SPINE Volume 27, Number 11, pp 1238–1244©2002, Lippincott Williams & Wilkins, Inc.

The Effect of Different Standing and Sitting Postures onTrunk Muscle Activity in a Pain-Free Population

Peter B. O’Sullivan, PhD,* Kirsty M. Grahamslaw, M Manip Ther, PT,†Michelle Kendell, M Manip Ther, PT,‡ Shaun C. Lapenskie, M Manip Ther, PT, FCAMT,§Nina E. Moller, M Manip Ther, PT,� and Karen V. Richards, M Manip Ther, PT¶

Study Design. A normative, single-group study wasconducted.

Objective. To determine whether there is a differencein electromyographic activation of specific lumbopelvicmuscles with the adoption of common postures in a pain-free population.

Summary of Background Data. Clinical observationsindicate that adopting passive postures such as swaystanding and slump sitting can exacerbate pain in indi-viduals with low back pain. These individuals oftenpresent with poor activation of the lumbopelvic stabiliz-ing musculature. At this writing, little empirical evidenceexists to document that function of the trunk and lum-bopelvic musculature are related to the adoption of stan-dardized standing and sitting postures.

Methods. This study included 20 healthy adults, withequal representation of the genders. Surface electromyo-graphy was used to measure activity in the superficiallumbar multifidus, internal oblique, rectus abdominis, ex-ternal oblique, and thoracic erector spinae muscles forfour standardized standing and sitting postures.

Results. The internal oblique, superficial lumbar mul-tifidus, and thoracic erector spinae muscles showed asignificant decrease in activity during sway standing (P �0.027, P � 0.002, and P � 0.003, respectively) and slumpsitting (P � 0.007, P � 0.012, and P � 0.003, respectively),as compared with erect postures. Rectus abdominis ac-tivity increased significantly in sway standing, as com-pared with erect standing (P � 0.005).

Conclusions. The findings show that the lumbopelvicstabilizing musculature is active in maintaining optimallyaligned, erect postures, and that these muscles are lessactive during the adoption of passive postures. The resultsof this study lend credence to the practice of postural re-training when facilitation of the lumbopelvic stabilizingmusculature is indicated in the management of specific spi-nal pain conditions. [Key words: abdominal muscles, backmuscles, electromyography, lumbar spine, posture, sacroil-iac joint, stability] Spine 2002;27:1238–1244

Postural retraining has traditionally been an integral partof physiotherapeutic intervention in the treatment of lowback pain (LBP). However, the relation between postureand LBP is largely unknown.11 It has been proposed thatdifferent postures that superficially appear similar maylead to altered muscle activation. This results in the gen-eration of very different forces in the skeletal system.5

Postural training works on the assumption that an opti-mally aligned skeletal system reduces stress in itsstructures.

Commonly adopted relaxed postures often are pas-sive in nature, with a predisposition toward sway stand-ing and slump sitting.15 It is proposed that these posturesrely on the passive lumbopelvic structures for the main-tenance of an upright position against gravity. As a re-sult, the requirement for muscle activity is diminished. Ithas been reported clinically that these passive posturesfrequently exacerbate LBP,27,38 and that the adoptionof these postures is associated commonly with motordysfunction of spine-stabilizing muscles such as thelumbar multifidus muscle and the deep abdominalmusculature.27

Bergmark9 proposed that the muscles controlling thetrunk could be classified into two groups. The first groupincludes muscles attached directly to the lumbar verte-brae that can provide spine segmental stability. The lum-bar multifidus, transversus abdominis, and internaloblique muscles are part of this group. The second groupconsists of large torque-producing muscles with no seg-mental attachment to the lumbar spine. These musclescontrol gross trunk movement and provide general trunkstability. They include the rectus abdominis, externaloblique, and thoracic erector spinae muscles.

It is recognized that cocontraction of spine-stabilizingmuscles, specifically the lumbar multifidus, transversusabdominis, and internal oblique muscles, has the abilityto enhance lumbopelvic stability. The lumbar multifidusmuscle is unique in its ability to enhance lumbar segmen-tal stability14,41 while dynamically stabilizing the sacro-iliac joint through its control of sacral nutation.36 Thetransversus abdominis and internal oblique muscles pro-vide a stabilizing influence on the lumbar spine via thethoracolumbar fascia and control of intraabdominalpressure.13,20,39 The anteroinferior portion of the inter-nal oblique and transversus abdominis muscles is capa-ble of generating compression and hence of increasingthe stability of the sacroiliac joints.32,37

From the *School of Physiotherapy, Curtin University of Technology,Shenton Park, Western Australia, the †Taylor Physiotherapy andSports Injury Clinic, Edinburgh, Scotland, United Kingdom,‡SportsMed Murdoch, Murdoch, Western Australia, §D. Freer andAssociates Physiotherapy, Barrie, Ontario, Canada, the �HaukelandUniversity Hospital, Bergen, Norway, ¶Ruth Jones Physiotherapy,Southampton, Hampshire, United Kingdom.Acknowledgment date: May 15, 2001.First revision date: September 17, 2001.Acceptance date: October 24, 2001.The manuscript submitted does not contain information about medicaldevices or drugs.No funds were received in support of this work. No benefits in anyform have been or will be received from a commercial party relateddirectly or indirectly to the subject of this manuscript.

1238

At this writing, no empirical evidence exists to docu-ment that the function of the stabilizing and torque-producing muscles in the lumbopelvic region is related tothe adoption of standardized standing and sitting pos-tures in subjects with or without LBP. The function of thelumbar spine muscles during standing and sitting hasbeen examined previously using electromyography(EMG), but there has been an overwhelming focus on theerector spinae muscle group to the exclusion of the sta-bilizing musculature.1,15,40 The methods used to analyzeposture in the existing literature often are found to becomplex and difficult to reproduce in a clinical set-ting,17,42 and in other studies, postural standardizationhas not been quantified.35,37,38 The comparability andclinical application of these studies is reduced by the lackof postural standardization.

In light of these limitations, this study was performedto determine whether a relation existed between activa-tion of specific lumbopelvic muscles and the adoption ofcommon standing and sitting postures in subjects with-out LBP. It was hypothesized that the internal obliqueand superficial lumbar multifidus muscles acting as sta-bilizers of the lumbopelvic region would show greateractivation in active postures (erect standing and sitting)than in passive postures (sway standing and slumpsitting).

Methods

Participants. This study recruited 20 adult subjects (10 menand 10 women) from the Perth metropolitan region. Theirmean age was 27.8 years (range, 21–37 years), mean height 174cm (range, 163–189 cm), and mean weight 67.7 kg (range,50–110 kg). Ethical approval and written informed consentwere obtained.

Subjects were excluded if they were pregnant, had a bodymass index greater than 30, had any reports of LBP (requiringmedication or consultation with a health professional) in thepreceding 6 months, had any known spine disorders, or expe-rienced pain in the test postures.

Equipment and Procedures. Markers were placed on theright lateral tip of the acromion, the midpoint of the greatertrochanter, the tip of the lateral malleolus, and the L1 spinousprocess by the same investigator.

The skin was prepared to reduce skin impedance below 5k�. Two Ag/AgCl surface electrodes (3M Health Care Prod-ucts, London, Canada) were placed unilaterally 25 mm apartand parallel to the muscle fibers over the following muscles:rectus abdominis (5 cm inferior to the xiphoid process and 3 cmlateral to midline), external oblique (lower edge of eighth rib),thoracic erector spinae (5 cm lateral to T9 spinous process),internal oblique (medial to the anterior superior iliac spine),26

and superficial lumbar multifidus (2 cm lateral to midline at theL4–L5 interspinous space).33 The internal oblique electrodealso may have picked up activity of the transversus abdominismuscle, which lies directly beneath it. A common earth elec-trode was placed over the iliac crest.

The EMG equipment consisted of a Bortec (Bortec Electron-ics, Calgary, Canada) and a Medelec amplifier (Oxford Instru-ments, Surrey, UK) connected to a Power Macintosh computer

(Apple Computer, Cupertino, CA). The gain setting was 3 K,and the signal was sampled at 1000 Hz. Data were processedand stored on Labview 5.1 Virtual Instruments Software (Na-tional Instruments, Austin, TX). Raw data were demeaned andband-pass filtered at 6 to 400 Hz) using a fourth order zero lagButterworth filter (National Instruments, Austin, TX). The useof two EMG amplifiers had no effect on the outcome analysisbecause the comparison was made only of individual muscleactivation between postural conditions. Electromyographic re-cordings of thoracic erector spinae muscles were acquired for15 subjects because of missing data.

During data collection, the study participants were barefootwith arms relaxed and lightly clasped in front of their body andfeet positioned 20 cm apart. They were instructed to focusstraight ahead at a designated point. An adjustable height treat-ment table was used for all the sitting positions. The hips andknees were flexed to 90°. Using standardized instructions, theparticipants were positioned by the same investigator for all thetrials.

Four standardized postures were examined: erect standing,sway standing, erect sitting, and slump sitting. The ability toreliably position subjects in the test postures was determined ina pilot study using five subjects undergoing two trials. Intra-class correlation values were between 0.86 and 0.94, showingexcellent reliability.

Erect standing was defined as a position in which the mark-ers on the acromion, greater trochanter, and lateral malleoluslined up to form an angle of approximately 180° (Figure 1). Tomove into sway standing, the participants relaxed and let thepelvis translate anteriorly relative to the trunk (Figure 2). Theangle between the markers then was remeasured. In this study,it was deemed necessary for the angle between erect and swaystanding to differ by a minimum of 15°.

Figure 1. Erect standing posture: 177° angle between markers.

1239Effect of Different Postures • O’Sullivan et al

Erect sitting was defined as a position in which the subjecthad a neutral pelvic tilt, neutral lumbar lordosis, and neutralthoracic kyphosis (Figure 3). To move into slump sitting, thetrunk was relaxed into flexion and the pelvis rotated posteri-

orly (Figure 4). The angle between the acromion, L1 spinousprocess, and greater trochanter was measured. In this study itwas deemed necessary for the angle between erect and slumpsitting to differ by a minimum of 20°.

To measure the postures objectively, a digital camera (Intel,Santa Clara, CA) was placed perpendicularly to the participant275 cm from the proximal foot for the standing postures and200 cm for the sitting postures. The same investigator took adigital photograph, which was imported into an IBM 1460ilaptop computer (IBM, New York, NY) running Scion Image(Scion, Fredrick, MD), an image-processing program that mea-sures angles between manually marked positions on a digitalimage.

Surface EMG measurements were amplitude normalized totwo standardized activities designed to elicit a stable submaxi-mal voluntary contraction. This was performed because nor-malization to a maximal voluntary contraction reportedly isunreliable,23 reducing the ability to detect small changes inlevels of motor activity during the performance of posturaltasks.38

The first standardized activity was the crook-lying bilateralleg raise, in which the heels were held 5 cm above the plinth,during which activity of the internal oblique, rectus abdominis,and external oblique muscles was measured for 3 seconds. Thesecond activity was prone-lying bilateral active knee flexion, inwhich the dorsum of the feet was held 5 cm above the plinth,during which the activity of thoracic erector spinae and super-ficial lumbar multifidus muscles was measured for 3 seconds.

The positions were tested in the following order: supinelying (used as a reference position), crook-lying bilateral legraise, prone-lying bilateral active knee flexion, erect standing,sway standing, erect sitting, and slump sitting. After a 15-

Figure 2. Sway standing posture: 151° angle between markers.

Figure 3. Erect sitting: 132° angle between markers.

Figure 4. Slump sitting: 93° angle between markers.

1240 Spine • Volume 27 • Number 11 • 2002

second delay to ensure stable conditions, three 3-second trialswere recorded for each position. A rest period of 1 minute wasgiven between test postures to eliminate fatigue.

The EMG data of the three trials were visually inspected. Toreduce the effects of postural sway, the trial with the least num-ber of spikes was selected for analysis. In the case of threeconsistent traces, the first trace was chosen to avoid bias. Theelectrocardiographic artifact then was extracted and the rootmean square of the EMG data calculated.

Statistical Analysis. A sample of 20 subjects was determinedto be sufficient for detecting an effect of moderate size with analpha level of 0.05 and a power level of 80%. Statistical anal-ysis was performed using SPSS statistical analysis software,Version 10.0 (SPSS, Chicago, IL). Data were analyzed for nor-mality, and parametric analysis was performed.

A paired t test was used to detect a difference in muscleactivation between erect and sway standing for all five muscles.

The test was repeated to detect a difference in muscle activationbetween erect and slump sitting.

Results

The mean and standard deviation of muscle activities(expressed as a percentage of submaximal voluntary con-traction) for erect versus sway standing and erect versusslumped sitting are presented in Figures 5 and 6,respectively.

Standing PosturesThere was a significant decrease in superficial lumbarmultifidus (t[19] � 3.631; P � 0.002), internal oblique(t[19] � 2.400; P � 0.027), and thoracic erector spinae(t[14] � 3.526; P � 0.003) activation as subjects movedfrom erect standing into sway standing. Conversely, therectus abdominis muscle (t[19] � �3.142; P � 0.005)

Figure 5. Error bars indicatestandard deviation. SLM � su-perficial lumbar multifidus; IO �internal oblique; EO � externaloblique; RA � rectus abdominis;TES � thoracic erector spinae

Figure 6. Error bars indicatestandard deviation. SLM � su-perficial lumbar multifidus; IO �internal oblique; EO � externaloblique; RA � rectus abdominis;TES � thoracic erector spinae

1241Effect of Different Postures • O’Sullivan et al

was significantly more active, and the external obliquemuscle (t[19]� �1.490; P � 0.153) showed a trend to-ward increased activity in the sway standing position,but the latter was not statistically significant.

Sitting PosturesThere was a significant decrease in superficial lumbarmultifidus (t[19] � 2.796; P � 0.012), internal oblique(t[19] � 3.0460; P � 0.007), and thoracic erector spinae(t[14] � 3.634; P � 0.003) activation as the participantsmoved from erect sitting into slump sitting. Neither therectus abdominis muscle (t[19] � 0.436, P � 0.668) northe external oblique muscle (t[19] � 0.842; P � 0.410)showed any significant difference in activity between thetwo sitting postures.

Discussion

The results of this study show there is a relation betweenthe activity of specific muscles in the lumbopelvic regionand the adoption of common standing and sitting pos-tures. Specifically, activation of the superficial lumbarmultifidus, internal oblique, and thoracic erector spinaemuscles decreased in passive sitting and standing pos-tures, but increased in erect postures, indicating a pos-tural stabilizing role for these muscles. An early objectiveof this study was to devise a method of standardizingpostural alignment and measurement. The standardizedpostures in this study closely represent commonly ob-served postures and were found to be reliable, quantifi-able, and capable of being reproduced in a clinicalsetting.

Standing PosturesA key finding of this study was the decrease in activationof the superficial lumbar multifidus, internal oblique,and thoracic erector spinae muscles and the concurrentincrease in rectus abdominis activation observed duringsway standing as compared with erect standing. It hasbeen proposed that as the pelvis shifts anteriorly to thethorax when the subject moves from erect to sway stand-ing, the line of gravity moves posteriorly to the lumbarvertebral bodies. The result is extension of the low lum-bar spine and lumbosacral junction.4 Maintenance of anupright position during sway standing then is achievedmainly through activation of the anterior abdominalmusculature.4 This was demonstrated in the currentstudy by the significantly increased activation of the rec-tus abdominis muscle in the sway standing posture. Thedecrease in superficial lumbar multifidus and thoracicerector spinae activity likely resulted from a lumbar ex-tensor moment acting through the trunk in the swaystanding posture, resulting in trunk flexor activation.

Interestingly, an increase in rectus abdominis activitywas associated with a decrease in internal oblique muscleactivity. Individual muscles of the abdominal wall havedifferent functions with regard to supporting uprightpostures, highlighting the fact that these muscles do notact as a homogeneous group. This notion is supported byprevious research reporting regional differences in the

function of the external oblique muscle.22 Perhaps theinternal oblique muscle functions more to support erectpostures through its influence on the lumbar spine via itscontrol of intraabdominal pressure and its attachment tothe thoracolumbar fascia,13,39 and to provide stability tothe sacroiliac joint in weightbearing through the forceclosure mechanism.37

Increased activation of the thoracic erector spinaemuscle during erect standing is in agreement with previ-ous findings,1 reflecting the muscle activity required tomaintain the thoracic spine in a neutral kyphosis againstgravity. It has been previously reported that the “backmuscles” show slight to continuous activity,2,16,40 mod-erate levels of activity,1,40 or no activity16 in erect posi-tions. The inconsistency in these previous findings mayreflect variance in subjects’ natural posture and the lackof postural standardization within the study design.

Sitting PosturesElectromyographic activity of the superficial lumbarmultifidus, internal oblique, and thoracic erector spinaemuscles was significantly lower during slump sitting thanduring erect sitting. A reduction in EMG activity of theerector spinae during slump sitting has been reportedconsistently in the literature, indicating a “flexion relax-ation” response of the back muscles.1–3,15,35 It appearsthat postural muscle activity decreases as the lumbopel-vic region becomes dependent on its passive structures tomaintain the position against gravity at end-range spineflexion.

Snijders et al38 also documented this relaxation phe-nomenon in the internal oblique muscle when comparingsupported upright sitting with crossed leg sitting. It wasproposed that the reduced activity of the internal obliquemuscle observed in the crossed leg posture reflected en-hanced passive system stability in the sacroiliac joints,with less need for the dynamic stability provided by theinternal oblique muscle.38 In the current study, the activ-ity levels of the external oblique and rectus abdominismuscles did not differ significantly between the two sit-ting postures, demonstrating the limited role that thesemuscles play in lumbopelvic stability under low loadconditions such as sitting.

Clinical RelevanceThe current findings clearly show a link between activityof the lumbopelvic postural stabilizing muscles and themaintenance of erect upright postures. These findingssuggest a close relation between the adoption of passivepostures and reduction in activity of the lumbo-pelvicstabilizing muscles. The authors speculate that the neuralcontrol system may adjust levels of motor activity in thelumbopelvic stabilizing muscles, depending on posture anddegree of load sharing with the passive lumbopelvic struc-tures. This speculation warrants further investigation.

The authors also speculate that individuals who ha-bitually adopt passive postures deactivate and poten-tially decondition the stabilizing muscles of the lum-bopelvic region. A decrease in trunk muscle efficiency has

1242 Spine • Volume 27 • Number 11 • 2002

been shown to increase load on the lumbar discs andligaments.18 This may leave the lumbopelvic region vul-nerable to strain, instability, or injury.12 A positive cor-relation has been shown between the presence of LBPand time spent watching television in large studies con-ducted with adolescents.6–8 The authors of these studiesproposed that this association might be the result of pro-longed sitting as well as poor posture or physical inac-tivity. Clearly, further research is required to clarify therelation between sustained passive postures, muscle ac-tivity, and back pain.

Cholewicki and McGill12 reported that lumbar mul-tifidus activity comprising no more than 3% of maximalvoluntary contraction was sufficient to ensure segmentalstability of the lumbar spine, with lower levels poten-tially compromising this stability. In the current study,the change in superficial lumbar multifidus activity ob-served between erect and passive postures was slightlyless than 35% of submaximal voluntary contraction. Al-though a direct comparison cannot be made between thetwo studies, the authors speculate that the changes inmuscle activity observed between erect and passive pos-tures in the current study reflect a substantial change inmotor activity that may have clinical significance.

The relation of posture to trunk muscle activity inupright postures, and LBP is not well established.10,25,30

One reason for this may be previous lack of posturalstandardization and classification of back pain condi-tions in the literature. It is recognized that postural sta-bilizing muscles such as the lumbar multifidus, internaloblique, and transversus abdominis muscles have an im-portant stabilizing influence on the lumbopelvic re-gion,32,39,41 reducing stress on the passive structures.18

Specific motor dysfunction of these muscles in the pres-ence of chronic LBP also is well documented.19,21,29,31

Further investigation is required to determine whether arelation exists between an individual’s habitual adoptionof passive postures and dysfunction of these muscles.With LBP, in which the passive structures of the lum-bopelvic region already are sensitized, the authors spec-ulate that adopting passive postures may result in exac-erbation of pain. Adopting erect postures, whichfacilitate key lumbopelvic stabilizing muscles, may resultin more effective load sharing with the active system,reducing focal end range stress on the sensitized passivestructures. Conversely, if a pain disorder is related toincreased extensor muscle activity, adopting more pas-sive postures may relieve pain. These hypotheses are thefocus of ongoing research.

The results of this study also may have implicationsfor motor retraining among individuals with specificlumbopelvic pain syndromes. Historically, considerableemphasis has been placed on muscle retraining in non-weightbearing or nonfunctional positions.24,34 Thisstudy showed that erect postural alignment in weight-bearing positions facilitates the stabilizing musculatureof the lumbopelvic region. In contrast, facilitation ofmuscle activity in nonweightbearing or poorly aligned

positions may hinder the transfer of improved lumbopel-vic muscle function into everyday activities. This sup-ports specific retraining of the stabilizing musculature,with concurrent postural reeducation, in the treatment ofspecific LBP conditions.27–29

In the current study, commonly expressed posturesobserved in clinical practice were studied to verifywhether postural change leads to different patterns ofmotor activity. The findings support the conclusion thatthe lumbopelvic stabilizing musculature is active inmaintaining erect postures, and that these muscles areless active during the adoption of passive postures.

Key Points

● The superficial thoracic erector spinae muscleand the anteroinferior portion of the internaloblique muscle are preferentially activated in erectstanding and sitting postures and less active in pas-sive trunk postures.● The neural control system appears to adjust thelevels of motor activity in the lumbopelvic stabiliz-ing muscles, depending on trunk posture.● The findings support the conclusion that adopt-ing passive postures during standing and sitting hasan inhibitory affect on the key stabilizing muscles inthe lumbopelvic region.● The findings support the conclusion that trainingof erect standing and sitting postures has a facilita-tory affect on the key stabilizing muscles in the lum-bopelvic region.

Acknowledgments

The authors acknowledge with appreciation the valuablecontributions of Dr. Marie Blackmore (statistician),Geoff Strauss (senior lecturer), Dr. Kathy Henderson (se-nior lecturer), and Paul Davey (research assistant) of theCurtin University of Technology.

References

1. Andersson B, Jonsson B, Ortengren R. Myoelectric activity in individuallumbar erector spinae muscles in sitting: A study with surface and wireelectrodes. Scand J Rehabil Med Suppl 1974;3:91–108.

2. Andersson B, Ortengren R, Nachemson AL, et al. The sitting posture: Anelectromyographic and discometric study. Orthop Clin North Am 1975;6:105–20.

3. Andersson E, Oddsson L, Grundstrom H, et al. EMG activities of the quad-ratus lumborum and erector spinae muscles during flexion–relaxation andother motor tasks. Clin Biomech 1996;11:392–400.

4. Asmussen E, Klausen K. Form and function of the erect human spine. ClinOrthop 1962;25:55–63.

5. Aspden R. Review of the functional anatomy of the spinal ligaments and thelumbar erector spinae muscles. Clin Anat 1992;372–87:

6. Balague F, Dutoit G, Waldburger M. Low back pain in school children: Anepidemiological study. Scand J Rehabil Med 1988;20:175–9.

7. Balague F, Nordin M, Skovron M, et al. Nonspecific low back pain amongschool children: A field survey with analysis of some associated factors.J Spinal Disord 1994;7:374–9.

8. Balague F, Troussier B, Salminen J. Nonspecific low back pain in childrenand adolescents: Risk factors. Eur Spine J 1999;8:429–38.

9. Bergmark A. Stability of the lumbar spine: A study in mechanical engineer-ing. Acta Orthop Scand 1989;230:20–4.

1243Effect of Different Postures • O’Sullivan et al

10. Bullock-Saxton J. Normal and abnormal postures in the sagittal plane andtheir relationship to low back pain. Physiother Pract 1988;4:94–101.

11. Bullock-Saxton J. Postural alignment in standing: A repeatability study. AustJ Physiother 1993;39:25–9.

12. Cholewicki J, McGill S. Mechanical stability of the in vivo lumbar spine:Implications for injury and chronic low back pain. Clin Biomech 1996;11:1–15.

13. Cresswell A, Grundstrom H, Thorstensson A. Observations on intraabdomi-nal pressure and patterns of abdominal intramuscular activity in man. ActaPhysiol Scand 1992;144:409–18.

14. Crisco J, Panjabi M. The intersegmental and multisegmental muscles of thelumbar spine. Spine 1991;16:793–9.

15. Dolan P, Adams M, Hutton W. Commonly adopted postures and their effecton the lumbar spine. Spine 1988;13:197–200.

16. Donisch E, Basmajian J. Electromyography in the deep back muscles in man.Am J Anat 1972;133:25.

17. During J, Goudfrooij H, Keessen W, et al. Toward standards for posture:Postural characteristics of the lower back system in normal and pathologicalconditions. Spine 1985;10:83–7.

18. Goel V, Kong W, Han J, et al. A combined finite-element and optimizationinvestigation of lumbar spine mechanics with and without muscles. Spine1993;18:1531–41.

19. Hides J, Stokes M, Saide M, et al. Evidence of lumbar multifidus musclewasting ipsilateral to symptoms in patients with acute/subacute low backpain. Spine 1994;19:165–72.

20. Hodges PW. Is there a role for transversus abdominis in lumbopelvic stabil-ity? Manual Ther 1999;4:74–86.

21. Hodges PW, Richardson CA. Inefficient muscular stabilization of the lumbarspine associated with low back pain: A motor control evaluation of trans-versus abdominis. Spine 1996;21:2640–50.

22. Landin Z, Murthy K, DeLuca C. Mechanical recruitment of low back mus-cles: Theoretical predictions and experimental validation. Spine 1988;14:927–38.

23. McGill S. Electromyographic activity of the abdominal and low back mus-culature during the generation of isometric and dynamic axial trunk torque:Implications for lumbar mechanics. J Orthop Res 1991;9:91–103.

24. McGill S. Low back stability: From description to issues for performance andrehabilitation. Exerc Sports Sci Rev 2000;29:26–31.

25. Mulhearn S, George K. Abdominal muscle endurance and its associationwith posture and low back pain: An initial investigation in male and femaleelite gymnasts. Physiotherapy 1999;85:210–16.

26. Ng JK, Richardson CA. Reliability of electromyographic power spectralanalysis of back muscle endurance in healthy subjects. Arch Phys Med Re-habil 1996;77:259–64.

27. O’Sullivan P. Lumbar segmental “instability”: Clinical presentation and spe-cific stabilizing exercise management. Manual Ther 2000;5:2–12.

28. O’Sullivan P, Twomey L, Allison G. Dynamic stabilisation of the lumbarspine. Crit Rev Phys Rehabil Med 1997;9:315–30.

29. O’Sullivan PB, Twomey L, Allison GT. Altered abdominal muscle recruit-ment in patients with chronic back pain following a specific exercise inter-vention. J Orthop Sports Phys Ther 1998;27:114–24.

30. Raine S. Variations of a Series of Physical Characteristics Related to the

Comfortable Erect Standing Posture and How These Are Affected by Age,Gender, Back Pain, and Physical Activity. Curtin University of Technology,Perth, Australia, 1995.

31. Richardson C, Jull G. Muscle control–pain control. What exercises wouldyou prescribe? Manual Ther 1995;1:2–10.

32. Richardson C, Snijders C, Hides J, et al. The Relationship Between theTransversely Orientated Abdominal Muscles, Sacroiliac Joint Mechanics,and Low Back Pain. Proceedings of the seventh scientific conference of theIFOMT, Perth, Australia, 2000.

33. Roy S, Deluca C, Casavant D. Lumbar muscle fatique and chronic low backpain. Spine 1989;22:992–1001.

34. Sahrmann S, Woolsey N, McDonnell M. Level III Course Notes: Diagnosisand Treatment of Muscle Imbalances and Associated Regional Pain syn-dromes. Washington University School of Medicine, Program in PhysicalTherapy, St. Louis, 1998;

35. Snijders C, Bakker M, Vleeming A, et al. Oblique abdominal muscle activityin standing and in sitting on hard and soft seats. Clin Biomech 1995;10:73–8.

36. Snijders C, Vleeming A, Stoeckart R. Transfer of lumbosacral load to iliacbones and legs: Part I. Biomechanics of self-bracing of the sacroiliac jointsand its significance for treatment and exercise. Clin Biomech 1993;8:285–94.

37. Snijders CJ, Ribbers MT, de Bakker HV, et al. EMG recordings of abdominaland back muscles in various standing postures: Validation of a biomechanicalmodel on sacroiliac joint stability. J Electromyogr Kinesiol 1998;8:205–14.

38. Snijders CJ, Slagter AHE, van Strik R, et al. Why leg crossing? The influenceof common postures on abdominal muscle activity. Spine 1995;20:1989–93.

39. Tesh K, Dunn J, Evans J. The abdominal muscles and vertebral stability.Spine 1987;12:501–8.

40. Valencia F, Munro R. An electromyographical study of the lumbar multifi-dus in man. Electromyogr Clin Neurophysiol 1985;25:205–21.

41. Wilke H, Wolfe S, Claes L, et al. Stability increase of the lumbar spine withdifferent muscle groups: A biomechanical in vitro study. Spine 1995;20:192–8.

42. Wrigley TV, Green RA, Briggs CA. Microcomputer video image processingtechnology in working posture analysis: Application to seated posture ofkeyboard operators. Appl Ergon 1991;22:2–8.

Address reprint requests to

Peter O’Sullivan, PhDSchool of Physiotherapy

Curtin University of TechnologySelby Street

Shenton Park, WA, 6008Australia

E-mail: P.OSullivan@curtin.edu.au

1244 Spine • Volume 27 • Number 11 • 2002