Journal Logo


Specificity of a Back Muscle Exercise Machine in Healthy and Low Back Pain Subjects


Author Information
Medicine & Science in Sports & Exercise: March 2010 - Volume 42 - Issue 3 - p 592-599
doi: 10.1249/MSS.0b013e3181b96029
  • Free


Excessive fatigability of lumbar paraspinal muscles is a predictor of a first episode of low back pain (LBP) (3) and also is a predictor of long-term back-related disability (13). In patients with chronic LBP (CLBP), poor back muscle endurance could be attributed to a higher proportion of type II fatigable fibers (26) and to the atrophy of lumbar muscles (17). It is hypothesized that back muscle fatigue would increase the risk of neuromuscular errors that would cause brief uncontrolled intervertebral movements and subsequent tissue strain injury, as proposed by the spinal stability hypothesis (15). As concluded in a recent review, progressive resistance exercises targeting the back muscles have been successful for increasing strength and/or endurance as well as for decreasing pain and/or disability among patients with LBP (27).

Progressive resistance exercises for back muscles can be categorized by the equipment used (27): 1) machines, 2) benches and roman chair, 3) free weights, and 4) floor and stability balls. Although each subcategory has advantages and disadvantages, only machines, although relatively expensive, provide the load ranges and progression necessary for patients (27). However, hip extensors are synergistically recruited during all types of back extension exercises, which could hamper the specificity of these exercises to isolate, overload, and train the muscles of interest, namely, the back muscles. In line with this reasoning, hip extensors have been shown, using surface EMG, to fatigue faster than back muscles in some studies investigating prone back extension exercises (7,19,28,30). However, although back muscle fatigue has been repetitively observed in exercise machines (9,21,24,33), only two studies contrasted EMG indices of fatigue from back and hip extensors (9,21). Both studies showed lower fatigue of gluteus maximus relative to back muscles in healthy controls (9,21) but only one investigated the biceps femoris (9), showing no significant difference with back muscles. Interestingly, patients with CLBP showed higher gluteus maximus fatigue than healthy controls so that no muscle fatigue difference was obtained between this muscle group and the back muscles (21). The authors also suggest that gluteus maximus was the limiting factor in this exercise that was performed until exhaustion, which points the possible lack of specificity of the machine used. However, these results apply for a machine providing at least some pelvis stabilization (cushion on the lower lumbar vertebrae in Kankaanpaa et al. 21 and posterior to the pelvis in da Silva et al. 9) and with the knees at 90°, which apparently does not represent an optimal exercise modality in terms of specificity (10).

In a previous study, we have tested various combinations of pelvic stabilization (three levels: no stabilization, posterior stabilization, and anterior + posterior stabilization) and lower-limb configurations [two levels of hip extensors length: knees at 90° (sitting) or 135° (semisitting)] to determine which one would maximize the specificity of back extension exercises in a machine (10). These exercise modalities represented machines promoted on the market or used in research settings. Posterior pelvic stabilization combined with knees at 135° (semisitting position) allowed the increase of the activation of back muscles and the decrease of the activation of hip extensors, respectively. However, to minimize the confounding effect of muscle fatigue on the EMG activation estimates, only five cycles were performed in each experimental condition, thus preventing an EMG analysis of muscle fatigue. Also, only healthy subjects were tested, thus precluding to generalize these findings to patients with CLBP. Patients with CLBP are known to produce various abnormal trunk muscles activation patterns (39). Although these exercise modalities impose a highly standardized motion pattern and thus decreased the likelihood of producing abnormal EMG patterns, the generalization of findings to patients with CLBP is still mandatory.

The main purpose of the present study was to examine whether a dynamic back muscle endurance exercise performed in a semisitting position induces more fatigue in back muscles than that in hip extensors in healthy subjects as well as in patients with CLBP. We hypothesized 1) that more fatigue would be induced in back muscles relative to hip extensors and 2) that these results would be applicable as much in patients with CLBP as in healthy controls.



Eighteen healthy volunteers (8 men) and 18 patients with nonspecific CLBP (9 men) aged between 20 and 55 yr were recruited (from October 2006 to May 2007) on a voluntary basis. This is a sample of convenience, and the 20- to 55-yr range was selected to target subjects potentially at work (a requirement of the funding agency). The basic demographic and anthropometric characteristics were comparable between these groups (Table 1). Healthy controls were recruited from words of mouth and newspaper ads in Montreal (Quebec), whereas patients with CLBP were recruited through the Montreal Rehabilitation Institute (referral from one of the authors, RV) and newspaper ads. Considering that exercises are considered safe in subjects with nonspecific CLBP (32), relatively broad inclusion and exclusion criteria were used. Consequently, no pain history (mechanism, onset, and duration), no investigation of comorbidities, and no systematic physical examination were used to select or to further describe the patients. Inclusion criteria for the patients with CLBP were as follows: lumbar or lumbosacral pain with or without proximal radicular pain (limited distally to the knees) and presence of chronic pain defined as a daily or almost daily pain for at least 3 months. One specific exclusion criterion was applied to healthy controls: back pain exceeding 1 wk in the preceding year. General exclusion criteria were as follows: surgery of the pelvis or spinal column, known congenital malformation of the spine or scoliosis, body mass index >30 kg·m−2, systemic-neurological-degenerative disease, one positive response to the Physical Activity Readiness Questionnaire, and pregnancy. The preenrollment screening and the laboratory assessments were carried out by the same research assistant, whereas the data analysis and interpretation were carried out by the authors. The subjects were informed about the experimental protocol and the potential risks of the study and gave written consent before their participation. The protocol and the consent form had been previously approved by the ethics committee of the Centre for Interdisciplinary Research in Rehabilitation of Greater Montreal (CRIR).

Demographic, physical, and clinical characteristics of the subjects.

The sample size was estimated using the best data available, namely, from a study (20) showing the effect of endurance exercises on an EMG fatigue index comparable with what has been used in the present study (using a similar dynamic fatigue test), namely, NIMFslp (details below). The before and after results were simply considered as independent (from two groups) to comply with the present statistical analyses (between-group factor, as further outlined below). Using the means (−21.5 and -13.9 Hz·min−1) and a similar standard deviation (7.2 Hz·min−1), a 1.05 effect size was calculated. Using a statistical power table corresponding to t-tests (bilateral testing) and considering an effect size of 1.00 (a little more conservative) and a power of 0.80, 17 subjects were required in each group. Eighteen subjects were assessed in each group but two healthy subjects were rejected because of the poor quality of their EMG signals.


The basic steps of the assessment were as follows: 1) questionnaires and basic anthropometric measures (details below); 2) surface electrode positioning; 3) submaximal contractions in the Biodex for warming-up and to get familiarized with the dynamic task [range of motion (ROM) and pace] using the metronome; 4) three isometric maximal voluntary contractions (MVC), separated by a minimum of 2-min rest (actually ranged between 2 and 3 min); 5) 10-min rest during which the subjective rating of muscle fatigue (details below) was explained to the subject; 6) dynamic back endurance exercise performed until exhaustion; and 7) subjective assessment of muscle fatigue.

Questionnaires and basic anthropometric measures.

To describe the subjects, in addition to basic anthropometric measurements, including four skinfold measures to estimate the percentage of body fat (12), we used five self-administered questionnaires. First, the physical activity level was assessed in all subjects (1). Then, four clinical and pain-related psychological variables were assessed in CLBP subjects: 1) pain intensity on the day of testing with a visual analog scale, 2) perception of disability with the Rolland-Morris Disability Questionnaire (34), 3) fear of movement or injury with the Tampa Scale of Kinesiophobia (14), and 4) pain catastrophizing with the Pain Catastrophizing Scale (38).

Surface EMG.

EMG signals were collected from 14 preamplified (gain = 1000) active surface electrodes (Model DE-2.3; Delsys Inc., Wellesley, MA) consisting of two silver bars (10 mm long, 1 mm wide) spaced 10 mm apart. EMG signals from the recording sites were band-pass filtered between 20 and 450 Hz, analog-to-digital converted at a sampling rate of 2048 Hz, and stored on a personal computer hard disk for later analyses.

After the skin at the electrode sites was shaved and abraded with alcohol, the electrodes were positioned bilaterally on the multifidus at the L4 level (MU-L4-left and MU-L4-right), on the iliocostalis lumborum at the L3 level (IL-L3-L and IL-L3-R), and on the longissimus at L1 (LO-L1-L and LO-L1-R) and T10 (LO-T10L and LO-T10-R) considering their muscle fiber directions (23). We acknowledge the difficulty of capturing the multifidus with surface electrodes and therefore assigned validity of the EMG signal to the landmark location rather than to the multifidus muscle itself. Six additional electrodes were positioned over the belly of the gluteus maximus (GM-L and GM-R), the biceps femoris (BF-L and BF-R) (19), and the vastus medialis (VM-L and VM-R) (4). VM was also considered because the semisitting position requires the participation of the quadriceps to stabilize the body during the exercise. To avoid movement artifacts related to direct contact of electrodes on the seat, holes were pierced into a cushion (thickness = 2.5 cm), which covered the whole seat, so as to allow GM and BF electrodes to be free from contact. A reference silver-silver chloride electrode was positioned over the T8 spinous process.

Tasks and exercise machine.

The three isometric MVC and the dynamic back endurance exercise were performed in a specialized dynamometer (Biodex Dynamometer Medical Systems III, Inc., Shirley, NY), combined with a chair designed for trunk muscle exercises, because it can simulate machines having a pin-loaded weight stack as available on the market in the isotonic mode. For instance, it allows posterior stabilization of the pelvis, a slightly forward tilting of the seat (15°), the positioning of lower limbs with knees at 135°, and the feet pressing against a footboard, which make the femurs to be driven toward the pelvis and thus securing the pelvis against the sacral pad (Fig. 1). However, instead of using the sacral pad or the pelvic-femur strap from the original system, a custom-designed sacral pad that is more like the ones on the market machines [relatively rigid foam (depth = 5 cm; height = 10 cm; width = 33 cm) glued on a plywood] was positioned on the sacrum to prevent the posterior motion of the pelvis. This pad was adjusted both horizontally (to generate a straight lumbar spine) and vertically (upper edge just covering the posterior superior iliac spines) so that the arm axis of the machine could be aligned with the L5/S1 joint. In all tasks, the thighs were firmly stabilized against the seat with a belt available with the system, the resistance scapular pad was positioned in the middle of the scapulas, and the hands were crossed on the opposite shoulders.

perimental setup showing the subject's position and stabilization in the Biodex dynamometer. A custom device designed to stabilize the pelvis anteriorly is also shown but was used only during MVC (not during the back endurance exercise). Please note that this picture was obtained during pilot testing, explaining why no EMG electrodes are shown. This picture has also been modified (elements eliminated in the background) to improve its clarity.

Three isometric MVC were performed with the trunk flexed at an angle of 5° in relation to vertical, which represents the middle of the ROM covered during the dynamic exercise. All MVC were performed progressively (3 s to reach the maximal, 1 s to maintain and relax). Only for MVC (not the dynamic endurance exercise), a custom-designed stabilization device (two adjusted pads mounted on a metallic armature) was positioned against the anterior superior iliac spines to prevent the anterior motion of the pelvis (Fig. 1). To maximize performance at each contraction, the measured L5/S1 extension moment was displayed in real time as visual feedback (on a monitor), and standardized verbal encouragements were given. Back strength (strength) was defined as the peak L5/S1 extension moment among the three MVC.

The dynamic back endurance exercise consisted of extending the trunk from a forward flexion posture of 25° to a −15° extension posture (dynamometer settings), for a 40° total ROM. Each flexion-extension cycle lasted 4 s (2 s of flexion and 2 s of extension) and was paced with a metronome (60 beats·min−1). The subjects generated efforts constantly during the concentric (extension) and eccentric (flexion) phases of each cycle using the isotonic mode of the system. The load corresponded to 60% of strength, as used in previous training studies in patients with CLBP (11,37). This should approximately lead to 15-25 repetitions maximal, but a limit of 60 repetitions was fixed, which was reached in only one case (healthy woman). Exhaustion was defined as the inability to reach the target angles in due time for three consecutive cycles or the inability to generate the required torque. Then, the number of successful cycles (repetitions) was recorded (NbRep).

Subjective rating of muscle fatigue.

Immediately after exhaustion, the subjects were asked to rate the fatigue of each muscle group using a Borg CR-10 scale (6) combined with a drawing of the body representing fives broad areas corresponding approximately to the positioning of the EMG electrodes: upper and lower back, buttocks, hamstrings, and quadriceps.

Signal processing.

All data processing was performed using Matlab subroutines (Version 7.0, release 14; MathWorks Inc., Natick, MA). From the EMG signals corresponding to the trunk ROM lying between 15° and −5° (to avoid the acceleration and deceleration portions of the concentric exertions during extension) and from all flexion-extension cycles, two EMG parameters were computed to assess muscle fatigue: the root mean square (RMS) amplitude and the instantaneous median frequency (IMF). A continuous wavelet transform (continuous Morlet, as described elsewhere [22]) was used to compute IMF values at each cycle, following the formulation of Bonato et al. (5), using successive (50% overlapped) 250-ms time windows (512 points). The corresponding values were afterward averaged to obtain one IMF value per cycle. Then, linear regression was applied to each of these time series (RMS and IMF values across cycles) to calculate the rate of decline in MF over time (MF/time slope: IMFslp) and the rate of increase in RMS over time (RMS/time slope: RMSslp), rates that were indicative of muscle fatigue. RMSslp and IMFslp were divided by their corresponding intercept value (obtained from the linear regression analysis) and multiplied by 100 to get normalized EMG indices of muscle fatigue (NIMFslp and NRMSslp), which account for subcutaneous tissue thickness differences between subjects. Finally, because no between-side differences were obtained (t-test), NIMFslp and NRMSslp scores were averaged bilaterally to reduce data to seven muscle groups (four back muscles, two hip extensors, and one VM) and to increase their reliability (23).

Interpretation of EMG indices of muscle fatigue.

Generally speaking, although NIMFslp is a more reliable index of muscle fatigue than NRMSslp (23), evidence of muscle fatigue is more obvious when NIMFslp is negative and NRMSslp is positive concomitantly (25), except when motor unit (MU) recruitment is not possible such as at higher force levels (>80% MVC), which is not the case here. In this case, the positive NRMSslp is indicative of MU recruitment without an increase of the muscle mechanical contribution. When only a negative NIMFslp is obtained, more subtle muscle fatigue is present, with an approximately zero NRMSslp value, indicating that no noticeable MU recruitment is required. On the other hand, when only a positive NRMSslp is obtained, MU recruitment is confirmed and a zero NIMFslp value indicates that this is not elicited by muscle fatigue, thus suggesting an increase of the muscle mechanical contribution. Obviously, these indications are valid only when the external net joint loading is constant during the exercise, which is the case in the present study.

Statistical analyses.

All statistical analyses were done with the NCSS statistical software for Windows (version 6.0), with an alpha of 0.05 as the level of statistical significance. Some of the variables (especially EMG indices of fatigue) were not normally distributed and thus were transformed to get normal distribution as verified with the Wilk-Shapiro test. However, the values reported in tables and figures are untransformed values. Although this study was not designed, in terms of sample size, to carry out gender comparisons, no gender differences were observed in any of the different fatigue criterion (Borg, NbRep, NIMFslp, and NRMSslp). Therefore, men and women were pooled to simplify all statistical analyses.

The t-tests were used to assess between-group (healthy vs CLBP) differences in anthropometric measures, Strength, and NbRep. To determine which muscle groups showed evidence of muscle fatigue, nontransformed NIMFslp and NRMSslp slope values were tested (Wilcoxon ranked test) to determine whether they were significantly different from zero. To further assess which muscle group was more associated with the end of the endurance exercise, Spearman correlations were carried out between these EMG indices of muscle fatigue (nontransformed values) and NbRep. To assess group and muscle differences in muscle fatigue using EMG indices (NIMFslp and NRMSslp), two-way ANOVA (2 group × 7 muscle) with repeated measures on the Muscle factor were carried out. Likewise, a two-way ANOVA (2 group × 5 muscle area) with repeated measures on the muscle-area factor was used to analyze Borg ratings. Post hoc analyses were performed, when necessary, using the Tukey test.


Strength and NbRep were not significantly different between healthy controls and patients with CLBP (Table 1). The nontransformed NIMFslp slope values were significantly lower than zero at all back muscle sites and for the VM, whereas NRMSslp slope values were significantly higher than zero invariably (Table 2 and Fig. 2). The highest significant correlations between the EMG indices of fatigue (NIMFslp and NRMSslp) and NbRep were obtained at L4 (NIMFslp = 0.68; NRMSslp = −0.65) and L3 (NIMFslp = 0.51), as detailed in Table 2.

Statistic results on the physiological significance of the EMG indices of fatigue.
Box plot representation corresponding to NIMFslp and NRMSslp results (untransformed values) for each group and each muscle site.

The ANOVA results showed no group × muscle interaction, neither for the EMG indices of fatigue nor for the Borg ratings (Table 3). Although a significant group main effect was observed for the Borg ratings (healthy < CLBP subjects), no group main effect was obtained with the EMG indices of fatigue. In fact, the effect sizes corresponding to NIMFslp were all below 0.31 (n = 7 muscles), which is relatively small. The Borg ratings showed a muscle main effect (Table 3), with the corresponding post hoc analyses revealing more muscle fatigue at the lower back than at the four other areas (upper back, buttocks, hamstrings, and quadriceps). Only the NIMFslp EMG parameter showed a muscle main effect (Table 3), with the post hoc analyses showing more fatigue at L4 than at T10, GM, BF, and VM.

ANOVA results (P values) corresponding to the comparisons between healthy and CLBP subjects (group factor) and between-muscle sites (muscle factor) for the EMG indices (NIMFslp and NRMSslp) and subjective perception (Borg) of muscle fatigue.


The main findings of the present study were that EMG revealed significantly more muscle fatigue of lower back muscles (L4 electrode sites) than at the other electrode sites, which was further corroborated by the subjective Borg scale assessment. Concerning the between-group differences, although the Borg assessment suggests that patients with CLBP were more fatigued, NbRep and EMG fatigue indices did not. However, no group × muscle interaction was disclosed for the Borg rating or for both EMG parameters.

Muscle main effect.

All back muscles as well as VM showed clear evidence of muscle fatigue (significantly negative NIMFslp and positive NRMSslp), whereas hip extensors showed evidence of MU recruitment without fatigue. However, between-muscle comparisons revealed more negative NIMFslp values at L4 than at T10, GM, BF, and VM, and the highest correlations between EMG indices of fatigue and NbRep were obtained at L4 (for NIMFslp and NRMSslp) and L3 (for NIMFslp only). Furthermore, these findings were corroborated by the subjective Borg scale assessment showing more muscle fatigue at the lower back than at the four other muscle areas (upper back, buttocks, hamstrings, and quadriceps). All these results point in the same direction, highlighting the specificity of this back endurance machine modality (posterior pelvic stabilization combined with semisitting position) to induce fatigue at the lower back. As mentioned earlier, NRMSslp is a less reliable EMG fatigue index (23) so that the between-muscle differences were not large enough to be detected with this parameter. This means that more subjects would probably be required to disclose significant differences using NRMSslp and also to disclose more subtle differences between the other muscle sites (e.g., between L4, L3 and L1) using the NIMFslp parameter.

The specific fatigue of lower back muscles can partly be explained by the increased back muscle activation with pelvic stabilization, as recently substantiated (10,36). This is also associated with the specific training of back muscles (16). However, lower-limb configuration is apparently also of importance. Previous findings in the same apparatus (Biodex), but with the knees at 90° (using a tibial pad), also showed clear low back muscle fatigue but also, contrary to the GM, a comparable fatigue level of the BF (9). In the present study, the fatigue level of the BF was apparently eliminated by positioning the knees at 135°. This position lengthens hamstrings (29), which increase their strength according to the length-tension relationship. Considering that the net moment at the hip theoretically remains the same, this in turn decreases the relative contribution (relative to their higher maximal strength) of these muscles (10). On the other hand, the EMG analyses of GM and BF muscle groups showed evidence of MU recruitment (without fatigue) that would most likely be attributable to an increased mechanical contribution of hip extensors. These muscles have the potential to extend the knee in this machine and to contribute, with the VM, to stabilize the body by pushing against the footboard. It is also possible that hip extensors try to compensate for lower back muscle fatigue by tensioning the passive tissues (sacrotuberous ligament and thoracolumbar fascia), linking hip extensors to lower back muscles (31).

The presence of VM fatigue was not unexpected with this exercise modality because the subjects have to push against the footboard to stabilize the body, which in turn facilitates the generation of extension torque at the lower back. However, according to ANOVA results (NIMFslp parameter), the level of fatigue is not as important as at the lower back, and the corresponding NIMFslp/NbRep correlation (r = 0.43) was not as high as at L4 (r = 0.68) and L3 (r = 0.51), thus suggesting that VM fatigue is less likely to limit the number of repetitions required to specifically overload the back muscles.

Group main effect.

A statistically significant group difference was obtained with the subjective fatigue criterion (Borg), whereas the more objective mechanical (NbRep) and EMG (NIMFslp and NRMSslp) fatigue criteria did not. This suggests that the increased perception of fatigue in patients with CLBP, and this independently of muscle groups (group × muscle interaction), could be explained by the difficulty to disconnect discomforts due to pain nociception emerging from lower back passive tissues from discomforts due to muscle fatigue per se.

Taking into account an equivalent strength between groups (CLBP executed "true" MVC regardless of their fears or pain), the 60% relative load was also likely equivalent, thus confirming that the patients with CLBP did not show abnormal back muscle fatigue. The sole study where similar conditions are reunited [equivalent strength, trunk at vertical (though the contraction was isometric), 80% MVC] or in other words the sole study where most confounding effects were apparently avoided showed a higher fatigability of back muscles as assessed with similar NIMFslp indices (35). We believe that these conflicting findings might be attributable to the lack of abnormal muscle composition in our patients because heterogeneous findings are reported relative to back muscle composition of patients with CLBP (8,26). Crossman et al. (8) suggest that not all patients may present abnormal back histomorphometric findings and the associated back muscle weakness and fatigability, depending on if they are "avoiders" or "confronters," as proposed in the fear-avoidance model (40). Maybe our CLBP sample was mostly composed of confronters, as suggested by the strength and physical activity level results (Table 1). In line with this explanation, avoiders have been shown to fatigue faster than confronters, which were comparable with healthy controls, again as assessed with similar NIMFslp indices (muscle sites at L4/L5 level) during an isometric extension contraction (2).

Two other comparable EMG studies, examining dynamic contractions in a machine with a pin-loaded stack but with the weight adjusted according to physical characteristics, did not obtain a group difference as long as back muscles are concerned (18,21). However, in one of these studies, patients with CLBP showed higher gluteus maximus fatigue than healthy controls so that no muscle fatigue difference was obtained when compared with back muscles (21). However, they used a sitting position (knees at 90°), and the stabilization pad was positioned at the lower back (at L3), making it difficult to ascertain whether the pelvis was posteriorly stabilized as in the present study. They suggested that patients with CLBP used the abovementioned mechanisms linking GM and back muscles through the thoracolumbar fascia, which might be possible in these exercise conditions. Nevertheless, the main finding of the present study regarding differences between groups is that no group × muscle interaction was disclosed for Borg or for the EMG parameters, suggesting that the present findings with regards to the specificity of this back muscle endurance exercise apply as much in patients with CLBP as in healthy controls.


As outlined previously, it appears always possible to find different findings in other CLBP subgroups, and the sample of subjects was not large enough to well represent the heterogeneity of patients with nonspecific CLBP. The fact that the present patients with CLBP were not significantly different from healthy controls according to strength and NbRep results might represent a limitation in this respect. Also, although age and anthropometric characteristics of the patients and controls were similar (Table 1), patients were not perfectly matched to healthy controls (not a case-control study). These results cannot necessarily be generalized to obese subjects for whom the use of stabilization devices could be difficult. No direct comparisons with other exercise modalities (e.g., 90° knee angle) were carried out with the use of the same subjects. Finally, no deep back muscles were investigated using intramuscular EMG, and a specialized dynamometer was used instead of a real machine with a pin-loaded stack.


Because back muscle endurance is an important clinical outcome as suggested by prospective studies (3,13), therapists and trainers in clinical or fitness centers should pay attention to design more specific exercise programs for the endurance training of back muscles, specially for patients with LBP (27). The results of this study should help in this respect. This does not mean that the training of other muscle groups is not of importance, but it is important to plan, during the rehabilitation process, a specific exercise to target the muscles that often show histomorphometric alterations (26) as well as atrophy (17), namely, the back muscles. Of course, whether such muscle atrophy is the consequence of reflex inhibition could possibly influence the effectiveness of this specific exercise. Further studies are required to verify whether this exercise modality is specific enough to reverse this situation and ultimately to determine whether increasing back muscle endurance would help to further improve clinical outcomes and protect against recurrences.

R. A. da Silva, a Ph.D. student, and this project were both funded by the Occupational Health and Safety Research Institute Robert-Sauvé (IRSST). Sylvie Nadeau is a senior research scientist from the Fonds de la recherche en santé du Québec. The authors gratefully acknowledge the assistance of Flavia O'Delloso for the recruitment of subjects and data collection as well as Michel Goyette and Daniel Marineau for technical support. Finally, they also greatly thank David McFadden and Jean-François Pilon for their help in data processing. The results of the present study do not constitute endorsement by the American College of Sports Medicine.

Disclosure statement: The authors disclose that this study was funded by the Occupational Safety and Health Research Institute Robert-Sauvé from Québec (Institut de recherche Robert-Sauvé en santé et en sécurité du travail; IRSST).


1. Baecke JAH, Burema J, Frijters JER. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982;36(5):936-42.
2. Biedermann HJ, Shanks GL, Forrest WJ, Inglis J. Power spectrum analyses of electromyographic activity. Discriminators in the differential assessment of patients with chronic low-back pain. Spine. 1991;16(10):1179-84.
3. Biering-Sorensen F. Physical measurements as risk indicators for low-back trouble over a one-year period. Spine. 1984;9(2):106-19.
4. Bilodeau M, Schindler-Ivens S, Williams DM, Chandran R, Sharma SS. EMG frequency content changes with increasing force and during fatigue in the quadriceps femoris muscle of men and women. J Electromyogr Kinesiol. 2003;13(1):83-92.
5. Bonato P, Roy SH, Knaflitz M, De Luca CJ. Time-frequency parameters of the surface myoelectric signal for assessing muscle fatigue during cyclic dynamic contractions. IEEE Trans Biomed Eng. 2001;48(7):745-53.
6. Borg GAV. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14(5):377-81.
7. Clark BC, Manini TM, Ploutz-Snyder LL. Derecruitment of the lumbar musculature with fatiguing trunk extension exercise. Spine. 2003;28(3):282-7.
8. Crossman K, Mahon M, Watson PJ, Oldham JA, Cooper RG. Chronic low back pain-associated paraspinal muscle dysfunction is not the result of a constitutionally determined "adverse" fiber-type composition. Spine. 2004;29(6):628-34.
9. da Silva R, Lariviere C, Arsenault AB, Nadeau S, Plamondon A. The comparison of wavelet- and Fourier-based electromyographic indices of back muscle fatigue during dynamic contractions: validity and reliability results. Electromyogr Clin Neurophysiol. 2008;48:147-62.
10. da Silva R, Larivière C, Arsenault AB, Nadeau S, Plamondon A. Pelvic stabilization and semi-sitting position increase the specificity of back exercises. Med Sci Sports Exerc. 2009;41(2):435-43.
11. Danneels LA, Vanderstraeten GG, Cambier DC, et al. Effects of three different training modalities on the cross sectional area of the lumbar multifidus muscle in patients with chronic low back pain. Br J Sports Med. 2001;35(3):186-91.
12. Durnin JVGA, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr. 1974;32(1):77-97.
13. Enthoven P, Skargren E, Kjellman G, Oberg B. Course of back pain in primary care: a prospective study of physical measures. J Rehabil Med. 2003;35(4):168-73.
14. French DJ, Roach P, Mayes S. Fear of movement in injured workers: The Tampa Scale of Kinesiophobia (EKT). Can J Behav Sci. 2002;34(1):28-33.
15. Granata KP, Gottipati P. Fatigue influences the dynamic stability of the torso. Ergonomics. 2008;51(8):1258-71.
16. Graves JE, Webb DC, Pollock ML, et al. Pelvic stabilization during resistance training: its effect on the development of lumbar extension strength. Arch Phys Med Rehabil. 1994;75(2):210-5.
17. Hides JA, Stokes MJ, Saide M, Jull GA, Cooper DH. Evidence of lumbar multifidus muscle wasting ipsilateral to symptoms in patients with acute/subacute low back pain. Spine. 1994;19(2):165-72.
18. Kankaanpaa M, Colier WN, Taimela S, et al. Back extensor muscle oxygenation and fatigability in healthy subjects and low back pain patients during dynamic back extension exertion. Pathophysiol. 2005;12:267-73.
19. Kankaanpaa M, Laaksonen D, Taimela S, Kokko SM, Airaksinen O, Hanninen O. Age, sex, and body mass index as determinants of back and hip extensor fatigue in the isometric Sorensen back endurance test. Arch Phys Med Rehabil. 1998;79(9):1069-75.
20. Kankaanpaa M, Taimela S, Airaksinen O, Hanninen O. The efficacy of active rehabilitation in chronic low back pain. Effect on pain intensity, self-experienced disability, and lumbar fatigability. Spine. 1999;24(10):1034-42.
21. Kankaanpaa M, Taimela S, Laaksonen D, Hanninen S, Airaksinen O. Back and hip extensor fatigability in chronic low back pain patients and controls. Arch Phys Med Rehabil. 1998;79(4):412-7.
22. Karlsson S, Gerdle B. Mean frequency and signal amplitude of the surface EMG of the quadriceps muscles increase with increasing torque-a study using the continuous wavelet transform. J Electromyogr Kinesiol. 2001;11(2):131-40.
23. Larivière C, Arsenault AB, Gravel D, Gagnon D, Loisel P. Evaluation of measurement strategies to increase the reliability of EMG indices to assess back muscle fatigue and recovery. J Electromyogr Kinesiol. 2002;12(2):135-46.
24. Lee C, Minamitani H, Ju K, Wakano K, Onishi S, Yamazaki H. Fatigue evaluation of lumbar muscles during repeated dynamic trunk exercise. Electromyogr Clin Neurophysiol. 1996;36(4):237-45.
25. Luttmann A, Jager M, Laurig W. Electromyographical indication of muscular fatigue in occupational field studies. Int J Ind Ergon. 2000;25:645-60.
26. Mannion AF, Weber BR, Dvorak J, Grob D, Muntener M. Fibre type characteristics of the lumbar paraspinal muscles in normal healthy subjects and in patients with low back pain. J Orthop Res. 1997;15(6):881-7.
27. Mayer J, Mooney V, Dagenais S. Evidence-informed management of chronic low back pain with lumbar extensor strengthening exercises. Spine J. 2008;8(1):96-113.
28. Moffroid MT, Haugh LD, Henry SM, Short B. Distinguishable groups of musculoskeletal low back pain patients and asymptomatic control subjects based on physical measures of the NIOSH low back atlas. Spine. 1994;19(12):1350-8.
29. Nemeth G, Ohlsen H. In vivo moment arm lengths for hip extensor muscles at different angles of hip flexion. J Biomech. 1985;18(2):129-40.
30. Plamondon A, Trimble K, Lariviere C, Desjardins P. Back muscle fatigue during intermittent prone back extension exercise. Scand J Med Sci Sports. 2004;14(4):221-30.
31. Pool-Goudzwaard AL, Vleeming A, Stoeckart R, Snijders CJ, Mens JM. Insufficient lumbopelvic stability: a clinical, anatomical and biomechanical approach to 'a-specific' low back pain. Man Ther. 1998;3(1):12-20.
32. Rainville J, Hartigan C, Martinez E, Limke J, Jouve C, Finno M. Exercise as a treatment for chronic low back pain. Spine J. 2004;4(1):106-15.
33. Robinson ME, Cassisi JE, O'Connor PD, MacMillan M. Lumbar iEMG during isotonic exercise: chronic low back pain patients versus controls. J Spinal Disord. 1992;5(1):8-15.
34. Roland M, Morris R. A study of the natural history of back pain. Part I: development of a baseline and sensitive measure of disability in low-back pain. Spine. 1983;8(2):141-4.
35. Roy SH, De Luca CJ, Casavant DA. Lumbar muscle fatigue and chronic lower back pain. Spine. 1989;14(9):992-1001.
36. San Juan JG, Yaggie JA, Levy SS, Mooney V, Udermann BE, Mayer JM. Effects of pelvic stabilization on lumbar muscle activity during dynamic exercise. J Strength Cond Res. 2005;19(4):903-7.
37. Smeets RJ, Vlaeyen JW, Hidding A, et al. Active rehabilitation for chronic low back pain: cognitive-behavioral, physical, or both? First direct post-treatment results from a randomized controlled trial. BMC Musculoskelet Disord. 2006;7(1):5.
38. Sullivan MJL, Bishop SR, Pivik J. The Pain Catastrophizing Scale: development and validation. Psychol Assess. 1995;7:524-32.
39. van Dieen JH, Selen LP, Cholewicki J. Trunk muscle activation in low-back pain patients, an analysis of the literature. J Electromyogr Kinesiol. 2003;13(4):333-51.
40. Vlaeyen JWS, Linton SJ. Fear-avoidance and its consequences in chronic musculoskeletal pain: a state of the art. Pain. 2000;85(3):317-32.


©2010The American College of Sports Medicine