Both the British and US Armies assess the endurance of military personnel’s abdominal and hip flexor muscles by counting the number of sit-ups that they can perform in 2 minutes. In both tests, personnel are instructed to raise and lower their trunk from the floor to the vertical, which involves both spine and hip flexion. Such sit-ups produce greater moments (19) and compressive forces (5) in the lower back when compared with curl-ups that involve flexion of the spine rather than the hips. Furthermore, although some studies (4,5) have reported that the magnitude of activity of the rectus abdominis (RA) is higher during sit-ups than during curl-ups, there are many (6,11,16,17,31) that have found activity to be similar or lower during sit-ups. When compared with sit-ups, curl-ups have also been shown to reduce activity in the hip flexors (11,12,16,19). Because of the lower compressive forces and muscle challenge that they provide, Axler and McGill (5) recommended curl-ups as exercises that may best suit individuals with low back pain.
Military personnel in both armies are instructed to maintain approximately 90° of knee flexion, and their ankles are held to ensure that the feet are in contact with the ground during the sit-ups. Since the first study by Walters and Partridge (35), numerous researchers have agreed that supporting the feet increases the activity of the hip flexors when compared with sit-ups or curl-ups that are performed with the feet unrestrained (4,16,26,30). Furthermore, Szasz et al. (34) reported that the amplitude of electrical activity in the RA and rectus femoris (RF) increased between the start, middle, and end of the US Army version of the test and that the increase was greater for the RF than for the RA.
High activation of the hip flexors, as has been evidenced in the type of sit-ups that are performed by the military personnel of both the British and US Armies, will likely result in high muscles forces that pull the pelvis into anterior tilt (21). The RF has a similar pattern of activation to the psoas during abdominal exercises (22), which is considered to exert large shear and compressive forces to the lumbar spine during activities such as sit-ups (8). Increased lumbar lordosis associated with anterior pelvic tilt can also lead to facet joint compression and shear to the pars interarticularis (8,33). Such forces can create vertebral rim strains, increased disc pressures and vertebral end plate deformation (14). Individuals who experience adverse mechanical loading of the intervertebral discs could, therefore, be more susceptible to adverse matrix changes and disc degeneration (32).
Because of the higher compression and shear forces that occur when performing sit-ups as opposed to other forms of abdominal exercise, they have been advised against for individuals with low back injury, lumbar instability, or weak abdominal muscles (5,11,12). Accordingly, the American College of Sports Medicine (3) and the Young Men’s Christian Association (YMCA) (36) both advise performing curl-ups as opposed to sit-ups. In comparison to the Army sit-up tests, the YMCA protocol assesses abdominal muscle endurance by counting the number of curl-ups that can be performed in one minute, and the American College of Sports Medicine test requires subjects to keep time with a metronome (40 b·min−1) and perform as many repetitions as possible (maximum 75).
Previous research outlined above has indicated that the type and duration of sit-ups used by the British and US Armies may provide a greater risk of low back problems than do different forms of sit-ups or curl-ups advocated by other organizations. Despite this, to the authors’ knowledge, only Szasz et al. (34) have investigated the changes in abdominal and hip flexor muscle activity during an armed forces sit-up test. Thus, we aimed to provide further information on how armed forces style sit-ups activate muscles and therefore load the musculoskeletal system in comparison with other forms of abdominal exercise, and how this changes with the duration of the exercise test. This new information would enhance the existing scientific evidence used to inform the recommendations made by such organizations regarding which form of exercise should be performed and for how long. Specifically, we did this by comparing the activity of 5 hip and trunk flexors between an army style sit-up and 4 commonly performed versions of a curl-up, and assessed the changes in muscle activity over a 2-minute exercise period. We hypothesized that hip flexor muscle activity would be greater in sit-ups and in exercises with the feet restrained, that abdominal muscle activity would be greater in exercises with the feet restrained, and that activity in hip flexors and abdominals would increase throughout the exercise duration.
Experimental Approach to the Problem
Surface electromyography was used to measure the muscle activity and, therefore, provide an indication of the loading on the musculoskeletal provided by a hip flexor and abdominal muscles. Muscle activity was compared between the British Army’s 2-minute sit-up section of its physical fitness test (PFT) and 4 variations of this that involved curl-ups instead of sit-ups (listed as exercises 1–5 below). Activation was also compared between the start, middle, and end of the 2-minute exercise period for each sit-up or curl-up. The order of exercises was randomized, and each testing session contained the performance of 1 of the 5 exercises, plus instruction of the technique required for the next exercise. Where possible, that is, where soldiers were not required to be on short-notice commitments, sessions took place at approximately the same time of the day to minimize diurnal variation. Sessions lasted approximately 30 minutes with a minimum of 48 hours of rest between each to reduce the effects of fatigue. The subjects were instructed to perform the same eating and drinking habits on the day of each test, and warm-up as per normal military PFT conditions.
Exercise 1 involved a bent knee sit-up as described in the Introduction and used in the British Army PFT (24) with feet anchored and arms crossed to touch the opposite shoulder (Figure 1A). The only variation was that the subject repeatedly raised and lowered their trunk from the floor to the vertical position at the prescribed cadence rather than at their preferred rate.
Exercise 2 involved a curl-up with feet anchored and arms crossed, performed to the equivalent of a 12-cm reach (Figure 1B). This distance was measured from the tip of the third finger and was based on that used in the Canadian Standard Test for Fitness (10,13). Before the test, the subjects performed the 12-cm reach and noted the position of the tops of their knees in relation to 3 horizontal lines that were drawn 10 cm apart on a screen in front on them.
Exercise 3 involved a curl-up as described for exercise 2 but with feet unanchored (Figure 1C).
Exercise 4 involved a curl-up as described for exercise 3, but with arms uncrossed and straight by side (Figure 1D). Rather than noting the position of their knees in relation to the 3 lines, as in exercises 2 and 3, a wooden baton was placed on a line on the ground 12 cm away from the tips of subjects’ middle fingers, to provide feedback.
Exercise 5 involved a curl-up as described for exercise 4, but with subjects instructed to perform lower abdominal hollowing (LAH). This exercise is designed to recruit the transversus abdominis (TrA) and the internal obliques (IO) and has been advocated in the clinical rehabilitation of patients with low back pain (28,29).
Twenty-three full-time serving male soldiers (mean ± SD age = 24.6 ± 5.9 years; height = 177.4 ± 6.1 cm; mass = 75.8 ± 7.6 kg) volunteered to take part in the study after responding to the advertisements in local military gymnasiums. All the subjects provided informed and written consent and were only included if they had no ongoing musculoskeletal injury or history of abdominal surgery or low back pain. Ethical approval was obtained from the Ministry of Defence (Navy) Personnel Research Ethics Committee and Manchester Metropolitan University’s Ethics Committee.
To ensure standardization, all the tests were performed to the sound of a metronome (Seiko SQ50: Thomann, Burgebrach, Germany). Pilot testing showed that the subjects (n = 4) were unable to attain the total score of 50 repetitions (mean = 42, SD = 4) in tests with feet unanchored, which is the score required to pass the full sit-up test. Thus, a metronome rate equivalent to a total of 40 repetitions (i.e., 20 reps·min−1) was selected for all the tests to allow comparison.
The subjects were instructed to start each repetition on the first beep of the metronome and to ensure that a beep was heard at the end of the upward phase of the movement and during body (scapulae) contact with the floor. The duration of a complete exercise, including raising and lowering, was 3 seconds. As with normal testing conditions, verbal time checks were provided to the subjects at 1 minute 30 seconds and at 1 minute 45 seconds. The count was recorded and compared with the data from an electrogoniometer (see below), attached to the subject to record hip angle, to ensure that the required rate of repetitions was performed.
The right hip joint angle of flexion and extension was measured during all the tests using an SG150 type electrogoniometer (Biometrics Ltd., Gwent, United Kingdom). Positioning was as in Figure 2, and the subjects reported no restriction of movement caused by its use.
Electromyographical signals from the upper RA (URA), lower RA (LRA), external oblique (EO), TrA/IO, and RF were detected using type SX230 sensors (Biometrics Ltd.). Each sensor consisted of 2 circular electrodes (10-mm diameter) with a fixed center-to-center distance of 20 mm. Signals were preamplified (gain = 1,000; input impedance >106 MΩ; bandwidth = 20–450 Hz; common mode rejection ratio >9 6dB; noise <5 μV) and sampled at 1,000 Hz. The sensors were connected to a junction box, which was linked to a base unit via a RS422 data transfer cable. The base unit was connected to a Latitude laptop PC (Dell, Ireland) using a USB cable.
Sensors were located in accordance with the recommendations listed in Table 1, fixed to the skin using T350 adhesive pads (Biometrics Ltd.) and further secured with micropore hypoallergenic tape (3M, Neuss, Germany). A reference electrode was attached over the ulnar styloid using a wrist strap.
To enable electromyograms (EMGs) to be compared between exercises, they were normalized. To facilitate this, EMGs from isometric maximal voluntary contractions (MVCs) were also recorded. The MVCs were performed in accordance with the recommendations listed in Table 2. Each MVC was held for 6 seconds once maximal effort was assessed by the principal investigator, typically within 1–3 attempts. The principal investigator provided resistance for all MVCs to ensure standardization of this process.
For each subject, three 6-second phases (2 full repetitions) were extracted from the raw EMGs from each exercise. These were labeled phase A (the third and fourth sit-ups), phase B (2 sit-ups at the midway stage), and phase C (the last 2 sit-ups of the test). The third and fourth sit-ups were chosen to allow for any adjustment to the rate of the metronome. Raw EMGs were full-wave rectified and integrated (iEMG) over the entire duration of each phase using DataLINK Analysis software PC version 5.05 (Biometrics Ltd.). This form of processing was preferred to, for example, peak activity because it includes all the recorded electrical activity generated by the muscle during the exercise. The iEMGs were then normalized by expressing them as a percentage of the iEMG from the isometric MVC of the same muscle, also over 6 seconds.
All data were analyzed using the SPSS Version 16.0 for Windows with an alpha (α) level set at p ≤ 0.05. The iEMG from each muscle and hip joint range of motion (ROM) was analyzed using factorial analyses of variance (ANOVAs; 5 tests × 3 phases) with a Bonferroni post hoc test (25). The Shapiro-Wilk test was significant for 60% of the data sets. Thus, all data were log transformed (log10) before being analyzed using the ANOVAs. Violations of the assumption of sphericity were corrected for using the Huynh-Feldt Adjustment. Based on 23 subjects, the statistical power would be 0.19, 0.67, 0.97, and 0.99 for effect sizes of 0.25, 0.50, 0.75, and 1.00.
Hip Joint Range of Motion
The mean ROM (Table 3) was significantly different between exercise type (F3.21, 70.7 = 54.7, p < 0.001), with exercise 1 (sit-up) having a greater ROM than the other exercises (all curl-ups). Despite the subjects being instructed to perform each of the curl-ups over the equivalent of a 12-cm reach, the effect of anchoring the feet resulted in a significantly greater ROM for exercise 2 in comparison with exercises 3–5.
No significant difference (F1.40, 30.8 = 2.18, p = 0.143) existed in the ROM between the 3 phases of all the exercises. However, a significant interaction (F5.29, 116 = 6.56, p < 0.001) occurred between exercise type and phase, with the ROM for exercise 2 (curl-up with feet restrained) increasing from phase A to phase C.
Differences Between Exercises
The iEMG from all muscles was significantly different between exercises (see Table 4 for statistical information and specific comparisons). The URA and LRA and the EO and TrA/IO were generally more active in exercise 2 than in the other exercises. In addition, exercise 3 (curl-up with arms crossed) required significantly more URA activity than did exercises 4 and 5 (curl-ups with arms straight by sides; Figures 3–7). The RF was significantly more active in exercise 1 than in the other exercises and more active in exercise 2 than in exercises 3–5 (Figure 7). Despite the RF being relatively inactive in exercise 3 (<10% MVC), it was significantly more active than in the other curl-ups where the feet were unanchored (exercises 4 and 5).
Differences Between Phases
The iEMG from all the muscles was significantly different between phases (Table 4). Figures 3–7 show that all muscles displayed more activity in the middle (phase B) than at the start (phase A) of the exercise duration, and most muscles were more active at the end of the duration (phase C) than in the middle (phase B).
Interaction Between Exercises and Phases
Significant interactions between exercise and phase existed for all the muscles analyzed (Figures 3–7 and Table 4). For all 4 abdominal muscles, activity generally increased in a linear fashion between phases in all but 1 of the exercises. In exercise 3, the increase in activity was either not as great between phases B and C as it was between phases A and B (URA and EO), or the activity actually decreased between phases B and C (LRA and TrA/IO). Activation of the RF increased in a linear fashion across phases in exercises 1 and 2 but remained fairly constant across phases for the other 3 exercises.
This study aimed to compare the magnitude of muscle activation, as an indication of muscle loading on the skeletal system between the sit-up exercise used by the British Army and 4 variations of this. We also aimed to assess how the muscle activity altered over the 2-minute duration of each exercise type. Abdominal muscles (URA, LRA, EO, and TrA/IO) were generally found to be most active during curl-ups in which the feet were restrained. Activity in the abdominal muscles was also similar in sit-ups and curl-ups in which the feet were unrestrained. Greatest activity in the RF, which was used to represent the hip flexors (5), occurred during exercises where the feet were restrained, with the sit-up being significantly more active than the curl-up in this case. The activity of all abdominal muscles increased after 1 minute of exercise in all 5 variations of the sit-up and curl-up. However, a significant interaction was caused by the activity in the LRA and TrA/IO decreasing between phases B and C, whereas activity in other abdominals increased over time. Activity in the RF also increased with duration, but only for exercises 1 and 2.
The 5 exercises needed to be performed on separate days to avoid the effects of fatigue. Because of the existing high number of visits to the laboratory, it was considered impractical to perform a reliability analysis, although this would have been desirable. However, to enable muscle activity to be compared between exercises, and between muscles, the EMGs recorded during the exercises were normalized by expressing them as a percentage of those recorded during MVCs. Although this is a method that is advocated by electromyographers (9), it is recognized that not all subjects will be able to maximally activate their muscles during MVCs (1). This was indeed the case for 3 of the muscles in this study, as the mean + SD for the normalized EMGs exceeded 100% (Figures 3, 5, and 7). However, this was mostly the case only at the end of the 2-minute exercise period when the muscles would likely have been close to if not maximally activated. This, and the established protocols used for performing MVCs (Table 2) and the practice period allowed gives the authors’ confidence that comparisons could be made between exercises. Moreover, EMGs were recorded throughout the duration of each exercise without the removal of electrodes. Thus, the limitation described above does not apply to comparisons made between phases of each exercise.
Apart from the EO during phase A, the abdominals were more active during the curl-up with feet restrained (exercise 2) than the sit-up (exercise 1). These findings agree with the limited amount of research (5) that has compared these exercises for the EO, but not for the URA. We also found that for the abdominal muscles, feet-restrained curl-ups exhibited greater activity than the other (nonrestrained) curl-ups for at least 2 of the 3 measurement points (i.e., start, middle, or end of the 2-minute duration). Despite these differences being only statistically significant for the TrA/IO, they contrast with those of Parfrey et al. (30) who reported generally greater URA and LRA EMGs for nonfixated curl-ups. Nevertheless, our finding that L/URA activity is generally lower during sit-ups than during nonrestrained curl-ups (exercise 3) does agree with other research that has compared muscle activation between these forms of exercise (6,11,16,17).
In the majority of cases, abdominal activity was reduced when, for curl-ups, the position of the arms was changed from ‘crossed’ to ‘straight by sides.’ The change in arm position (i.e., between exercise 3 and exercises 4 and 5) shifted the center of gravity of the upper body toward the hips and, therefore, required less muscular torque to complete the exercise. Because the moment arms of the abdominal muscles would not have changed, this would explain the lower neural drive recorded with the arms by the side.
As expected, instructing the subjects to hollow their abdominals during curl-ups had the desired effect of increasing activity of the TrA/IO, although this difference was not as statistically significant as it was in the EO. Manual muscle testing was used to ensure both the correct placement of electrodes and that minimal crosstalk existed between the EO and TrA/IO electrode sites. Thus, increased activation of the EO during exercise 5 implies that the subjects may have performed more of an abdominal bracing than hollowing maneuver (2). Nevertheless, increased activation of these muscles with little or no effect on the L/URA suggests that the motor control of the spine could be improved in curl-ups with LAH. Despite not being significant, these findings reinforce the recommendations of O’Sullivan et al. (28,29) that LAH should be used in the rehabilitation of patients with chronic low back pain.
As previously stated, the exercises that best recruited the abdominal muscles were those that restrained the feet. However, these exercises also resulted in the greatest activity in the RF. Our findings agree with those of previous research that has compared abdominal exercises with feet restrained and not restrained (4,16,27,30,35). The greater RF activity in the sit-up than the feet-restrained curl-up (mean across phases and subjects = 46 and 32% MVC for exercises 1 and 2, respectively) also serves to reinforce previous research that has warned against using sit-ups for individuals with low back injury, lumbar instability, or weak abdominal muscles (5,11,12). Moreover, high activation of the RF during feet-restrained curl-ups relative to curl-ups without restraint (mean for exercises 2, 3, 4, and 5 = 32, 5, 4, and 4% MVC) implies that large and potentially injurious compressive and shear forces would also be experienced during such exercises (8). Despite attempts to standardize the ROM, it is possible that the difference in RF activity between foot-restrained and nonrestrained curl-ups could partly be because of the greater hip ROM recorded with the feet restrained (mean difference of 8, 15, and 20° for phases A, B, and C). Some of this increased ROM would have been brought about by active hip flexion, in addition to spinal flexion, which would have required increased activation of the hip flexors. However, this limitation serves to demonstrate that restraining the feet during a curl-up will invariably result in undesired hip flexion and increased activity of the hip flexors. Thus, feet-restrained curl-ups should be added to the list of exercises that should not be performed by individuals at risk of injury.
For each exercise, repetitions of trunk and hip flexion and extension were performed over the same ROM and for the same duration, as set by the metronome. Thus, the pattern of torque required to perform each repetition would likely have been the same at the beginning, middle, and end of the 2-minute exercise period, and during repetitions in between. Assuming that the change in muscle moment arms during each repetition was the same across the exercise period, the pattern of force required by the abdominal and hip flexor muscles would also have remained the same at 0, 1, and 2 minutes. Despite this, the EMG amplitude increased for most muscles between the start and the end of most exercise variations. Increased neural drive to the muscle has previously been reported for sustained submaximal contractions (15), although these have been isometric rather than nonisometric as in our study. This increase in muscle activity is generally thought to represent the recruitment of new motor units to supplement the reduction in force because of impaired excitation-contraction coupling of previously recruited units (7).
Curl-ups with neither feet restrained nor LAH produced the lowest increases in muscle activity across phases (mean for all muscles for exercise 3 = 8% and exercise 4 = 13%). The RF showed the largest increase in activity of all the muscles analyzed. Not surprisingly, this occurred only during exercises that involved the feet being restrained, where its activity was also highest and, therefore, fatigue likely to be greatest. Our findings are in agreement with those of Szasz et al. (34), who reported that the increase in RF activity was greater than that of the RA over the 2-minute duration of the foot-restrained sit-up used by the US Army. Moreover, the increases that were observed in our study for the sit-up (21 and 29% between phases A and B, and B and C, respectively) were surpassed by those seen during the foot-restrained curl-up (49 and 41%).
As discussed above, such increases likely occurred to maintain the muscle force required to perform the exercises as subjects’ muscles became fatigued. Coupled with the fact that the exercise rate and ROM remained the same, it is likely that the higher compressive and shear forces that occur during feet-restrained abdominal exercises (5) are also maintained throughout the period of exercise. The same cannot necessarily be said for sit-ups that conform to the protocol required by the British or US Army, which are not necessarily performed at a standard rate. Such exercises would, undoubtedly at times, be performed at a greater rate than that set, for example, by a metronome. Higher angular accelerations of the trunk and hip would require larger forces in the trunk and hip flexors and, most likely, increased compressive and shear forces in the lower back. Should organizations continue to assess the endurance of abdominal muscles by counting the number of completed feet retrained sit-ups in a given time period, as is currently the case in the US and British Army, further research should be carried out into the effect of cadence on the biomechanics of the exercise and the implications that this may have on forces in the lower back and the risk of injury.
Abdominal muscles were most activated during curl-up exercises in which the feet were restrained. Thus, individuals who are seeking a higher muscle challenge should adopt this form of exercise. However, restraint of the feet during curl-ups and sit-ups activates the hip flexors (RF in this study) to significantly greater levels than observed in curl-ups without restraint of the feet. Previous research has indicated that high activation of the hip flexors increases compressive and shear forces in the lower back. Thus, in agreement with previous recommendations, individuals with low back injury, lumbar instability, or weak abdominal muscles should avoid abdominal exercises in which the feet are restrained (5,11,12). Instead, such individuals should perform curl-ups with the feet unrestrained, after appropriate assessment and progression of motor control performance for this level of exercise. These exercises activate the abdominals to levels that are similar to those experienced in feet-restrained sit-ups, as used by the British Army.
Activity of the abdominals and hip flexor increased between the start, middle, and end of the 2-minute duration of most exercises. Based on previous research that has investigated isometric contractions, this is likely to be required to compensate for the reduction in force provided by motor units that have already been recruited. Thus, for exercises that are performed at a constant rate, as in this study, muscle force and possibly compressive and shear forces are unlikely to increase with exercise duration. Further research is required to investigate the pattern of muscle activity and forces in abdominal exercises where the rate is not controlled, as used by the British and US Armies.
1. Allen GM, Gandevia SC, McKenzie DK. Reliability of measurements of muscle strength and voluntary activation using twitch interpolation. Muscle Nerve 18: 593–600, 1995.
2. Allison GT, Godfrey P, Robinson G. EMG signal amplitude assessment during abdominal bracing and hollowing. J Electromyogr Kinesiol 8: 51–57, 1998.
3. American College of Sports Medicine. ACSM’s Guidelines for Exercise Testing and Prescription. Philadelphia, PA: Lippincott, Williams & Wilkins, 2009.
4. Anderson EA, Nilsson J, Ma Z, Thorstensson A. Abdominal and hip flexor muscle activation during various training exercises. Eur J Appl Physiol 75: 115–123, 1997.
5. Axler CT, McGill SM. Low back loads over a variety of abdominal exercises: Searching for the safest abdominal challenge. Med Sci Sports Exerc 29: 804–811, 1997.
6. Beim GM, Giraldo JL, Pincivero DM, Borror MJ, Fu FH. Abdominal strengthening exercises: A comparative EMG study. J Sport Rehabil 6: 11–20, 1997.
7. Bigland-Ritchie B, Cafarelli E, Vøllestad NK. Fatigue
of submaximal static contractions. Acta Physiol Scand Suppl 128: 137–148, 1986.
8. Bogduk N, Pearcy M, Hadfield G. Anatomy and biomechanics of psoas major. Clin Biomech 7: 109–119, 1992.
9. Burden AM. How should we normalize electromyograms obtained from healthy participants? What we have learned from over 25 years of research. J Electromyogr Kinesiol 20: 1023–1035, 2010.
10. Cole B, Finch E, Gowland C, Mayo N. Physical Rehabilitation Outcome Measures. Baltimore, MD: Williams & Wilkins, 1995.
11. Escamilla RF, Babb E, DeWitt R, Jew P, Kelleher P, Burnham T, Busch J, D'Anna K, Mowbray R, Imamura RT. Electromyographic analysis of traditional and non-traditional abdominal exercises: Implications for rehabilitation and training. Phys Ther 86: 656–671, 2006.
12. Escamilla RF, McTaggart MSC, Fricklas EJ, DeWitt R, Kelleher P, Taylor MK, Hreljac A, Moorman CT. An electromyographic analysis of commercial and common abdominal exercises: Implications for rehabilitation and training. J Orthop Sports Phys Ther 36: 45–57, 2006.
13. Faulkner RA, Springings EJ, McQuarrie A, Bell RD. A partial curl-up protocol for adults based on an analysis of two procedures. Can J Sports Sci 14: 135–141, 1989.
14. Frei H, Oxland TR, Nolte LP. Thoracolumbar spine mechanics contrasted under compression and shear loading. J Orthop Res 20: 1333–1338, 2002.
15. Fuglevand AJ, Zackowski KM, Huey KA, Enoka RM. Impairment of neuromuscular propagation during human fatiguing contractions at submaximal forces. J Physiol 460: 549–572, 1993.
16. Guimaraes AC, Vaz MA, De Campos MI, Marantes R. The contribution of the rectus abdominis and rectus femoris in twelve selected abdominal exercises. An electromyographic study. J Sports Med Physical Fitness 31: 222–230, 1991.
17. Halpern AA, Bleck EE. Sit-up exercises: An electromyographic study. Clin Orthop Relat Res 145: 172–178, 1979.
18. Hermens HJ, Freriks B, Merletti R, Stegeman D, Blok J, Rau G, Disselhorst-Klug C, Hägg G, eds. European Recommendations for Surface Electromyography: Results of the SENIAM Project. Enschede, Netherlands: Roessingh Research and Development, 1999.
19. Juker D, McGill S, Kropf P, Steffen T. Quantitative intramuscular myoelectric activity of lumbar portions of psoas and abdominal wall during a wide variety of tasks. Med Sci Sports Exerc 30: 301–302, 1998.
20. Lehman GJ, McGill S. Quantification of the differences in electromyographic activity magnitude between the upper and lower portions of rectus abdominis muscle during selected trunk exercises. Phys Ther 81: 1096–1101, 2001.
21. Levangie CC, Norkin PK. Joint Structure and Function: A Comprehensive Analysis (4th ed.). Philadelphia, PA: FA Davis Company, 2005.
22. McGill S, Juker D, Kropf P. Appropriately placed surface EMG electrodes reflect deep muscle activity (psoas, quadratus lumborum, abdominal wall) in the lumbar spine. J Biomech 29: 1503–1507, 1996.
23. Miller MI, Medeiros JM. Recruitment of internal oblique and transversus abdominis muscles during the eccentric phase of the curl-up exercise. Phys Ther 67: 1213–1217, 1987.
24. Ministry of Defence (MOD). Military Annual Training Tests (MATT2). London, United Kingdom: MOD, 2006.
25. Munro BH. Statistical Methods for Health Care Research. Philadelphia, PA: Lippincott, Williams & Wilkins, 2001.
26. Ng JK, Kippers V, Richardson CA. Muscle fibre orientation of abdominal muscles and suggested surface EMG electrode positions. Electromyogr Clin Neurophysiol 38: 51–58, 1998.
27. Norris CM. Abdominal muscle training in sport. Br J Sports Med 27: 19–27, 1993.
28. O’Sullivan P, Twomey L, Allison GT. Altered abdominal muscle recruitment in patients with chronic back pain following a specific exercise intervention. J Orthop Sports Phys Ther 27: 114–124, 1998.
29. O’Sullivan P, Twomey L, Allison G, Sinclair J, Miller K, Knox J. Altered pattern of abdominal muscle activation in patients with chronic low back pain. Aust J Physiother 43: 91–98, 1997.
30. Parfrey KC, Docherty D, Workman RC, Behm DG. The effects of different sit- and curl-up positions on activation of abdominal and hip flexor musculature. Appl Physiol Nutr Metab 33: 888–895, 2008.
31. Partridge MJ, Walters CE. Participation of the abdominal muscles in various movements of the trunk in man. An electromyographic study. Phys Ther Rev 39: 791–800, 1959.
32. Roughley PJ. Biology of intervertebral disc aging and degeneration. Spine 19: 2691–2699, 2004.
33. Sahrmann SA. Diagnosis and Treatment of Movement Impairment Syndromes. St. Louis, MO: Mosby, 2002.
34. Szasz A, Zimmerman A, Frey E, Brady D, Spaletta R. An electromyographical evaluation of the validity of the two-minute sit-up section of the Army Physical Fitness Test in measuring abdominal strength and endurance. Mil Med 167: 950–953, 2002.
35. Walters CE, Partridge MJ. Electromyographic study of the differential action of the abdominal muscles during exercise. Am J Phys Med 36: 259–268, 1957.
36. YMCA of the USA. Fitness Testing and Assessment Manual. Champaign, IL: Human Kinetics, 2000.