ARTICLE IN PRESS

Manual Therapy 11 (2006) 54–60

www.elsevier.com/locate/math

Original article

The response of the transverse abdominis and internal oblique

muscles to different postures
Anne-Marie Ainscough-Potts, Matthew C Morrissey, Duncan Critchley
Division of Applied Biomedical Research, Guy’s, King’s and St. Thomas’ School of Biomedical Sciences, Guy’s Campus,
3.19 Shepherd’s House, London SE1 1UL, UK
Received 21 June 2004; received in revised form 23 February 2005; accepted 10 March 2005

Abstract

The purpose of this study was to consider how the deep abdominal muscles responded to alterations in seated stability.
The thickness of the right transverse abdominis (TrA) and internal oblique (IO) muscles were measured with ultrasound
imaging in 30 healthy human subjects (mean age 27.7, years 22 females) in supine lying, relaxed sitting on a chair with both
feet on the ground, relaxed sitting on a gym ball with both feet on the ground and sitting on a gym ball lifting the left foot
off the floor. Measurements were taken at the end of both inspiration and expiration. The results showed that muscle thickness
expressed as a percentage of the actual muscle thickness in supine lying did not differ between relaxed sitting on a chair and sitting
on a gym ball for either muscle (P ¼ 0:01220:054) where Bonferroni corrected P-value for significance ¼ 0.002. Raising the foot off
the floor produced a significant increase in thickness for TrA and IO, when compared with the other seated postures (Po0:001). It
was also found that both muscles were thicker at the end of expiration (Po0:001) which has also been established by other
authors.
These findings suggest that both deep abdominal muscles respond in the same way to postural changes. It also demonstrates that
these muscles are automatically targeted by significantly decreasing the base of support, but in normal subjects sitting on a gym ball
is not sufficient to increase their activity.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: Abdominal muscles; Ultrasound scanning; Posture

1. Introduction Clinicians attempt to strengthen these muscles by

Treating low back pain is a complex activity and the
British Association of Chartered Physiotherapists in
Manipulation (1996) gave an overview of the options
that may be considered. This group concluded that there
was a place for supervised exercises for patients with
chronic low back pain. Currently, there is interest in
strengthening the deep abdominals, which are thought
to help protect the lower spine and prevent recurrence of
pain (Hides et al., 2001, Richardson et al., 1999).
Corresponding author. Tel.: +44 207 848 6319;
fax: +44 207 848 6325.
offering specific challenges to the body that elicit their
contraction.
Investigations by Hodges and Richardson (1996
and 1998), using fine wire EMG, found that the
transverse abdominis (TrA) became active in normal
subjects prior to limb movement. However, when
back pain was present they found the activity of
TrA was delayed. They concluded that the early
activation of TrA may be a protective mechanism
for the lumbar spine, which is lost when patients
have back pain. Norris (1995a) also suggested that the
active muscular system should be addressed, and high-
lights the need to address the deep abdominals that are

1356-689X/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2005.03.007
 ARTICLE IN PRESS
A.-M. Ainscough-Potts et al. / Manual Therapy 11 (2006) 54–60
stabilizers, as well as the superficial ones that act as
prime movers.
Attempts have been made by clinicians to target the
deep abdominal muscles with exercises such as abdom-
inal hollowing and Richardson and Jull (1995) advo-
cated four-point kneeling as the easiest method of
activating TrA. In designing a rehabilitation programme
for TrA, many physiotherapists have noted that it can
be difficult to elicit a contraction using starting positions
and verbal instructions alone and have been looking for
possible ways of triggering an automatic response.
Many authors have advocated sitting on a large inflated
gym ball as a rehabilitation tool. Lewis and Hawke
(1983) advocated sitting on a gym ball to help re-educate
pelvic rocking whereas Scott et al. (1983) suggested that
this would improve joint position sense, balance and
posture. Carriere (1999) proposed that using the gym
ball might improve proprioception. Norris (1995b)
advocated using a gym ball as part of a spinal
rehabilitation programme to trigger abdominal muscle
activity. Gym balls are frequently employed in clinical
practice with the aim of facilitating deep abdominal
muscle activity; however, there has been little research to
support their use in this way.
Fine wire EMG is an invasive procedure, so there has
been a development towards using real time ultrasound
scanning for measurements of abdominal muscle activ-
ity, where changes in muscle thickness are believed to be
indicative of relative muscle activity levels. Misuri et al.
(1997) used real time ultrasound measurements of the
abdominal muscles, and found an increase in thickness
during expiration, when the muscles were actively
contracting. Bunce et al. (2002) also found that
ultrasound was a reliable tool for measuring
changes in TrA muscle thickness between supine
lying, standing and walking. McMeeken et al. (2004)
compared needle EMG recordings of TrA contraction
with real time ultrasound changes in thickness
and found there was a good correlation at all levels
of activity. Hodges et al. (2003) used surface and
fine wire EMG recording and compared the response
with real time ultrasound thickness changes in the
abdominal muscles. They found that at submaximal
contractions there was a good correlation in TrA and
internal oblique (IO).
In another study using ultrasound, Critchley and
Coutts (2002) found that TrA changes in thickness
during abdominal hollowing in four-point kneeling.
These findings help support the use of their non-invasive
technique to measure abdominal muscle thickness and
estimate relative muscle activity.
The purpose of this study was to evaluate possible
changes in the thickness of the right TrA and IO in
different positions, standardized as a percentage of the
muscle thickness found in supine lying. The postures
investigated were relaxed sitting on a chair, sitting on a
55
gym ball and sitting on the same gym ball raising the left
foot off the ground.
2. Method
2.1. Subjects
Thirty healthy subjects (22 females and 8 males)
ranging in age from 18 to 50 with a mean (standard
deviation (SD)) age of 27.7 (8.6) were recruited from
physiotherapy students and staff from King’s College
London. Their height had a mean (SD) of 1.706 m
(89.9 m) and a body mass mean (SD) of 67.2 Kg/m2
(12.4 Kg/m2). All subjects completed a questionnaire
that recorded their age, gender, height, weight, levels of
physical activity and a brief history of any previous back
problems.
Subjects had no history of low back pain in the past 6
months. Two females had suffered previous back pain.
One had post-natal back pain 12 years ago which
received mobilizations and the second had pain after
lifting a box, which lasted 3 days, and then recovered
spontaneously. Four males had previous problems. One
had right sciatica 12 years ago, and left sciatica 6 years
ago now resolved. Three others mentioned that they had
previous back problems that spontaneously recovered
and required no treatment.
Women who were pregnant or thought they might
have been pregnant at the time were excluded from the
study.
Ethics approval was gained from the Guy’s Hospital
Research Ethics Committee and informed written consent
was received from all subjects prior to data collection.
2.2. Equipment
A portable real time ultrasound scanner (Aloka 55D-
900 ultrasound monitor Aloka Co. Ltd., Tokyo, Japan)
was used to measure muscle thickness of TrA and IO on
the right side of the abdominal wall in B-mode with a
7.5 MHz linear head, and a connected screen that
showed the image. By using the callipers measures of
muscle thickness were obtained in millimetres. The
scanner was calibrated approximately every 7 days. The
calipers readings are compared with a standard distance
through water (Cardiff Test Tool; Diagnostic Sonar
Ltd., Livingstone, Scotland) to ensure reliable and
accurate readings.
2.3. Testing postures
The positions used for the recordings were (a) supine
lying with arms folded out of the way (Fig. 1), (b) sitting
comfortably in a chair 43 cm high (with no arm rests),
supported against a back rest, with arms folded; hands
 ARTICLE IN PRESS
56 A.-M. Ainscough-Potts et al. / Manual Therapy 11 (2006) 54–60
Fig. 1. Supine lying.
Fig. 2. Relaxed sitting on chair, both feet on the ground.
gently resting on the opposite shoulders and the feet
together on the floor (Fig. 2), (c) sitting comfortably
with a straight back on a 65 cm in diameter gymnasium
ball, with arms folded and gently resting as above and
feet together on the floor (Fig. 3) and (d) the same ball
sitting position as (c) but lifting the left foot off the floor
by approximately 10 cm (Fig. 4). When sitting on the
ball, a 70 kg person reduced the ball height to 43 cm as
measured using a tape measure—the same as the seat
height of the chair. Each position was held long enough
for the examiner to have a clear picture of the muscle
Fig. 3. Relaxed sitting on gym ball, both feet on the ground.
thickness at the end of inspiration and expiration,
usually no more than 2 min. The subjects were requested
to adopt each of the experimental positions in a random
order using a random square Latin table.
2.4. Ultrasound measurement of muscle thickness
The thickness of both muscles were measured on the
right side. The abdominal wall was exposed and the
linear head, covered with conducting gel, was placed on
the skin halfway between the anterior superior iliac
spine and the lower rib cage in the anterior axillary line
(Strohl et al., 1981) where the three deep abdominal
layers could be readily identified. Initially, supine lying
was used to familiarize the examiner with the individual
subject’s muscle topography. However, no recording
was made of the muscle thickness (Fig. 1).
The recording of the muscle thickness in supine lying
was taken at the time when that posture was measured
during the randomized presentation of the postures. The
scan site was not marked on the skin but a knowledge of
the local muscle topography was used to return to the
same place as the position of the muscle was expected to
alter between lying and sitting. The curser points
measured the muscle thickness between the fascial bands.
 ARTICLE IN PRESS
A.-M. Ainscough-Potts et al. / Manual Therapy 11 (2006) 54–60
Fig. 4. Sitting on a gym ball, lifting the left foot.
2.5. Analysis
The data was analysed using the SPSS statistical
software version 11.50 (SPSS Inc. 233 South Wacker
Drive, 11th Floor, Chicargo, IL 60606-6307). Statistical
significance was set at P ¼ 0:05. Paired t-tests were used
to detect any differences between inspiration and
expiration for the TrA and IO muscles, as well as
thickness differences between the muscles. t-Tests were
used on the data from both muscles at the end of both
inspiration and expiration to detect differences in
response between the different postures.
Intrarater reliability was tested using the method
described by Rankin and Stokes (1998) using intraclass
correlation coefficients (ICC). Two results of muscle
thickness (in millimetres) from ten subjects, in each of
the postures, during both respiratory states, were used
to calculate the ICCs.
3. Results
For each respiratory state and in each of the postures,
two results for ten subjects gave ICC results between
0.97 and 0.99. The remainder of the all the results
were calculated from the mean of three observations.
57
Possible order effects were evaluated and no statistically
significant effects were noted (range of P-values ¼
0.98–0.99).
Table 1 shows the means, standard deviations and
ranges of the muscle thickness recorded in millimetres
for both muscles at the end of inspiration and expiration
in each of the four experimental positions. These data
demonstrate the wide range of abdominal muscle
thickness that exists in a normal population. Inspiration
was compared with expiration and there was a
significant difference (Po0:001) for both muscles with
the muscles being larger in size during expiration.
Table 2 shows the results of significance tests when the
difference in thickness is expressed as a percentage of the
individual muscle thickness in lying. For both muscles in
both respiratory states there is no statistically significant
difference when relaxed sitting is compared with sitting
on the gym ball (P ¼ 0:01220:054). When either relaxed
sitting or sitting on the gym ball was compared with
sitting on the gym ball and lifting the left leg off the
ground there was a significant increase in muscle
thickness (Po0:001).
Fig. 5 depicts the mean (SD) of the results for each
muscle when the experimental posture was expressed as
a percentage of the thickness of the individual’s muscle
in lying as a way of standardizing the results. The large
standard deviations indicate that some people have a
decrease in thickness between lying and relaxed sitting
and lying and sitting on the gym ball, but not between
lying and sitting on the gym ball and raising the left foot
off the floor.
Fig. 6 uses the standardized results and shows the
means (SD) of comparing the different experimental
positions. A Bonferroni correction of dividing the
significance of 0.05 by the number of postural observa-
tions (24) was carried out on the data in order to avoid
reporting false positive results and setting the signifi-
cance to Po0:002. This was in line with the recommen-
dations of Portney and Watkins (2000). Sitting on
the gym ball was not statistically significant when com-
pared with relaxed sitting in either respiratory state
(P ¼ 0:01220:054). Sitting on the gym ball, and lifting
the contra-lateral foot off the ground, was statistically
significant when compared to sitting, relaxed sitting or
sitting on the gym ball in both respiratory states
(Po0:001) and  indicates a statistical significant
difference in Table 2.
4. Discussion
The results showed that both muscles, in both
respiratory states, acted in the same way. Using the
normalized data relaxed sitting in the chair showed no
significant difference when compared with sitting on the
gym ball but a significant increase in thickness was
 ARTICLE IN PRESS
58 A.-M. Ainscough-Potts et al. / Manual Therapy 11 (2006) 54–60

Table 1
Muscle thickness in the different postures (N ¼ 30)

Position Muscle and respiratory state Mean SD Range min–max

Lying TrA inspiration 3.98 0.91 2.63–7.10

TrA expiration 4.23 0.99 2.7–7.27
IO inspiration 9.05 2.95 5.03–15.93
IO expiration 9.62 3.1 5.10–16.27
Relaxed sitting TrA inspiration 4.43 1.31 2.23–8.03
TrA expiration 4.83 1.46 2.23–8.90
IO inspiration 9.44 2.82 5.47–16.27
IO expiration 10.04 2.98 5.83–17.37
Sitting on the gym ball TrA inspiration 4.77 1.56 1.67–7.40
TrA expiration 5.10 1.70 1.73–8.27
IO inspiration 10.18 3.5 5.67–18.70
IO expiration 10.71 3.58 5.67–19.57
Sitting on the gym ball with the left foot TrA inspiration 6.32 2.01 2.57–10.30
raised approx 10 cm off the floor TrA expiration 6.71 2.07 2.53–10.20
IO inspiration 12.58 3.97 5.93–21.50
IO expiration 13.12 3.97 6.37–21.0

All measurements are in millimetres.

Table 2
t-Test results of the effects of different postures when abdominal muscle thickness is expressed as a percentage of the thickness of the muscle in lying
(P ¼ 0:002)

Transverse abdominis muscle percentage difference

Posture 1 Posture 2 Mean difference (SD) between P-value

postures 1 and 2

Transverse abdominis muscle percentage difference

During inspiration
Sitting on a gym ball Relaxed sitting 8.34 (22.72) 0.054
On the gym foot raised Relaxed sitting 47.95 (35.07) 0.001*
On the gym foot raised Sitting on a gym ball 39.6 (31.23) 0.001*
During expiration
Sitting on a gym ball Relaxed sitting 5.45 (23.87) 0.221
On the gym foot raised Relaxed sitting 44.82 (35.79) 0.001*
On the gym foot raised Sitting on a gym ball 39.37(26.59) 0.001*
Internal oblique muscle percentage difference
During inspiration
Sitting on a gym ball Relaxed sitting 6.96 (14.15) 0.012
On the gym foot raised Relaxed sitting 35.31 (20.72) 0.001*
On the gym foot raised Sitting on a gym ball 28.34 (19.42) 0.001*
During expiration
Sitting on a gym ball Relaxed sitting 6.13 (13.22) 0.017
On the gym foot raised Relaxed sitting 33.5 (21.43) 0.001*
On the gym foot raised Sitting on a gym ball 27.37 (22.67) 0.001*

*Statistically significant difference (using Bonferroni correction of 0.05/24 P ¼ :002).

found when the subjects sat on the gym ball and raised
their left foot.
The results of the raw data were expressed as a
percentage of the recordings taken in supine lying, the
method used previously by Critchley and Coutts (2002)
and McMeeken et al. (2004). This normalized the data,
so t-tests could be used to compare differences between
the means and identify whether significant differences
existed between the postures. These are shown in Fig. 5.
The experimental positions of relaxed sitting and sitting
on the gym ball, A and B, show large standard
deviations which fall below zero in all cases imply that
some subjects decrease their muscle recruitment when
sitting in a chair or just sitting on a gym ball. The results
when they sat on the gym ball and raised the left leg (C)
still show large standard deviations but they do not go
 ARTICLE IN PRESS
A.-M. Ainscough-Potts et al. / Manual Therapy 11 (2006) 54–60
Fig. 5. Posture result means (SD) represented as a percentage of the
thickness in lying (mm).
Fig. 6. Comparisons of the different percentage postures with means
(SD) with  indicating Po0:002.
below zero, which suggests that this posture is most
likely to increase activity in both muscles.
Table 2 and Fig. 6 give the results of the comparisons
of the mean differences between the postures. They
show that there is little difference between relaxed sitting
and just sitting on the gym ball, but when these two
postures are compared with sitting on the gym ball and
raising the left leg, there is a significant difference in
thickness in both muscles demonstrating that there is an
increase in activity.
From observing the data in Table 1, it is clear that
there is a general trend for the muscles to increase in
thickness as stability of the postures gradually decreases,
during both respiratory states. These results reinforce
the proposal by Richardson et al. (1999) that a
progression for improving the activity of IO and TrA
was to decrease the stability of the base of support, and
saw sitting on a gym ball as an ‘advanced posture’
compared with crook lying, due to the inherent
instability of this position.
59
However, when the data from this study was normal-
ized there was no significant difference found between
relaxed sitting and sitting on a gym ball in normal
subjects.
The results of this current study suggest that, in
normal subjects, as postures become less stable, the
inner abdominal muscles TrA and IO respond in the
same way during both respiratory states. Relaxed sitting
in a chair and sitting on a gym ball produced a similar
response from the TrA and IO, and comparisons of the
means show no statistical significance. As the muscles
responded in a similar way, the study demonstrates that
sitting on a gym ball is not sufficient to trigger a reaction
in TrA alone.
When the two seated postures are compared with
sitting and lifting the contra-lateral foot off the floor, the
increase in thickness in both muscles is statistically
significant, and happens without conscious effort. Once
again, both muscles responded so they do not appear to
work independently during these activities. However,
this study does not take into account any feed forward
mechanism that might be occurring in the TrA, which
has been found by Hodges and Richardson (1996 and
1998). In their studies they found that TrA became
active in anticipation of activities of limb movements,
and may act as a brace stabilizer of the low back prior to
activities.
Both the TrA and IO muscles in all the positions were
found to be thicker at the end of expiration, as
compared to inspiration with P ¼ 0:001 This agrees
with the findings of De Troyer et al. (1990) who found
TrA was thicker at the end of expiration and Misuri et
al. (1997) who found an increase in both muscles. So this
current study confirms the contribution that these
muscles make towards the respiratory process.
These interpretations presuppose that an increase in
thickness represents activity in the deep abdominal
muscles. This is the case in low-level (o20% maximum
voluntary contraction) static effort (Hodges et al., 2003)
and for concentric contraction at all levels of effort in
TrA (McMeeken et al., 2004). The most demanding
posture in the present study, lifting a single foot, is likely
to involve low-level static contraction of the deep
abdominal muscles which increases confidence in the
validity of these results.
The present investigation reports changes in muscle
thickness in both deep abdominal muscles at one imag-
ing point only. Recent anatomical studies (Urquhart
et al., 2005) have described morphological and hence
possible functional subdivisions within TrA and OI. The
possibility of such a functional subdivision is supported
by the differences in thickness change between lying and
standing seen in different parts of OI (Beith et al., 2001).
Further investigations of muscle change in thickness in
different portions of the deep abdominal muscles might
clarify their functional role and ultrasound lends itself
 ARTICLE IN PRESS
60 A.-M. Ainscough-Potts et al. / Manual Therapy 11 (2006) 54–60
well to making such multiple measures quickly and
simply.
Clinicians wishing to activate the deep abdominal
muscles need to consider a choice of starting position
that induces instability, but this study suggests that both
TrA and IO respond in the same way to alterations in
postural demands, rather than acting independently.
5. Conclusion
Sitting on a gym ball with the feet supported is not
enough to trigger a significant increase in right-sided
deep abdominal muscle thickness, indicating stronger
contraction, than relaxed supported sitting. In this study
the left foot needed to be raised which indicates a further
decrease in stability was required to produce statistical
significance at a subconscious level. It would be of
interest for further studies to compare the effects of the
training using the gym ball, on the possible change in
size of the muscles.
References
British Association of Chartered Physiotherapists in Manipulation.
Back pain management. The evidence for physiotherapy.
Chartered Society of Physiotherapy 1996:11–7.
Beith ID, Critchley DJ, Copeman E, Newham DJ. Changes in
thickness of the left and right human abdominal muscles in
standing and lying. Journal of Physiology 2001;531:147P.
Bunce SM, Moore AP, Hough AD. M-mode ultrasound: a reliable
measure of transverse abdominis thickness? Clinical Biomechanics
2002;17:315–7.
Carriere B. The ‘Swiss Ball’ physiotherapy 1999;83(10):552–61.
Critchley DJ, Coutts FJ. Abdominal muscle function in chronic low
back pain. Physiotherapy 2002;68(6):322–31.
De Troyer A, Estenne M, Ninane V, Van Gansbeke D, Goirini M.
Transverse abdominus muscle function in humans. Journal of
Applied Physiology 1990;68:1010–6.
Hides JA, Jull GA, Richardson CA. Long term effects of specific
stabilizing exercises for first-episode back pain. Spine 2001;
26(11):243–8.
Hodges PW, Richardson CA. Inefficient muscular stabilization of the
lumbar spine associated with low back pain. Spine 1996;21(22):
2640–50.
Hodges PW, Richardson CA. Delayed postural contraction of
transverse abdominis in low back pain associated with movements
of the lower limb. Journal of Spinal Disorders 1998;11(1):46–56.
Hodges PW, Pengel LHM, Herbert RD, Gandevia SC. Measurement
of muscle contraction with ultrasound imaging. Muscle and Nerve
2003;27:682–92.
Lewis FMK, Hawke JR. Orthopaedic treatments—1. The spine.
Physiotherapy 1983;69(3):76–7.
McMeeken JM, Beith ID, Newham DJ, Milligan P, Critchley DJ. The
relationship between EMG and change in thickness of transverses
abdominis. Clinical Biomechanics 2004;19:337–42.
Misuri G, Colagrande S, Gorini M, Iandelli I, Mancini M, Duranti,
Scano G. In vivo ultrasound assessment of respiratory function of
abdominal muscles in normal subjects. European Respiratory
Journal 1997;10:2861–7.
Norris CM. Spinal stabilisation 3. Stabilisation mechanisms of the
lumbar spine. Physiotherapy 1995a;81(2):72–8.
Norris CM. Spinal stabiliation 5. An exercise programme to enhance
lumbar stabilisation. Physiotherapy 1995b;81(3):138–46.
Portney LG, Watkins MP. Foundations of clinical research. Applica-
tion to practice. 2nd ed. Englwood Cliffs, NJ: Prentice Hall; 2000.
p. 461–6 [chapter 21].
Rankin G, Stokes M. Reliability of assessment tools in rehabilitation:
an illustration of appropriate statistical analysis. Clinical Rehabi-
litation 1998;12:187–99.
Richardson C, Jull G. Muscle control–pain control. What exercises
would you prescribe? Manual Therapy 1995;1:2–10.
Richardson C, Jull G, Hodges P, Hides J. Therapeutic exercise for
spinal segmental stabilization in low back pain. New York, NY:
Churchill Livingstone; 1999. p. 153–4 [chapter 10].
Scott JJ, Pruce SPD, Wilson DJ. Orthopaedic Treatments—2. The
upper and lower limbs. Physiotherapy 1983;69(3):78–9.
Strohl KP, Mead J, Bankett RB, Loring SH, Kosch PC. Regional
differences in abdominal muscle activity during various man-
oeuvres in humans. Journal of Applied Physiology 1981;51:1471–8.
Urquhart DM, Barker PJ, Hodges PW, Story IH, Briggs CA. Regional
morphology of the transversus abdominis and obliuus internus
and externus abdominis muscles. Clinical Biomechanics 2005;
20:233–41.