Secondary Logo

Journal Logo

Original Research

Core Muscle Activation During Dynamic Upper Limb Exercises in Women

Tarnanen, Sami P.1; Siekkinen, Kirsti M.2; Häkkinen, Arja H.1,3; Mälkiä, Esko A.1; Kautiainen, Hannu J.4; Ylinen, Jari J.3

Author Information
Journal of Strength and Conditioning Research: December 2012 - Volume 26 - Issue 12 - p 3217-3224
doi: 10.1519/JSC.0b013e318248ad54
  • Free



The core is a functional center of the kinetic chain (18). Core stability can be viewed as the product of motor control and the muscular capacity of the lumbo-pelvic-hip complex (19,26). Neuromuscular control of the core is a requirement not only for proper functioning of the spine but also the hip and lower extremities. Deficiencies in neuromuscular control and muscle capacity may result in uncontrollable motion and injury (35). Greater core stability may benefit performance by providing a foundation for greater force production in the upper and lower extremities (34). Thus, core muscle exercises have a theoretical basis in the rehabilitation of various musculoskeletal disorders, injury prevention, and improved performance (1,23,28,35).

In core muscle exercises, the emphasis should be on training muscle activation in functional positions and motions (5). A variety of occupational pushing and pulling tasks are present in manual material handling; for example, in industry or warehouse and distribution settings that expose the core to horizontal forces. Also, several everyday functions and sporting activities, such as tennis, crosscountry skiing, and rowing, demand the controlled use of abdominal and back muscles while working with the upper limbs (6,15,21,23,27). Because push or pull movement work tasks are often associated with low back injuries (16), the supporting capacity of the core stabilizing muscles may not be adequate for preventing trunk torque.

Electromyography (EMG) has been used to determine muscle activity during traditional trunk muscle exercises. These exercises are often performed in a supine hook lying, prone, side-lying, or 4-point kneeling position. The mean averaged EMG amplitudes of the trunk muscles during floor exercises are typically <50% of the maximal voluntary contraction amplitude and, therefore, may be inadequate to enhance muscle strength characteristics (3,9,11,31,33). Core muscle activity has been previously examined during upper limb exercises performed in the sitting or standing position, but the effectiveness of the respective exercises is unclear because resistance has been applied as either a constant load, a manual resistance provided by an assistant, or using an elastic band so that the magnitude of the load has been undefined (2–4). Furthermore, muscle activation levels are unclear for functional exercises that mimic the demands of occupational tasks, daily tasks, sport activities, or where the resistance level is individualized.

The purpose of this study was to determine core muscle activity during 1 repetition maximum (1RM) in 7 dynamic upper limb push and pull exercises. We hypothesized that the activation level of core muscles varies based on upper limb position and movement direction and that pelvis fixation increases core muscle activity during upper limb exercises.


Experimental Approach to the Problem

The activity of the core muscles was recorded by surface EMG during maximal voluntary isometric trunk flexion, lateral flexion, and extension contractions to establish comparative reference values. The core muscle activity during maximal dynamic upper limb exercises with fixated pelvis was normalized to the reference values. Normalized activity values were used in the evaluation of the intensity of muscle activation. The 2 upper limb exercises were performed also without pelvis fixation to test the hypothesis regarding the effect of fixation.


Twenty healthy women were recruited from among hospital staff. The subjects were excluded if they had any neuromuscular, orthopedic, or cardiorespiratory problems preventing the type of physical exertion required for this study. The mean age of the subjects was 38.1 (SD 7) years and mean body mass index was 23 (SD 2). The physical activity level of the subjects was determined using a written questionnaire and a Time-Weighted Average Power calculated for each subject's daily active time (work, leisure time physical activity, and controlled exercises) using the MetPro software program (MetPro 2.03.9 MX, Kuntoväline Inc. Helsinki, Finland) (30). The scale unit is from 1.5 (light work) to 9.5 (extremely strenuous work) METs (24). The mean physical activity level of the subjects was 3 (SD 1) METs. The study was performed in the biomechanics laboratory. The subjects were informed of the study protocol and possible risks and discomforts related to the measurements. All the participants signed an informed consent form before participation. The study protocol was approved by the Ethical Committee of the Central Finland Central Hospital.


Eight-channel ME3000P8 surface EMG (Mega Electronics Ltd., Kuopio, Finland) was used for measurements. Raw EMG data were recorded with a sampling frequency of 1,000 Hz and the band-pass filtered using a bandwidth of 8–500 Hz (Butterworth). A differential amplifier was used to strengthen and filter the measured signal and to reduce the noise with a common mode rejection ratio >110 dB, a root mean square of noise <1.6 μV, and an amplification level of 412. The amplifier's feed impedance was >10GΩ. The surface EMG signal was converted to a digital format with a 12-bit analog-to-digital converter and saved on a computer for analysis. The raw EMG signal was rectified and averaged. The average amplitude level (in microvolts) of every exercise was calculated as the average (8) of the data segments during an analysis period (100 milliseconds). A 4-second time period was selected for analysis from the point where the electrical activity was greatest during each isometric exercise. The starting and finishing points of the upper limb exercises were determined from simultaneous electromyographic and video (DCR-HC96E, Sony, Tokyo, Japan) analysis, and the whole period was used for analysis.

Round, single-use silver/silver chloride (Ag/AgCl) surface electrodes (M-00-S, Medicotest Inc., Ölstykke, Denmark) were used. The skin at the electrode attachment sites was shaved, cleaned with sandpaper, and wiped with alcohol to decrease skin impedance. Electrode pairs were positioned on both sides of the body on the following core muscles in the direction of the muscle fibers (10,13,25): rectus abdominis, 1 cm above the navel and 2 cm laterally from the midline; obliquus externus abdominis, just below the curvature of the ribs; longissimus, 3 cm laterally from the L1 spinous process, and multifidus, 2 cm laterally from the L5 spinous process.

Familiarization Session

The heights of the isometric strength measurement frame and pelvis fixation frame were adjusted individually. The participants practiced dynamic upper limb exercises with very light load until the performance technique was correct. The weights were increased gradually on the weight stack system of a pull machine (544, Frapp, Joensuu, Finland) to find 1RM. The smallest increase was 1.25 kg. The average time between familiarization and the actual measurement session was 8 days (range 4–17 days). The participants were instructed to avoid strenuous physical activity before the actual measurement session.

Measurement Session

After the attachment of the electrodes, a period of 15 minutes was allowed to pass before the commencement of measurements. Isometric reference trunk exercises were measured first, followed by upper limb 1RM exercises (Figure 1). Both exercise series were carried out in a random order, and 3 submaximal warm-up trials were performed for each exercise. To avoid muscle fatigue, a rest period of at least 1 minute was used between each effort and 5 minutes between the reference and dynamic upper limb exercises (22). The subjects were verbally encouraged and the speed for each 1RM exercise dictated by a metronome (50 b·min−1).

Figure 1
Figure 1:
Illustration of the exercises: A) the starting position of the bilateral shoulder extension (exercise 5), B) the final position of the bilateral shoulder extension, C) the starting position of the shoulder horizontal adduction (exercise 7), D) the final position of the shoulder horizontal adduction, E) the starting position of the shoulder horizontal abduction (exercise 9), F) the final position of the shoulder horizontal abduction, G) the starting position of the shoulder flexion (exercise 10), H) the final position of the shoulder flexion, I) the starting position of the shoulder extension (exercise 11), J) the final position of the shoulder extension. The subject has given permission for the use of this figure.

Reference Exercises

Isometric reference exercises were performed in the standing position to flexion (exercise 1), extension (exercise 2), and lateral flexion (exercises 3 and 4) in the same manner as in a previous study (32). Following a command to commence each exercise, the subjects performed with maximal effort for approximately 5 seconds. After warm-up, 2 maximal efforts were performed and the maximal strength measured using a strain-gauge dynamometer (DS Europe, Milan, Italy). If strength increased >10% from the first effort, one additional effort was measured. Reliability of isometric trunk strength tests and trunk EMG measurements has been evaluated in earlier studies (7,8,17).

Upper Limb Exercises

The dynamic 1RM upper limb exercises were performed in the standing position using a pull machine (Figure 1). The subject's left leg was in the front so the heel of the left foot was in the same line as the toes of the right foot. The lateral distance between the feet was 20 cm, and the left knee was in a slight flexion. Exercises were performed in a manner that retained the neutral posture of the lumbar spine. The studied exercises were bilateral shoulder extension (exercise 5), shoulder horizontal adduction without fixation (exercise 6), shoulder horizontal adduction (exercise 7), shoulder horizontal abduction without fixation (exercise 8), shoulder horizontal abduction (exercise 9), shoulder flexion (exercise 10), and shoulder extension (exercise 11). Exercise 5 was performed with both arms at the same time and exercises 6–11 with the right arm. The pelvis was fixated to the measurement frame with a belt at the level of the greater trochanter, but shoulder horizontal adduction and abduction were also done without fixation.

Statistical Analyses

The results are presented as means with SDs. Activity levels of each muscle were normalized by being expressed as a percent contribution of the activity during reference exercises. Repeated measures analysis of variance (ANOVA) was used to compare the surface EMG of each muscle, which was elicited during the different exercises. Paired t-tests were used to compare the force values and EMG activities of the different upper limb exercises and between the sides. The same tests were used to compare the muscle activity with or without pelvis fixation. No adjustment was made for multiple testing, but this information can be obtained by multiplying the actual p-value by the number of comparisons made. The alpha level was set to p ≤ 0.05 to determine statistical significance. All statistical analyses were performed using the SPSS statistical software program (SPSS Inc., Chicago, IL, USA, Version 13.0).


Great individual variation was observed in both the trunk isometric force and maximal load in upper limb 1RM dynamic exercises (Table 1). The highest resistance in the upper limb dynamic exercises was during right shoulder extension and the lowest in right shoulder horizontal adduction without fixation. The resistance was considerably greater during pulling (exercises 9 and 11) compared with pushing (exercises 7 and 10; p < 0.001). The load was higher in exercises 7 and 9, in which the pelvis was fixed, compared to nonfixed exercises 6 and 8 (p < 0.001).

Table 1
Table 1:
Force production values for reference exercises of the trunk and load in upper limb exercises when carried out with or without pelvis fixation.*

The highest activation in the rectus abdominis (Figure 2) was measured during fixed bilateral shoulder extension; the activity on the right side was 103% and the left side 96% compared with the reference value during trunk flexion. In addition, a moderate level of bilateral activation was measured during unilateral shoulder horizontal adduction with fixation.

Figure 2
Figure 2:
The averaged surface electromyography (sEMG) amplitude (microvolts) of the left and right rectus abdominis muscles during maximal isometric trunk efforts and dynamic upper limb exercises; nf = no fixation. Dots present means and whiskers representSDs.

The highest bilateral activity of the obliquus externus abdominis (Figure 3) with fixation was measured during shoulder horizontal adduction and bilateral shoulder extension. Activation of the ipsilateral side was noticeable during shoulder flexion. The relative muscle activity during shoulder horizontal adduction was 65% on the ipsilateral side and 75% on the contralateral side compared to trunk flexion, and corresponding values during bilateral shoulder extension were 61 and 68%. During shoulder flexion, ipsilateral activity was 60% of the activity during trunk flexion.

Figure 3
Figure 3:
The averaged surface electromyography (sEMG) amplitude (microvolts) of the left and right obliquus externus abdominis muscles during maximal isometric trunk efforts and dynamic upper limb exercises; nf = no fixation. Dots present means and whiskers representSDs.

During unilateral shoulder extension with fixation, the activity on the contralateral longissimus (Figure 4) was 67% of the reference value during trunk extension. The greatest activity on the ipsilateral side was 60%, which was measured during shoulder horizontal abduction with fixation. However, the activity of the longissimus during isometric trunk extension was significantly greater than activation during all dynamic upper limb exercises (p < 0.05).

Figure 4
Figure 4:
The averaged surface electromyography (sEMG) amplitude (microvolts) of the left and right longissimus muscles during maximal isometric trunk efforts and dynamic upper limb exercises; nf = no fixation. Dots present means and whiskers representSDs.

Shoulder extension and horizontal abduction, both with fixation, activated the multifidus to the greatest extent (Figure 5). During shoulder extension, the activity of the multifidus was 80% on the contralateral side and 53% on the ipsilateral side compared to trunk extension. During shoulder horizontal abduction, the activity of the multifidus was 70% on the contralateral side and 55% on the ipsilateral side. The activity of the multifidus muscles during isometric trunk extension was significantly greater than ipsilateral activation in all dynamic exercises (p < 0.05) and contralateral activation in 5 of 7 exercises.

Figure 5
Figure 5:
The averaged surface electromyography (sEMG) amplitude (microvolts) of the left and right multifidus muscles during maximal isometric trunk efforts and dynamic upper limb exercises; nf = no fixation. Dots present means and whiskers representSDs.

During unilateral shoulder horizontal adduction without pelvis fixation, average activation of the rectus abdominis was 64% lower (p < 0.001) and activation of the externus obliquus abdominis 44% (p < 0.001) lower than during the fixed exercise. The activity levels of the longissimus and multifidus during unfixed shoulder horizontal abduction were 43% (p < 0.001) and 35% (p < 0.001) lower, respectively, than the same exercise with fixation.


By changing the direction of movement and the lever arm of an upper limb while maintaining the neutral position of the lumbar spine, each of the studied muscle groups reached an activation level at which muscle endurance and strength development might be expected. The greatest activity in back muscles was achieved during shoulder extension with fixation, especially on the contralateral side, and the greatest abdominal muscle activity during bilateral shoulder extension and shoulder horizontal adduction. The activity in trunk flexor and extensor muscles during pelvis fixation was greater than during unfixed exercises. The finding was expected but has not been reported previously in the literature.

The activity of core muscles during dynamic upper limb pushing and pulling exercises were previously measured only in 3 studies. In Santana et al. (29), the activity of back and abdominal muscles was measured during 1RM shoulder horizontal adduction, reporting that the activities of the contralateral side obliquus internus and externus abdominis were >60% of maximal activity. In our study, the activity level of the obliquus externus abdominis in horizontal adduction was 65% on the ipsilateral side and 75% on the contralateral side compared to trunk flexion but activity level decreased on average 44% when exercise was done without pelvis support. In Fenwick et al. (12), only the upper erector spinae (T9 level) exceeded 40% of the maximal voluntary contraction in unilateral shoulder extension, whereas in this study, the activity of the longissimus was 67% during shoulder extension. The higher activity level in our study may be explained by the use of pelvis fixation and a higher resistance level. The findings of thisstudy are comparable with the results of our previous study (32) in which the effect of isometric upper limb exercises on the core stabilizing muscles was studied. This observation was surprising because dynamic movements are more challenging for maintaining motor control.

Core stability exercises can be divided into 2 main groups: exercises that aim to increase the coordination and control of the trunk muscles to improve control of the lumbar spine and pelvis, and exercises that aim to increase the strength and endurance capacity of the trunk muscles to meet the demands of control. In addition, balance and movement efficiency are also necessary to optimize trunk function, which is integral to performance and sport-related skills (14,26). To add muscle mass and strength, the recommended resistance is 60–80% of the maximal voluntary contraction (20). In dynamic upper limb exercises with a fixed pelvis, an activity level of >60% of the maximum activity measured during reference movements was achieved; thus, these exercises are applicable to muscle endurance and strength training. Pelvis fixation increases trunk muscle activity, but it may also add load to the spinal skeleton and other supporting structures. On the other hand, fixation may decrease twisting displacement during exercise. Thus, the use of fixation should always be considered individually and take into account possible injury history and the individual's ability to perform the exercises with the correct technique.

To maintain spine stability, the contralateral muscles react by increasing muscle activity. In this study, dynamic unilateral upper limb exercises, the ipsilateral and contralateral sides of the muscles were activated in different ways. The largest differences between the ipsilateral and contralateral side were caused by shoulder flexion and extension exercises for the obliquus externus abdominis, longissimus, and multifidus muscles, and the smallest differences in the rectus abdominis. These differences in unilaterally performed limb exercises have also been noted in previous studies (4,12,29). No significant differences were found between the sides during bilateral shoulder extension with fixation and shoulder horizontal abduction without fixation.

One limitation of this study was that the performance speed varied between 1.1 and 3.1 seconds, with an average of 1.8 seconds, showing the difficulty in maintaining a preset speed. Most study subjects had not performed these kinds of exercises previously, which is why the familiarization phase was used. However, this inexperience may be manifest in suboptimal activation of directly involved muscles and excessive co-contractions of the back and abdominal musculature. The activity levels achieved during these upper limb exercises to develop trunk stabilization were achieved during 1RM performance. When several repetitions are performed, it is likely that activity level of core muscles is lower.

Practical Applications

The core musculature has 2 main functions: control lumbopelvic stability and create proximal stability during upper and lower limbs movements. In the integrated core and shoulder exercises, these functions can be trained simultaneously. By performing core exercises in the standing position with free weights or pulleys, they mimic specific core muscle function patterns that are needed during occupational tasks, sport activities, or functional daily tasks. Thus, the transfer effect of functional training for real life demands may be better than in traditional abdominal and back muscle floor training. Because the neutral position of the lumbar spine is maintained during these upper limb exercises, they are also suitable for rehabilitation purposes when movement of the lumbar spine is painful or when movement of the lumbar spine has to be avoided.

Upper limb exercises performed in the standing position activate abdominal and back muscles by producing torque within the torso. By changing upper limb positions (lever arm) and the direction of movement (pushing vs. pulling), selective core muscle involvement may be achieved. If the focus of training is to increase the endurance and strength characteristics of the core muscles, then pelvis fixation should be used. The use of pelvis fixation may also be justifiable at the beginning of an exercise program or if there is a problem with the performance technique. Without pelvis fixation, the focus of exercise moves toward coordination and movement pattern training.


This study was funded by the Academy of Finland and Central Finland Central Hospital, Jyväskylä, Finland. Esko Mälkiä has a decision-making position at Kuntoväline Inc. (MetPro 2.03.9 MX, Kuntoväline Inc. Helsinki, Finland).


1. Akuthota V, Nadler SF. Core strengthening. Arch Phys Med Rehabil 85: S86–S92, 2004.
2. Arokoski JP, Kankaanpää M, Valta T, Juvonen I, Partanen J, Taimela S, Lindgren KA, Airaksinen O. Back and hip extensor muscle function during therapeutic exercises. Arch Phys Med Rehabil 80: 842–850, 1999.
3. Arokoski JP, Valta T, Airaksinen O, Kankaanpää M. Back and abdominal muscle function during stabilization exercises. Arch Phys Med Rehabil 82: 1089–1098, 2001.
4. Behm DG, Leonard AM, Young WB, Bonsey AC, MacKinnon SN. Trunk muscle electromyographic activity with unstable and unilateral exercises. J Strength Cond Res 19: 193–201, 2005.
5. Borghuis J, Hof AL, Lemmink KA. The importance of sensory-motor control in providing core stability: implications for measurement and training. Sports Med 38: 893–916, 2008.
6. Chow JW, Shim JH, Lim YT. Lower trunk muscle activity during the tennis serve. J Sci Med Sport 6: 512–518, 2003.
7. Dankaerts W, O'Sullivan PB, Burnett AF, Straker LM, Danneels LA. Reliability of EMG measurements for trunk muscles during maximal and sub-maximal voluntary isometric contractions in healthy controls and CLBP patients. J Electromyogr Kinesiol 14: 333–342, 2004.
8. Danneels LA, Cagnie BJ, Cools AM, Vanderstraeten GG, Cambier DC, Witvrouw EE, De Cuyper HJ. Intra-operator and inter-operator reliability of surface electromyography in the clinical evaluation of back muscles. Man Ther 6: 145–153, 2001.
9. Davidson KL, Hubley-Kozey CL. Trunk muscle responses to demands of an exercise progression to improve dynamic spinal stability. Arch Phys Med Rehabil 86: 216–223, 2005.
10. DeFoa JL, Forrest W, Biedermann HJ. Muscle fibre direction of longissimus, iliocostalis and multifidus: Landmark-derived reference lines. J Anat 163: 243–247, 1989.
11. Ekstrom RA, Osborn RW, Hauer PL. Surface electromyographic analysis of the low back muscles during rehabilitation exercises. J Orthop Sports Phys Ther 38: 736–745, 2008.
12. Fenwick CM, Brown SH, McGill SM. Comparison of different rowing exercises: Trunk muscle activation and lumbar spine motion, load, and stiffness. J Strength Cond Res 23: 1408–1417, 2009.
13. Hermens HJ, Freriks B, Merletti R, Hägg G, Stegeman D, Blok J, Rau G, Disselhorst-Klug C. European Recommendations for Surface Electromyography: Deliverable of the SENIAM Project. Enschede, The Netherlands: Roessingh Research and Development, 1999.
14. Hodges PW. Core stability exercise in chronic low back pain. Orthop Clin North Am 34: 245–254, 2003.
15. Holmberg HC, Lindinger S, Stöggl T, Eitzlmair E, Müller E. Biomechanical analysis of double poling in elite cross-country skiers. Med Sci Sports Exerc 37: 807–818, 2005.
16. Hoozemans MJ, van der Beek AJ, Frings-Dresen MH, van Dijk FJ, van der Woude LH. Pushing and pulling in relation to musculoskeletal disorders: a review of risk factors. Ergonomics 41: 757–781, 1998.
17. Hutten MM, Hermend HJ. Reliability of lumbar dynamometry measurements in patients with chronic low back pain with test-retest measurements on different days. Eur Spine J 6: 54–62, 1997.
18. Kennedy DJ, Noh MY. The role of core stabilization in lumbosacral radiculopathy. Phys Med Rehabil Clin N Am 22: 91–103, 2011.
19. Kibler WB, Press J, Sciascia A. The role of core stability in athletic function. Sports Med 36: 189–198, 2006.
20. Kraemer WJ, Ratamess NA. Fundamentals of resistance training: progression and exercise prescription. Med Sci Sports Exerc 36: 674–688, 2004.
21. Lett KK, McGill SM. Pushing and pulling: personal mechanics influence spine loads. Ergonomics 49: 895–908, 2006.
22. Matuszak ME, Fry AC, Weiss LW, Ireland TR, McKnight MM. Effect of rest interval length on repeated 1 repetition maximum back squat. J Strength Cond Res 17: 634–637, 2003.
23. McGill SM. Low back stability: from formal description to issues for performance and rehabilitation. Exerc Sport Sci Rev 29: 26–31, 2001.
24. Mälkiä EA. MET based questionnaire for the study of physical activity. In: Assessment of Function and Movement. Selected Papers. Third Nordic Symposium on Physiotherapy 1996. Mälkiä E., Sihvonen S., eds. Jyväskylä, Finland: Paino-Porras Oy, 1996. pp. 91–103.
25. Ng JK-F, Kippers V, Richardson CA. Muscle fibre orientation of abdominal muscles and suggested surface EMG electrode position. Electromyogr Clin Neurophysiol 38: 51–58, 1998.
26. Okada T, Huxel KC, Nesser TW. Relationship between core stability, functional movement, and performance. J Strength Cond Res 25: 252–261, 2011.
27. Pollock CL, Jenkyn TR, Jones IC, Ivanova TD, Garland SJ. Electromyography and kinematics of the trunk during rowing in elite female rowers. Med Sci Sports Exerc 41: 628–636, 2009.
28. Saeterbakken AH, van den Tillaar R, Seiler S. Effect of core stability training on throwing velocity in female handball players. J Strength Cond Res 25: 712–718, 2011.
29. Santana JC, Vera-Garcia FJ, McGill SM. A kinetic, and electromyographic comparison of the standing cable press and bench press. J Strength Cond Res 21: 1271–1279, 2007.
30. Sjögren T, Nissinen K, Järvenpää S, Ojanen M, Vanharanta H, Mälkiä E. Effects of a workplace physical exercise intervention on the intensity of headache and neck and shoulder symptoms and upper extremity muscular strength of office workers: A cluster randomized controlled cross-over trial. Pain 116: 119–128, 2005.
31. Souza GM, Baker LL, Powers CM. Electromyographic activity of selected trunk muscles during dynamic spine stabilization exercises. Arch Phys Med Rehabil 82: 1551–1557, 2001.
32. Tarnanen SP, Ylinen JJ, Mälkiä EA, Siekkinen KM, Kautiainen HJ, Häkkinen AH. Effect of isometric upper-extremity exercises on the activation of core stabilizing muscles. Arch Phys Med Rehabil 89: 513–521, 2008.
33. Vezina MJ, Hubley-Kozey CL. Muscle activation in therapeutic exercises to improve trunk stability. Arch Phys Med Rehabil 81: 1370–1379, 2000.
34. Willardson JM. Core stability training: applications to sports conditioning programs. J Strength Cond Res 21: 979–985, 2007.
35. Zazulak B, Cholewicki J, Reeves NP. Neuromuscular control of trunk stability: Clinical implications for sports injury prevention. J Am Acad Orthop Surg 16: 497–505, 2008.

electromyography; abdominal muscles; back

© 2012 National Strength and Conditioning Association