pectoralis major muscles. Acta Morphol Neerl Scand
6. Ekholm J, Arboreliu UP, Hillerel L, et al: Shoulder muscle EMG and resisting moment during
diagonal exercise movements resisted by
weight-and-pulley circuit. Scand J Rehabil Med
7. Warwick R, Williams PL (eds): Gray's Anatomy:
36th British Edition. Philadelphia, PA, WB
Saunders Co, 1980, pp 534-542
8. Neter J, Wasserman W: Applied Linear Statistical Models. Homewood, IL, Richard D Irwin
Inc, 1974, p 777
9. Moritani T, DeVries HA: Re-examination of the
relationship between the surface integrated
IEMG and force of isometric contraction. Am J
Phys Med 57:263-277, 1978
10. Costill DL, Daniels W, Evans W, et al: Skeletal
muscle enzymes and fiber composition in male
and female track athletes. J Appl Physiol
11. Broome HL, Basmajian JV: The function of the
teres major muscle: An electromyographic
study. Anat Rec 170:309-310,1971
12. Ito H, Iwasaki T, Yamada M, et al: Electromyography of latissimus dorsi muscle. Rinsho
Rigaka Ryoko (Journal of Japanese Physical
Therapy Association) 3:23-36, 1976 (translated by Paul D. Andrew)
13. Soderberg GL, Dostal WF: Electromyographic
study of three parts of the gluteus medius
muscle during functional activities. Phys Ther
14. Pare E, Stern JT, Schwartz JM: Functional
differentiation within the tensor fasciae latae. J
Bone Joint Surg [Am] 63:1457-1471,1981
15. Beam JG: An electromyographic study of the
trapezius, deltoid, pectoralis major, biceps and
triceps muscle during static loading of the upper limb. Anat Rec 140:103-108,1961
Downloaded from by guest on August 29, 2014
Electromyographic Analysis of Selected Muscles
during Sitting Push-ups: Effects of Position and Sex
Debra S Anderson, Martha F Jackson, Debra S Kropf
and Gary L Soderberg
PHYS THER. 1984; 64:24-28.
Permissions and Reprints
Information for Authors
Downloaded from by guest on August 29, 2014