[ research report ]

RACHEL J. PARK, PT1,2 • HENRY TSAO, PT, PhD2 • ANDREW CLAUS, PT, PhD1
ANDREW G. CRESSWELL, Med Dr1 • PAUL W. HODGES, PT, PhD2

Changes in Regional Activity of the Psoas

Major and Quadratus Lumborum With
Voluntary Trunk and Hip Tasks and
Different Spinal Curvatures in Sitting
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.

D
eeply located trunk muscles with segmental attachments to the discrete fascicles within a single muscle
lumbar vertebrae, such as psoas major (PM) and quadratus may generate torque in opposite direc-
tions. Consequently, anatomically dis-
lumborum (QL), have complex functions at the lumbar spine
crete regions within each muscle may
and are a common target for clinical interventions for low
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

have differential actions on the lumbar

back pain.10,23 Due to their close proximity and relative orientation to spine, which may be an important con-
the instantaneous axes of rotation of the lumbar motion segments, sideration when designing treatments
that target these muscles.
Recent data have shown differential
TTSTUDY DESIGN: Cross-sectional controlled trunk extension than flexion, whereas activity of
activation of discrete regions of PM and
laboratory study. PM-v was greater during hip flexion than trunk
QL in tasks that require the production of
TTOBJECTIVES: To investigate the function of dis- efforts. Activity of QL-p was greater during trunk
extension and lateral flexion, whereas QL-a showed trunk torque in different directions.19 The
crete regions of psoas major (PM) and quadratus
lumborum (QL) with changes in spinal curvature greater activity during lateral flexion. During sitting fascicles of PM arising from the lumbar
Journal of Orthopaedic & Sports Physical Therapy®

and hip position.
TTBACKGROUND: Anatomically discrete regions
of PM and QL may have differential function
on the lumbar spine, based on anatomical and
biomechanical differences in their moment arms
between fascicles within each muscle.
TTMETHODS: Fine-wire electrodes were inserted
with ultrasound guidance into PM fascicles arising
from the transverse process (PM-t) and vertebral
body (PM-v) and anterior (QL-a) and posterior
(QL-p) layers of QL. Recordings were made on 9
healthy participants, who performed 7 tasks with
maximal voluntary efforts and adopted 3 sitting
postures that involved different spinal curvatures
and hip angles.
TTRESULTS: Activity of PM-t was greater during
tasks, PM-t was more active when sitting with a
short lordosis than a flat (less extended) lumbar
spine posture, whereas PM-v was similarly active
in both sitting postures.
TTCONCLUSION: Activity of PM-t was more
affected by changes in position of the lumbar
spine than the hip, whereas PM-v was more
actively involved in the movement of the hip rather
than that of the lumbar spine. Moreover, from
its anatomy, PM-t has a combined potential to
extend/lordose the lumbar spine and flex the hip,
at least in a flexed-hip position. J Orthop Sports
Phys Ther 2013;43(2):74-82. Epub 5 September
2012. doi:10.2519/jospt.2013.4292
TTKEY WORDS: fine-wire electromyography,
lumbar spine, postural control, respiration
transverse process (PM-t) (FIGURE 1A) are
biased toward an extension moment arm,
in contrast to the more anteriorly posi-
tioned fascicles of PM from the vertebral
body (PM-v) (FIGURE 1A), which are biased
toward a flexion moment arm. These ob-
servations are consistent with predictions
based on anatomical and biomechanical
data.5 Studies of generalized activity of
QL3,16 and selective recordings of pos-
terior (from the iliac crest to transverse
processes of L1-L4 [QL-p]) (FIGURE 1B)
and anterior (from the iliac crest to the
12th rib [QL-a]) (FIGURE 1B)19 regions of
QL show activity of QL during trunk lat-
eral flexion. However, QL-p was active to

1
The University of Queensland, NHMRC Centre of Clinical Research Excellence in Spinal Pain, Injury and Health, School of Health and Rehabilitation Sciences, Brisbane,
Queensland, Australia. 2The University of Queensland, Centre for Sensorimotor Neuroscience, School of Human Movement Studies, Brisbane, Queensland, Australia. Paul
Hodges is supported by grant ID 1002190 and ID 401598 from the National Health and Medical Research Council. The protocol used for the current study was approved by
the Institutional Medical Research Ethics Committee at the University of Queensland in Brisbane, Australia. The authors certify that they have no affiliations with or financial
involvement in any organization or entity with a direct financial interest in the subject matter or materials discussed in the manuscript. Address correspondence to Dr Paul W.
Hodges, NHMRC Centre of Clinical Research Excellence in Spinal Pain, Injury and Health, School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane
4072, Queensland, Australia. E-mail: p.hodges@uq.edu.au t Copyright ©2013 Journal of Orthopaedic & Sports Physical Therapy

74 | february 2013 | volume 43 | number 2 | journal of orthopaedic & sports physical therapy

43-02 Park.indd 74 1/16/2013 5:29:35 PM

 a greater percentage of maximum volun-
tary contraction (MVC) than QL-a.19 Fur-
thermore, QL-p may generate a greater
lumbar extensor torque than QL-a, due
to its more posteriorly situated fascicles
relative to the instantaneous axes of rota-
tion of the lumbar segments.21
Although discrete function of the fas-
cicles of PM and QL has been observed
when motion is limited to the upright
trunk (by restraint of the pelvis), inclu-
sion of the hip and changes in spinal cur-
vature are likely to further distinguish the
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.

function of these fascicles. For instance,

the PM can produce a hip flexor mo-
ment, which has been suggested to be
the primary function of this muscle, due
to a presumed limited capacity to exert
substantial torques at the lumbar spine.5
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

The function of PM at the hip would be

predicted to change with hip angle, due
to changes in the moment arm of the
muscle relative to the hip joint, along
with length-dependent changes in force-
producing capacity. QL has also been ar-
gued to exert minor torque at the spine.21
However, electromyographic (EMG)
studies consistently show activation of
Journal of Orthopaedic & Sports Physical Therapy®

QL on both the ipsilateral and contralat-

eral sides of the spine during lateral flex-
ion of the spine relative to the pelvis. 3,16
The mechanical advantage of regions
of the PM and QL will also likely change
as a function of lumbar spine curvature.
Changes from lordosis (extension) to ky-
phosis (flexion) would alter the effective
moment arms of these muscles due to
changes in the location of the instanta-
neous axes of rotation.5 Biomechanical
data5 have been interpreted to suggest
that PM has a small net extensor torque FIGURE 1. Anatomical depictions of the muscle regions of PM and QL muscles investigated in this study. (A)
at the upper 3 lumbar vertebrae and a Regions of PM arising from the transverse processes (PM-t) and the vertebral body/intervertebral disc (PM-v). (B)
Anterior (QL-a) and posterior (QL-p) fascicles of the QL muscle. Abbreviations: PM, psoas major; PM-t, transverse
small net flexor torque on the lower 2
process portion of psoas major; PM-v, vertebral body portion of psoas major; QL, quadratus lumborum; QL-a,
lumbar vertebrae when the lumbar spine anterior layer of quadratus lumborum; QL-p, posterior layer of quadratus lumborum. Adapted with permission from
is in a midrange lordotic posture. As bio- Park et al.19
mechanical and EMG studies13,17,22 have
been interpreted to suggest a vital con- On the basis of the available anatomi- cated PM-v would be more active during
tribution of the PM to control of lumbar cal and biomechanical data,5,12,21 the fol- trunk and hip flexion efforts, whereas the
curvature in sitting, it is important to in- lowing hypotheses regarding the function more posteriorly located PM-t would be
vestigate whether activity of specific re- of PM and QL were made. We hypothe- more active in trunk extension efforts.
gions of PM differs with changes in spinal sized that when the spine is in a midrange However, due to its smaller moment arm
curvatures. lordotic posture, the more anteriorly lo- relative to the hip joint compared to that

journal of orthopaedic & sports physical therapy | volume 43 | number 2 | february 2013 | 75

43-02 Park.indd 75 1/16/2013 5:29:37 PM

 [ research report
of the PM-v, PM-t may have a limited role
at the hip, except in positions of greater
hip flexion. We also predicted that the ac-
tivity of the different components of PM
would differ with changes in sitting pos-
ture. As per previous studies,2,13 we pre-
dicted that both PM-t and PM-v would
likely be minimally active in slumped
sitting postures (kyphosis). We also pre-
dicted that PM-v would be affected by
sitting postures that involved changes in
hip flexion angle, whereas PM-t activity
would increase in sitting postures with
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.
increasing angles of lumbar extension
(ie, lumbar lordotic curve). Unlike the
QL-a, due to its more posteriorly located
fascicles, we hypothesized that QL-p ac-
tivity would be greater in a sitting posture
with a lordotic lumbar curve and anterior
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
rotation of the pelvis.
Further complicating the action of
these muscles, activity of PM and QL may
be modulated with respiration. QL-a may
provide support for contraction of the
diaphragm by bracing the 12th rib.21 Fur-
ther, the superior portion of PM interdig-
itates with the medial arcuate ligament
of the diaphragm and may be influenced
Journal of Orthopaedic & Sports Physical Therapy®
by respiration. We hypothesized that this
potential contribution of PM and QL
to respiration may be affected by subtle
changes in lumbar posture due to chang-
es in mechanical advantage.
Understanding the regional physiol-
ogy of PM and QL with different tasks
and changes in spinal curvature may
have important clinical significance. In
clinical practice, these muscles have tra-
ditionally been considered as a single en-
tity. Although some recent clinical work
has been presented to argue for regional
variation in activity, this is being promot-
ed in the clinical literature without scien-
tific underpinning. Evidence of regional
differences in PM and QL activity would
provide a critical foundation for a clinical
rationale for training methods that must
be studied for their clinical utility before
being advocated as a recommendation for
change in practice.
The aim of this study was to inves-
tigate the activity of discrete regions of
76 | february 2013 | volume 43 | number 2 | journal of orthopaedic & sports physical therapy
43-02 Park.indd 76
PM and QL with changes in spinal cur-
vature and hip position, and to determine
whether activity of these muscle regions
is affected by respiratory phase during
these tasks. To provide a reference for
comparison of the activation of PM and
QL, EMG recordings were also made
from typical trunk flexor (obliquus ex-
ternus [OE] and obliquus internus [OI]
abdominis) and extensor (erector spinae
[ES]) muscles, which have a predict-
able response to changes in posture8 and
respiration.25
METHODS
Participants
N
ine healthy volunteers (7 male,
2 female; mean  SD age, 23  3
years; height, 169  5 cm; weight,
62  8 kg) participated in this study. Par-
ticipants were excluded if they reported
having any circulatory, cardiorespiratory,
orthopaedic (including any pain or dys-
function involving the spine and hip), or
neurological condition; recent or current
pregnancies; or a history of back pain or
surgery, including surgery involving in-
cision of the abdominal wall. All proce-
dures were approved by The University
of Queensland and were conducted in
accordance with the Declaration of Hel-
sinki. Participants were involved in an
earlier study.19
Electromyography
EMG of PM and QL was recorded with
intramuscular fine-wire bipolar elec-
trodes. The electrodes were fabricated
using 2 Teflon-coated, 75-μm, stainless-
steel wires. One mm of the 2 wires was
exposed by scraping off their Teflon. The
wires were threaded into a hypodermic
needle (for the PM, 0.70 × 150 mm; for
the QL, 0.65 × 70 mm) and bent back to
form hooks at 1 and 2 mm from each end.
This staggering was to prevent contact
between the electrodes, though it was
impossible to know the exact interelec-
trode distance during recording, because
this depended on the final stable position
of each wire, which involved minor dif-
]
ferences in the location of each electrode.
A researcher with more than 8 years of
experience with fine-wire electrode inser-
tion placed the sterilized fine-wire elec-
trodes into the right PM-t, PM-v, QL-a,
and QL-p at the L3-4 level with ultra-
sound guidance at 5 to 10 MHz (LOGIQ
9; GE Healthcare, Waukesha, WI), us-
ing an aseptic technique.19 The anatomy
of different regions of each muscle was
confirmed with observation on cadavers
and ultrasound examination on healthy
volunteers. The middle layer of QL was
not investigated in this study, because
its anatomy was found to be highly vari-
able and difficult to delineate from the
other layers. Surface electrodes (Ag/AgCl
discs, 10-mm diameter, 20-mm fixed in-
terelectrode distance; Noraxon USA Inc,
Scottsdale, AZ) were placed in a bipolar
configuration over the muscle belly of the
right ES, 2 cm lateral to the L4 spinous
process7 and over OE and OI/transver-
sus abdominis (TrA), according to stan-
dard landmarks.18 Prior to placement
of the surface electrodes, the skin was
thoroughly cleaned and lightly abraded.
A ground electrode was placed over the
right rib cage.
Recordings from both fine-wire and
surface electrodes were acquired using
a TeleMyo 2400 G2 Telemetry System
(Noraxon), with a common-mode rejec-
tion ratio of less than 100 dB, an input
impedance of greater than 100 mΩ, and a
base gain of 500 dB. Data were amplified
2000 times, band-pass filtered between
10 Hz and 1.5 kHz, using the TeleMyo
2400 G2 Telemetry System, and sampled
at 2 kHz using a Power1401 data-acqui-
sition system with Signal Version 3 soft-
ware (Cambridge Electronic Design Ltd,
Cambridge, UK). The signal quality from
both fine-wire and surface electrodes was
visually checked to ensure that the data
were not clipped and the recordings had
adequate signal-to-noise ratios.
MVC Tasks
A series of isometric MVCs was per-
formed for 3 seconds against manual re-
sistance (FIGURE 2). For PM, 3 tasks were
1/16/2013 5:29:38 PM

Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
FIGURE 2. Seven tasks performed with maximal voluntary efforts against manual resistance (direction of manual
resistance indicated by arrow). Trunk rotation to the left was performed by pushing the arms from left to right and
the knees from right to left and trunk rotation to the right (not illustrated) was performed with the subject in the
same position, but with the resistance applied in the opposite direction at the forearms and knees.
Journal of Orthopaedic & Sports Physical Therapy®
performed in supine: trunk flexion with
bent knees and the lumbar spine in con-
tact with the support, right hip flexion
at 30° with the knee straight, and right
hip flexion at 90° with the knee flexed.
Participants performed isometric lateral
flexion of the trunk in sidelying for QL.
For ES, participants extended their trunk
isometrically in prone. In supine, with
knees bent and shoulder flexed at 90°,
participants rotated their trunk isometri-
cally to the left for OI and to the right for
OE, against manual resistance applied to
the knees and hands. Three repetitions
were completed for each MVC task with
verbal encouragement.
Sitting Posture
EMG signal was also recorded during 3
sitting postures. Participants sat on a flat-
surface stool, with its height adjusted to
each participant’s popliteal crease. A pic-
ture of the target slump (kyphotic curves
journal of orthopaedic & sports physical therapy | volume 43 | number 2 | february 2013 |
43-02 Park.indd 77
at thoracolumbar and lumbar regions),
flat (minimal curve at thoracolumbar
and lumbar regions), and short lordosis
(flat/kyphotic thoracolumbar with a lor-
dotic lumbar curve) postures was shown.
Verbal descriptions and manual guidance
were provided to adopt each posture. For
the short lordosis posture, participants
were taught to sit toward the front of
their ischial tuberosities and perine-
um, with a forward tilt of the sacrum
(FIGURE 3). Nine 20-second trials (3 rep-
etitions of each posture in random order)
were recorded, with a rest of approxi-
mately 60 to 120 seconds between trials
to minimize possible fatigue. Partici-
pants were encouraged to maintain quiet
breathing during trials. The respiratory
phase was recorded using an inductance
plethysmograph (Inductotrace; Ambula-
tory Monitoring, Inc, Ardsley, NY) placed
around the chest to monitor rib cage
movement.
Kinematic data of spinal curvature at
3 regions of the spine were collected using
an optical tracking 3-D motion analysis
system (Vicon; OMG plc, Oxford, UK).
The system included 8 cameras and com-
mercial tracking software (Vicon Nexus;
OMG plc). Data were collected at 100 Hz,
and the system was calibrated to record
from a volume of 5.6 m3, within which
the subject remained during testing. Only
calibrations with average residual errors
of less than 1.0 mm in all cameras were
accepted. Five reflective markers were
adhered with double-sided tape to the
skin over the spinous processes at T1, T5,
T10, L3, and S2.
Data Analysis
EMG and kinematic data were exported
and analyzed using MATLAB Version
7.5 (The MathWorks, Inc, Natick, MA).
EMG data were high-pass filtered at 30
Hz (fourth-order Butterworth) to mini-
mize movement artifacts. Electrocardio-
gram was removed from the EMG data
with a modified turning-point filter.1,7 For
the MVC tasks, root-mean-square EMG
amplitude was calculated over a 3-second
region selected from the greatest activ-
ity for each MVC task and normalized
to the maximal activity recorded during
the MVC tasks. For the sitting task, chest
wall movement data from the inductance
plethysmograph were low-pass filtered at
30 Hz (second-order Butterworth) and
times of peak expiration and inspiration
were identified. The root-mean-square of
the EMG data was calculated 0.5 seconds
prior to peak expiration and inspiration
and averaged over the 3 repetitions. EMG
data were normalized to the greatest ac-
tivity recorded for the MVC tasks. How-
ever, to allow for greater sensitivity for
comparison among postures, EMG data
of each muscle were also normalized to
the peak activity recorded in the 3 sitting
postures. This normalization method
scales the data for each posture to be-
tween 0 and 1, with 1 reflecting the peak
across postures. This method provides
greater sensitivity than MVC normaliza-
tion, as MVCs are inherently variable. Al-
77
1/16/2013 5:29:39 PM
 [ research report ]
position. This angle was referred to as
the “estimated hip angle.”

Statistical Analysis
Statistical analysis was undertaken us-
ing STATISTICA Version 8 (StatSoft,
Inc, Tulsa, OK). For MVC tasks, root-
mean-square EMG of each muscle was
compared between MVC tasks using a
repeated-measures analysis of variance.
For the sitting postures, EMG amplitude
for each muscle (n = 7) was compared
between 3 postures (n = 3) and respi-
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.

ratory phases (n = 2), using a repeated-

measures analysis of variance. Post hoc
testing was undertaken with the Duncan
multiple-range test, and significance was
set at P<.05. Partial eta-squared was cal-
culated to estimate the effect size, with
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

values of 0.01 or greater and less than

0.06 indicating a small effect, values of
0.06 or greater and less than 0.14 a me-
dium effect, and values of 0.14 or greater
a large effect.9

RESULTS
MVC Tasks
Journal of Orthopaedic & Sports Physical Therapy®

W
hen participants (n = 7) per-
formed the MVC tasks (normal-
ized to peak amplitude across
MVC tasks), there was differential activ-
ity of the PM and QL regions (task-by-
FIGURE 3. (A) Participants adopted and maintained 3 different sitting postures: slump, flat, and short lordosis.
muscle interaction: P<.001, partial η2 =
Markers were placed on skin over T1, T5, T10, L3, and S2 for calculation of thoracic, thoracolumbar, lumbar, and 0.40) (FIGURE 4). EMG activity of PM-t
estimated hip angles in the sagittal plane. (B) Spinal angles are shown with mean and 95% confidence interval for was greater during trunk extension than
each of the postures. trunk flexion (post hoc, P<.01) or hip flex-
ion with the hip in 90° of flexion (post
though normalization to MVC amplitude ration of the recording: T1-T5 and T5- hoc, P<.05). PM-t EMG did not differ be-
allows comparison between muscles and T10, T5-T10 and T10-L3, and T10-L3 tween tasks that involved flexion of the
individuals (accounting for issues related and L3-S2 (FIGURE 3).8 Kyphotic angles trunk or hip (at 90° and 30° of hip flex-
to electrodes), it introduces its own varia- were described as positive, and lordotic ion) and was active to approximately 20%
tion to the data.24 Of the 9 participants, angles as negative. In addition, the an- of maximum during these tasks (post hoc,
2 were excluded from analysis of MVC gle between the L3-S2 segment and the all P>.16). Unlike PM-t, there was no dif-
tasks because 2 of the MVC tasks were horizontal reference was calculated to ap- ference in the activity of PM-v between
not completed. proximate the angle of the hip. Assuming trunk extension and flexion (post hoc, P
Kinematic data of spinal position were that the angle of the femur relative to the = .18), but PM-v EMG was greater during
exported and downsampled to 50 Hz. horizontal did not change between pos- hip flexion at 90° than the trunk flexion
Thoracic, thoracolumbar, and lumbar an- tures, changes in this angle would reflect task (post hoc, P<.01).
gles in the sagittal plane were calculated changes in the angle of the femur with Differential activity was also identified
as the angle between segments between respect to the lower lumbar spine, thus between regions of QL. QL-p EMG was
markers, averaged over the 5-second du- providing interpretation regarding hip greater during trunk lateral flexion and

78 | february 2013 | volume 43 | number 2 | journal of orthopaedic & sports physical therapy

43-02 Park.indd 78 1/16/2013 5:29:40 PM

 120 120 120
EMG Amplitude, % MVC

100 100 100

80 80 80
60 60 60
40 40 40
20 20 20
0 0 0
–20 –20 –20
F HipF 30° HipF 90° E R(L) R(R) LF F HipF 30° HipF 90° E R(L) R(R) LF F HipF 30° HipF 90° E R(L) R(R) LF

PM-t PM-v QL-a QL-p ES OE OI

FIGURE 4. EMG signal amplitude (normalized to peak amplitude) across 7 MVC tasks. Data are mean  SD. Abbreviations: E, trunk extension; EMG, electromyography;
ES, erector spinae; F, trunk flexion; HipF, hip flexion; LF, trunk lateral flexion; MVC, maximum voluntary contraction; OE, obliquus externus abdominis; OI, obliquus internus
abdominis; PM, psoas major; PM-t, transverse process portion of psoas major; PM-v, vertebral body portion of psoas major; QL, quadratus lumborum; QL-a, anterior layer of
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.

quadratus lumborum; QL-p, posterior layer of quadratus lumborum; R(L), trunk left rotation; R(R), trunk right rotation.

16
angle of the hip relative to the lower lum-
* bar spine was more flexed in short lor-
14
dosis than the other postures (post hoc,
12 all P<.001) and more extended in the
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

10 slump posture (post hoc, all P<.001). The

EMG Amplitude, % MVC

lumbar angle was more lordotic in short

8
lordosis than the other postures (post
6 †
hoc, all P<.001) and more kyphotic in
4 † the slump posture (post hoc, all P<.001).
† The thoracolumbar angle was more ky-
2
photic in the slump posture (post hoc, all
0
P<.001) and more lordotic in the short
–2 lordosis posture (post hoc, all P<.05)
Journal of Orthopaedic & Sports Physical Therapy®

*
–4 than the other postures. All postures were
PM-t PM-v QL-a QL-p ES OE OI kyphotic at the thoracic angle, but more
kyphotic in slump than the other tasks
Slump Flat Short lordosis (all P<.001).
EMG signal of PM-t was greater when
FIGURE 5. EMG signal amplitude normalized to peak amplitude of MVC across 3 different sitting postures: slump,
sitting in a short lordosis than sitting in a
flat, and short lordosis. Data are mean  SD. *Difference between EMG signal amplitude of specific muscles for
a given posture (P<.05). †Difference between EMG signal amplitude of the given muscle between sitting postures. flat posture with a less extended lumbar
Abbreviations: EMG, electromyography; ES, erector spinae; MVC, maximum voluntary contraction; OE, obliquus spine (interaction of posture by muscle:
externus abdominis; OI, obliquus internus abdominis; PM, psoas major; PM-t, transverse process portion of psoas P<.01, partial η2 = 0.23; post hoc, P<.01)
major; PM-v, vertebral body portion of psoas major; QL, quadratus lumborum; QL-a, anterior layer of quadratus (FIGURE 5). In contrast, PM-v, with its hy-
lumborum; QL-p, posterior layer of quadratus lumborum.
pothesized bias toward lumbar flexion,
had similar EMG signal amplitude in
extension tasks (approximately 80% of tasks (post hoc, all P<.01). Both OE and these 2 postures (post hoc, P = .34). De-
maximum) compared to the other tasks OI/TrA were more active during trunk spite this difference, PM-t and PM-v were
(post hoc, all P<.01), whereas QL-a EMG flexion, lateral flexion, and left (OE) or active to a similar percentage of MVC (ap-
was greater during trunk lateral flexion right (OI/TrA) trunk rotation than dur- proximately 8%; post hoc, P = .62), and
compared to the other tasks (post hoc, ing trunk extension (post hoc, all P<.05). EMG signal amplitude of PM-t and PM-v
all P<.001). The potential differentiation was greater than that of the other muscles
in role of activity of the different regions Changes in Sitting Postures in the short lordosis posture (post hoc,
of PM and QL can be interpreted with All participants successfully adopted the all P<.05). In the flat posture, PM-v and
respect to the predictable responses of 3 target spinal postures. This was con- PM-t (approximately 6%) were active to
ES and OE. ES activation was greater firmed by differences in spinal angles a similar proportion of maximum as OI
during trunk extension (similar to PM-t between postures (interaction of posture (post hoc, all P>.05), but to a greater per-
but unlike PM-v) compared to the other by angle: P<.001, partial η2 = 0.89). The centage of maximum than the other mus-

journal of orthopaedic & sports physical therapy | volume 43 | number 2 | february 2013 | 79

43-02 Park.indd 79 1/16/2013 5:29:41 PM

 [ research report ]
cles (post hoc, all P<.05). In the slump
1.2
posture, all muscles were active to a * *
1.1
similar low proportion of MVC, less than NS

EMG Amplitude (Normalized to Peak RMS EMG)

1.0 NS
approximately 3% (post hoc, all P>.50).
0.9 NS
EMG signal amplitude of QL-a, QL-p, ES,
OE, and OI/TrA did not differ between 0.8

sitting postures (post hoc, P>.13). 0.7

NS
0.6 NS
Effect of Respiration 0.5
When data were normalized to MVC,
0.4
there was no effect of respiration on
0.3
trunk muscle EMG in any sitting pos-
0.2
tures (interaction of respiration by
posture by muscle: P = .075, partial η2
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.

0.1

= 0.15). However, when data were nor- 0.0

PM-t PM-v QL-a QL-p ES OE OI
malized to peak EMG amplitude across
the postures, which provides greater Slump Flat Short lordosis
sensitivity to detect difference as a result
of elimination of variability introduced FIGURE 6. EMG amplitude (normalized to peak amplitude) in different respiratory phases across 3 sitting postures.
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

by potential variation in MVC activity,
there was a significant interaction be-
tween respiration, posture, and muscle
(P<.05, partial η2 = 0.18) (FIGURE 6). This
was explained by differences between
the response of regions of QL to changes
in posture as a function of respiratory
phase. EMG signals of both QL-a and
QL-p were greater in the short lordosis
Journal of Orthopaedic & Sports Physical Therapy®
Blue shapes show EMG during expiration and orange shapes show EMG during inspiration. Data are mean  SD.
The EMG signal amplitudes for each muscle across the 3 sitting postures during either expiration or inspiration
are different (P<.05), except as shown with NS, indicated by dotted lines. Only ES and OE were modulated by
respiration, with greater activity of ES in inspiration and that of OE in expiration, both in the short lordosis sitting
posture. *P<.05. Abbreviations: ES, erector spinae; NS, nonsignificant; OE, obliquus externus abdominis; OI,
obliquus internus abdominis; PM, psoas major; PM-t, transverse process portion of psoas major; PM-v, vertebral
body portion of psoas major; QL, quadratus lumborum; QL-a, anterior layer of quadratus lumborum; QL-p,
posterior layer of quadratus lumborum; RMS, root-mean-square.
changes in the lumbar spine and hip po- On the other hand, greater activity of PM

than flat posture during the inspiratory
phase (post hoc, all P<.05) but not during
expiration (post hoc, all P>.22). Greater
EMG signal amplitude in the flat than in
the slump posture was also observed for
QL-p (post hoc, P<.001) but not for QL-a
(post hoc, P = .27) during inspiration. ES
EMG signal amplitude was greater in
inspiration than in expiration (post hoc,
P<.05), and OE EMG signal amplitude
was greater in expiration than in inspira-
tion (post hoc, P<.001) in the short lor-
dosis posture but not the other postures
(post hoc, all P>.20).
DISCUSSION
sitions were largely consistent with our
predictions. Furthermore, differential ac-
tivity was apparent in the everyday task
of sitting. For ergonomic applications, ac-
tivity of PM-t appears to have a specific
advantage for sitting in a short lordotic
posture, as the fascicles of PM-t have a
combined potential to extend/lordose the
lumbar spine and flex the hip (at least in
a flexed-hip position). These results sug-
gest that functional tasks involve differ-
ential neural control of regions of PM
and QL.
Differential Function of Regions of PM in
MVC and Sitting Tasks
Consistent with our hypothesis, PM-t
could be explained by its poor mechani-
cal advantage and the necessity to acti-
vate to a higher proportion of maximum.
Our data cannot discriminate between
these possibilities. Activity of PM-t did
not change as a function of hip positions;
similar activity was observed during
maximal hip flexion in both 30° and 90°.
Taken together, these data imply that the
activity of PM-t is more affected by posi-
tion of the lumbar spine than that of the
hip.
Greater activity of PM-v during hip
flexion at 90° than during trunk tasks
supports our hypothesis that, unlike PM-
t, PM-v is more active as a hip flexor than
active at the lumbar spine. Consistent

T
he findings of this study sup-
port our hypothesis that recruit-
ment of discrete regions of PM and
QL differs based on their anatomical
location. The differential changes in ac-
tivity of anatomically discrete fascicles of
these muscles between MVC tasks with
EMG was greater in lumbar extension with Basmajian’s4 observation for iliacus,
than maximal trunk and hip flexion ef- which he assumed would also apply to
forts. On the one hand, this suggests a psoas, we observed no difference in PM
role at the lumbar spine and contrasts activity between tasks that involved op-
the prediction of a biomechanical study5 posite directions of hip rotation (a com-
that PM lacks the mechanical advantage ponent of the trunk rotation task with
to generate torque at the lumbar spine. manual resistance applied to the knee

80 | february 2013 | volume 43 | number 2 | journal of orthopaedic & sports physical therapy

43-02 Park.indd 80 1/16/2013 5:29:42 PM

 and outstretched arms). Although we did
not study hip rotation in isolation, trunk
rotation was evaluated by application of
manual resistance to the thigh. As this
would necessitate generation of rotation
torque at the hip and spine, the lack of
difference in activation between oppo-
site directions of hip torque may imply a
role in control of the hip joint rather than
generation of hip torque.
The discrete activity of PM regions
during the maximal tasks was largely in
agreement with differential changes as
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.
a function of changes in lumbar curva-
tures and hip angles in 3 sitting postures.
Previous EMG studies2,13,14 have shown
increased activity of PM in the upright
posture compared to the slump posture.
Our data extend this observation, show-
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
ing greater changes in activity of PM
with subtle changes in spine curvature
than those in other trunk and abdominal
muscles. Both PM-t and PM-v were more
active in short lordosis and flat sitting
postures compared to a slump posture.
However, as hypothesized, only PM-t was
more active in the short lordosis than in
the flat posture. Together with its anato-
Journal of Orthopaedic & Sports Physical Therapy®
my, this finding implies that PM-t may
have the potential to meet a concurrent
demand to extend the lumbar spine and
maintain flexion of the hip in the short
lordosis posture (FIGURE 7). The present
finding is consistent with the previous
argument that the PM functions to main-
tain control of the spine in the upright
posture.17,20,22
Differential Function of Regions of QL
As hypothesized, based on the anatomi-
cal features, QL-a EMG was greater dur-
ing trunk lateral flexion than the other
tasks, whereas QL-p was more active
during trunk extension than the other
tasks, except lateral flexion of the trunk.
This differential activity may explain the
conflicting findings from previous EMG
studies. McGill et al16 showed increased
QL EMG during lifting, which required
trunk extension torque, whereas Anders-
son et al3 showed inactivity of QL during
a similar task. It is plausible that the data
journal of orthopaedic & sports physical therapy | volume 43 | number 2 | february 2013 |
43-02 Park.indd 81
FIGURE 7. Proposed model of dual function of PM-t
at the hip and spine. PM-t has potential to extend
the lumbar spine (increasing the lordosis) and flex
the hip, whereas action of the PM-v would tend
to produce trunk flexion (decreasing the lordosis)
and/or hip flexion. X indicates the approximate
instantaneous axes of rotation of the lumbar motion
segments (data from Bogduk et al5). Abbreviations:
PM, psoas major; PM-t, transverse process portion of
psoas major; PM-v, vertebral body portion of psoas
major.
of McGill et al16 represented activity of
QL-p or ES, whereas data of Andersson et
al3 represented that of QL-a. Unlike the
activity of QL-p during trunk extension
in the present study, previous recordings
with static extension loads of 15% body
weight showed minimal activation.19 A
possible explanation is that recruitment
in this direction may be restricted to
higher efforts. However, similar to the
findings of the previous studies,3,16 our
data showed increased activity of both
QL-a and QL-p during lateral flexion of
the trunk. Contrary to our hypothesis but
in agreement with previous EMG obser-
vations,3,16 activity of QL was not affected
by subtle changes in spine curvature or
pelvic rotations in the sagittal plane with
changes in sitting posture.
Respiration
EMG signal amplitude of the discrete re-
gions of PM and QL did not change in
association with respiration in any of the
3 sitting postures. However, differences
between activity of QL-a and QL-p with
changes in sitting posture were accentu-
ated by respiratory phase; that is, differ-
ence in the EMG of QL regions between
postures was only present during inspira-
tion. This may be explained by greater ac-
tivity of QL during this phase in specific
postures, as a result of changes in breath-
ing dynamics with changes in posture.15
The short lordosis posture is likely to fa-
vor QL, as this posture is associated with
even distribution of motion of the abdo-
men and lower rib cage, which is gener-
ated by diaphragm contraction. Thus, it
could be argued that activation of QL-a
may aid in stabilizing the attachment of
the diaphragm at the 12th rib.6 QL-p is
also likely to assist ES in extending the
spine during inspiration in the short lor-
dosis and flat postures. In contrast, dur-
ing the slump sitting posture, abdominal
compliance is reduced; hence, in the
slump sitting posture there is increased
dependence on thoracic expansion and
less activity of the diaphragm for inspi-
ration than in the short lordosis posture.15
ES and OI/TrA EMG modulated with
respiratory phases only in the short lor-
dosis posture. The contribution of ES and
OI/TrA to respiration may be greater in
short lordosis, due to the mechanical
advantage of muscles in generating re-
spiratory movements and breathing pat-
tern, which involves spinal extension in
inspiration and abdominal displacement
in expiration.
Methodological Issues
Recordings of PM were only made at a
single lumbar level. As the function of
PM has been argued to vary between lev-
els,5 these findings cannot be extrapolat-
ed to muscle fascicles with more caudal
and cranial origins. As with all invasive
studies of this type, the number of par-
ticipants was relatively small. Despite
the small sample size, the effective sizes
of all comparisons were large (partial
η2>0.18).9
Clinical/Ergonomic Implications
The findings of this study have several
implications for clinical practice and er-
gonomics. First, PM-t may be important
to consider in rehabilitation for the con-
trol of sitting postures, due to its specific
advantage in the short lordosis sitting
posture, with the potential to simultane-
ously flex the hip and extend the lumbar
81
1/16/2013 5:29:44 PM
 [ research report
spine. As specific coordination of discrete
regions of PM is required to maintain/
adapt varying lumbar curvature and hip
angles, interventions targeting PM are
likely to require consideration of differ-
ent components of PM. Second, clinical
interventions have been designed to re-
duce activity of psoas in people with back
pain. These include muscle stretching,11
injection of Botulinum toxin A into PM,10
and treatment of trigger points.23 Further
investigation is required to determine
whether such treatments are effective,
Downloaded from www.jospt.org at on October 26, 2023. For personal use only. No other uses without permission.
whether the interventions should target
specific regions of PM, and whether these
interventions affect neuromuscular con-
trol of the spine in sitting.
CONCLUSION
Copyright © 2013 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
O
ur findings show differential
control of discrete regions of PM
and QL during functional tasks.
The unique activity of separate regions of
PM should be considered in future stud-
ies to better understand its contribution
to spinal control. t
Journal of Orthopaedic & Sports Physical Therapy®
KEY POINTS
FINDINGS: The recruitment of anatomical-
ly discrete regions of PM and QL differs
based on their location during tasks that
involve changes in trunk and hip posi-
tions and in different sitting postures.
IMPLICATIONS: Coordination of discrete
regions of PM is required to maintain/
adapt varying lumbar curvature and
hip angles in sitting. Clinical interven-
tions targeting PM are likely to require
consideration of different components
of PM.
CAUTION: The present study recorded PM
EMG at a single lumbar level (L3-4).
Caution is required when extrapolating
these findings to muscle fascicles with
more caudal and cranial origins.
ACKNOWLEDGEMENTS: Paul Hodges is support-
ed by a Senior Principal Research Fellowship
from the National Health and Medical Re-
82 | february 2013 | volume 43 | number 2 | journal of orthopaedic & sports physical therapy
43-02 Park.indd 82
]
search Council of Australia. The authors thank
Leanne Hall, Ryan Stafford, and Wolbert Van
den Hoorn for assistance in data collection.
REFERENCES
1. A minian K, Ruffieux C, Robert P. Filtering by
adaptive sampling (FAS). Med Biol Eng Comput.
1988;26:658-662.
2. Andersson E, Oddsson L, Grundström H, Thor-
stensson A. The role of the psoas and iliacus
muscles for stability and movement of the
lumbar spine, pelvis and hip. Scand J Med Sci
Sports. 1995;5:10-16.
3. Andersson EA, Oddsson LI, Grundström H,
Nilsson J, Thorstensson A. EMG activities of
the quadratus lumborum and erector spinae
muscles during flexion-relaxation and other
motor tasks. Clin Biomech (Bristol, Avon).
1996;11:392-400.
4. Basmajian JV. Electromyography of iliopsoas.
Anat Rec. 1958;132:127-132.
5. Bogduk N, Pearcy M, Hadfield G. Anatomy and
biomechanics of psoas major. Clin Biomech
(Bristol, Avon). 1992;7:109-119. http://dx.doi.
org/10.1016/0268-0033(92)90024-X
6. Boyd W, Blincoe H, Hayner JC. Sequence
of action of the diaphragm and quadratus
lumborum during quiet breathing. Anat Rec.
1965;151:579-581.
7. Cholewicki J, Panjabi MM, Khachatryan A. Stabi-
lizing function of trunk flexor-extensor muscles
around a neutral spine posture. Spine (Phila Pa
1976). 1997;22:2207-2212.
8. Claus AP, Hides JA, Moseley GL, Hodges PW.
Different ways to balance the spine: subtle
changes in sagittal spinal curves affect re-
gional muscle activity. Spine (Phila Pa 1976).
2009;34:E208-E214. http://dx.doi.org/10.1097/
BRS.0b013e3181908ead
9. Cohen J. Statistical Power Analysis for the
Behavioral Sciences. 2nd ed. Hillsdale, NJ: Law-
rence Erlbaum Associates; 1988.
10. De Andres J, Adsuara VM, Palmisani S, Vil-
lanueva V, Lopez-Alarcon MD. A double-blind,
controlled, randomized trial to evaluate the
efficacy of botulinum toxin for the treatment of
lumbar myofascial pain in humans. Reg Anesth
Pain Med. 2010;35:255-260. http://dx.doi.
org/10.1097/AAP.0b013e3181d23241
11. Janda V. Muscles, central nervous motor regula-
tion and back problems. In: Korr M, ed. The Neu-
robiologic Mechanisms in Manipulative Therapy.
New York, NY: Plenum Press; 1978:27-41.
12. Jemmett RS, Macdonald DA, Agur AM. Anatomi-
cal relationships between selected segmental
muscles of the lumbar spine in the context of
multi-planar segmental motion: a preliminary
investigation. Man Ther. 2004;9:203-210. http://
dx.doi.org/10.1016/j.math.2004.07.006
13. J uker D, McGill S, Kropf P, Steffen T. Quantitative
intramuscular myoelectric activity of lumbar
portions of psoas and the abdominal wall during
a wide variety of tasks. Med Sci Sports Exerc.
1998;30:301-310.
14. Keagy RD, Brumlik J, Bergan JL. Direct electro-
myography of the psoas major muscle in man. J
Bone Joint Surg Am. 1966;48:1377-1382.
15. Lee LJ, Chang AT, Coppieters MW, Hodges PW.
Changes in sitting posture induce multipla-
nar changes in chest wall shape and motion
with breathing. Respir Physiol Neurobiol.
2010;170:236-245. http://dx.doi.org/10.1016/j.
resp.2010.01.001
16. McGill S, Juker D, Kropf P. Quantitative intra-
muscular myoelectric activity of quadratus
lumborum during a wide variety of tasks. Clin
Biomech (Bristol, Avon). 1996;11:170-172.
17. Nachemson A. Electromyographic studies on
the vertebral portion of the psoas muscle;
with special reference to its stabilizing func-
tion of the lumbar spine. Acta Orthop Scand.
1966;37:177-190.
18. Ng JK, Kippers V, Richardson CA. Muscle fibre
orientation of abdominal muscles and sug-
gested surface EMG electrode positions. Electro-
myogr Clin Neurophysiol. 1998;38:51-58.
19. Park RJ, Tsao H, Cresswell AG, Hodges PW.
Differential activity of regions of the psoas
major and quadratus lumborum during sub-
maximal isometric trunk efforts. J Orthop Res.
2012;30:311-318. http://dx.doi.org/10.1002/
jor.21499
20. Penning L. Psoas muscle and lumbar spine sta-
bility: a concept uniting existing controversies.
Critical review and hypothesis. Eur Spine J.
2000;9:577-585.
21. Phillips S, Mercer S, Bogduk N. Anatomy and
biomechanics of quadratus lumborum. Proc Inst
Mech Eng H. 2008;222:151-159.
22. Santaguida PL, McGill SM. The psoas major
muscle: a three-dimensional geometric study. J
Biomech. 1995;28:339-345.
23. Travell JG, Simons DG. Myofascial Pain and Dys-
function: The Trigger Point Manual. Baltimore,
MD: Williams & Wilkins; 1983.
24. Urquhart DM, Hodges PW. Differential activity of
regions of transversus abdominis during trunk
rotation. Eur Spine J. 2005;14:393-400. http://
dx.doi.org/10.1007/s00586-004-0799-9
25. Urquhart DM, Hodges PW, Story IH. Postural ac-
tivity of the abdominal muscles varies between
regions of these muscles and between body po-
sitions. Gait Posture. 2005;22:295-301. http://
dx.doi.org/10.1016/j.gaitpost.2004.09.012
@
MORE INFORMATION
WWW.JOSPT.ORG
1/16/2013 5:29:44 PM