- Lid geworden
- 7 okt 2002
- Berichten
- 22.009
- Waardering
- 3.621
The Journal of Strength and Conditioning Research: Vol. 17, No. 3, pp. 475–483.
Electromyographic Comparison of the Upper and Lower Rectus Abdominis During Abdominal Exercises
Kathryn M. Clark, Laurence E. Holt, and Joy Sinyard
The School of Health and Human Performance, Dalhousie University, Halifax, Nova Scotia, Canada B3H 3J5
Introduction
The abdominal musculature plays a key role in the movement and stability of the lumbar spine. When the muscles, which contribute to the stability of this inherently unstable area, are compromised and their stabilizing role is negatively influenced, there are significant implications with regard to the development of low back pain (23). For this reason, many health professionals advocate abdominal
musculature exercises for the prevention or rehabilitation from back pain (1, 11, 15, 19, 20, 24). Abdominal exercises are also popular in fitness training programs and are important for athletic training (13, 19). In a survey conducted at San Jose State University in California, aerobics participants ranked the abdominal musculature as the most important body part to train (15).
While most researchers and fitness professionals are aware that the rectus abdominis (RA) is one muscle, it is general practice to indicate to clientele that they are indeed “working the upper or lower abs” while performing certain exercises. As a result, clients are led to believe they are recruiting more motor units from one portion of the RA than the other or that the upper fibers of the RA (URA) and the
lower fibers of the RA (LRA) are actually separate muscles. Although the RA has been studied extensively using surface electromyography (EMG), most studies do not directly compare the URA to the LRA (2, 4, 6, 10, 23, 24).
Whether or not exercise type can differentially influence the upper and lower fibers of the RA has consistently been an area of debate among clinicians, researchers, and fitness professionals. Allusions have been made in the literature that the different segments of the RA offer different stabilizing potentials (22). Among the studies that do compare the 2 muscle sites (8, 11, 16, 20), the differences in the exercises used and the methods of collecting, analyzing, and reporting the data make it difficult to see any similarities between the 2 sites of interest or to draw any significant conclusions. In many of the studies in which the authors have seen differences in electrical activity between the URA and the LRA, there is no agreement as to where and how these variations occur (2, 3, 8–11, 16, 20). Many
studies have found the curl up to elicit higher EMG activity from the URA than the LRA (16, 20, 24), whereas the reverse curl up and double leg raise have been touted as primarily LRA exercises (5, 11, 20). Other studies have found no significant difference in EMG activity between the 2 muscle sites (2, 9).
It was thought there might be a difference in response to the exercises from the URA and the LRA due to the segmental innervation of the anterolateral abdominal musculature from the ventral rami of the lower 6 or 7 thoracic nerves. The segmentation of the RA by its tendinous inscriptions also implies the ability to selectively isolate different portions of this muscle (20).
In contrast, the RA is a strap-like muscle with its insertions on the coastal cartilages of the 5th, 6th, and 7th ribs and the xiphoid process superiorly and to the crest of the pubis, the pubic tubercle, and the front of the pubis symphysis inferiorly. Some fibers have also been found to extend beyond the limits of the segments, crossing outside the tendinous inscriptions (12). From a biomechanical
perspective, this evidence suggests that the RA might not have the ability to contract more forcibly at one end than the other.
Methods
Experimental Approach to the Problem
This study looked at 6 RA exercise protocols using surface EMG. Our purpose in this study was to determine the effect of 4 traditional and 2 new abdominal exercises on the URA and LRA as well as to determine which of the 6 exercises elicited the highest electrical response from the RA musculature. Four of the exercises, i.e., the curl up, the reverse curl up, the Ab Trainer Curl up and the leg
lowering exercises, replicate previous research (1, 4, 8, 11, 14, 19, 20, 23, 24). In addition, we have tested the Sissel ball curl up and Sissel ball roll out using a stability ball, a piece of equipment used frequently in both the exercise and rehabilitative industries, which has not previously been studied in this manner.
Surface EMG was chosen for its accessibility and noninvasive nature. As the abdominal muscles are organized in layers, crosstalk from nearby or underlying muscles, which may confound the validity of surface EMG signals, becomes an important issue. A thorough review of the anatomy and fiber orientation of the anterolateral abdominal musculature demonstrates areas where the abdominals may
be recorded that will minimize crosstalk. Careful placement of electrodes is also a key element in minimizing crosstalk effects (7, 18).
The bipolar electrode configuration chosen for this study is almost always used in order to eliminate movement, electrocardiogram, and 60-cycle artifacts (7, 21). In this configuration, two electrodes are placed over the site of interest 2 cm apart, and a third, common-mode reference, or ground, electrode is placed away from the testing site in an electrically neutral environment, usually on a bony prominence
and in this case, on the right anterior superior iliac spine of the pelvis. This configuration also allows the signal to be specific to the sites in question. As the 2 electrodes get farther apart, the signal has been shown to become more regional and less specific (7).
Subjects
Eight volunteer subjects, 4 men and 4 women, participated in this study. All subjects were apparently healthy adults with no history of low back pain or abdominal surgery, as determined from a verbal questionnaire given prior to testing. All subjects had extensive experience performing multiple abdominal exercise protocols and were familiar with the equipment. Subjects were participants in structured physical activity at Dalhousie University, were physically active an average of 5 times a week, and trained their abdominal musculature an average of 3 times a week. The mean age of the subjects was 31.75 ± 8.63 years. Their mean weight was 75.41 ± 12.77 kg
and their mean height was 174.34 ± 9.34 cm. Before participation, all subjects provided written informed consent. This study received ethical approval from a Dalhousie University committee prior to data collection.
Experimental Design
To measure the action potentials of the URA and LRA, 4 surface electrodes were placed on the abdominal musculature of the subjects.
The muscle and ground electrode sites were prepared according to Cram et al. (7) and electrodes were applied according to Ng et al. (18) (Figure 1 ).
Two sites on the RA were tested using Hewlett Packard pregelled, disposable Ag/AgCl electrodes. The URA site was 3 cm above and 2 cm to the right of the umbilicus, over the muscle belly, and parallel to the muscle fibers of the URA. The LRA site was 3 cm below and 2 cm to the right of the umbilicus on an 8° inferomedial angle, over the muscle belly, and parallel to the muscle fibers of the lower portion of the RA (Figure 1 ) (18).
The raw EMG signal was amplified 1,000 times using two 2-channel bioamps. The signal was then passed through an EMG processor, where it was again amplified 5 times and filtered to eliminate artifacts that would otherwise alter the EMG signal; the band pass was set at 20–1,000 Hz. This filtered signal was full-wave rectified and linear enveloped. The EMG signal was sampled using Lab
View, Gen99 software (National Instruments, Austin, TX). A 12-bit National Instruments Card analogue to digital (A/D) converter was used to convert the signal to a digital format. The signal was then converted to volts and the root mean square (RMS) calculated.
Reference Contractions
The utility of surface EMG has often been questioned in the field of biomechanics (8, 17). Because of differences in subcutaneous fat, muscle geometry, skin impedance, trunk flexibility, age, coïtus, and other variables, absolute EMG amplitudes are not comparable between individuals, activities, or muscles. During dynamic movements in particular, comparison of amplitude measurements across muscle groups can be quite misleading. In order to circumvent this problem and to be able to make statements about the relative intensity of an EMG signal, it is necessary to normalize the EMG data (7, 17, 21).
The most common method of normalization is to have subjects perform 1 maximal voluntary contraction (MVC) as a reference. The EMG values subsequently obtained are then presented as a percentage of this MVC (21). Mirka (17) has shown that errors occur as a result of normalization to reference values collected from only 1 position for various muscles of the trunk. The study suggests that relating EMG recorded during a dynamic task to a reference value taken at one joint angle will render inaccurate results due to changes in the portion of the muscle within the viewing area of the electrode, changes in length/tension relationship of the muscle, and changes in the moment arm of the muscle around the joint of rotation. Mirka (17) proposes that, in order to be able to quantitatively compare the relative magnitude of the EMG signals of a dynamic movement under different joint angles, the maximum EMG signal used to normalize the data should be taken at several joint angles throughout the range of motion (ROM).
In order to ensure accuracy, we had each subject perform 4 different reference contractions in randomized order and normalization of their respective EMG activity was based on the average of these 4 reference contractions.
Each subject was initially asked to perform 1 curl up (CU) (Figure 3a ) in order to determine his or her normal range of motion (ROM) during cervical and thoracic spinal flexion.
Three isometric maximal voluntary contractions (MVC) were measured while the subject pushed against the resistance of a nylon strap, placed across the torso and arms just below the clavicle, which anchored them to the plinth. The subjects' arms and hands were placed across the chest, on top of the nylon strap in order to facilitate a maximal contraction and to prevent the strap from slipping. The
MVCs were measured at (a) 0° to the horizontal in the position of initiating a CU, (b) the halfway point of the subjects' measured ROM for the CU (15–20° to the horizontal), and (c) the subjects' measured full ROM for the CU (30–40° to the horizontal). The subjects were asked to push as hard as they could against the strap for 4 seconds (Figure 2a ).
A fourth, dynamic MVC was measured using the concentric phase of the CU. The subjects, no longer anchored to the plinth, were asked to push against a force manually applied to his or her shoulders while they were allowed to move through their full ROM over 4 seconds; arms and hands were again placed across the chest (Figure 2b ).
Subjects performed 3 trials each of the 4 MVCs and were required to take a 2-minute rest between trials. Angles were measured using a goniometer attached to the plinth at the level of the superior aspect of the iliac crest.
The RMS measures of each of the 3 MVC trials were visually inspected in order to find the highest average EMG activity over a 0.5-second time interval. Intraclass correlation coefficients (R) were calculated within each reference contraction across the 3 trials to determine the reliability of the MVC values. As a result of high R and p values indicating they were not significantly different, the 4 MVCs
were then averaged and applied as a single reference contraction (AMVC) for normalization (Table 1 ).
Exercises
The subjects were asked to perform 3 trials each of 6 randomized abdominal exercises. Each trial was performed continuously over 6 seconds with a 2-second concentric phase, a 2-second isometric phase, and a 2-second eccentric phase. All of the exercises and the concentric/isometric/eccentric contraction pattern are very common to general fitness classes and muscular conditioning programs
practiced in Halifax, Nova Scotia, university athletic centers. The trials were timed using a stopwatch, by a certified fitness instructor who counted out loud to the subjects. This protocol was selected because of its similarity to the timing used during many group fitness classes. A VHS video recorder was used to videotape all test sessions in order to confirm the timing of the 3 phases of the exercises.
No warm up was permitted and no more than 2 practice trials were allowed before each set of 3 test trials. A 1-minute rest was required between every trial.
The curl up (CU) (Figure 3a ) was performed lying supine on a plinth with both feet flat on the plinth and an included knee angle of 90°. The fingertips were placed at the base of the skull to support the head, with the elbows wide and lateral to the head. The subjects were instructed to lift the head and shoulders toward the ceiling over 2 seconds, to hold the new position for 2 seconds, and to slowly
return to the starting position over 2 seconds.
The Sissel ball curl up (BCU) (Figure 3b ) was performed lying supine on an inflatable, rubber Sissel ball, 55 cm in diameter. The fingertips were placed at the base of the skull, the elbows wide and lateral to the head. The subjects were instructed to keep both feet flat on the floor and to position themselves so their lower back was supported by the ball, their torso parallel to the floor. Correct positioning was accomplished by subjective feedback from a fitness professional. Once the starting position was achieved, they were then instructed to lift the head and shoulders toward the ceiling over 2 seconds, to hold the new position for 2 seconds, and to slowly return to the starting position over 2 seconds.
The Ab Trainer curl up (ATCU) (Figure 3c ) was performed using the Ab Trainer abdominal training device. This device consists of tubular metal shaped to roll forward and backward as the person performs a CU. The device extends above the exerciser and includes armrests and a headrest. The exercise was performed lying supine on a plinth with both feet flat on the plinth and an included knee
angle of 90°. The elbows rested anterolaterally to the head on the armrests and in this case, the headrest supported the head. The subjects were instructed to lift the head and shoulders toward the ceiling over 2 seconds, to hold the new position for 2 seconds, and to slowly return to the starting position over 2 seconds. They were also asked to maintain head contact with the headrest of the device.
The leg lowering exercise (LL) (Figure 3d ) was performed lying supine on a plinth with the legs in the air. In the starting position, the hips were flexed at a 90° included angle and the knees were flexed at a 45° included angle. The fingertips were placed at the base of the skull, the elbows wide and lateral to the head. The upper torso, arms, and head remained in contact with the plinth for the duration
of the exercises. The subjects were instructed to slowly lower their legs over 2 seconds so their heels were lightly touching, but not resting on, the plinth, while maintaining both the 45° angle at the knee joint and lower back contact with the plinth. They were then instructed to hold the lowered position for 2 seconds and to slowly return to the starting position over 2 seconds. The sequence of contraction for the LL was reversed in that the eccentric phase was performed first followed by the isometric and concentric phases.
The Sissel ball roll out (BRO) (Figure 3e ) was performed kneeling on the floor with a small mat under the knees for comfort. The forearms were placed on a 55-cm Sissel ball and the subject was instructed to roll out over 2 seconds into a fully extended position (i.e., the hips were fully extended and the shoulders were in a flexed position next to the head). They were then asked to hold the extended
position for 2 seconds and to slowly return to the starting position over 2 seconds. The sequence of contraction for the BRO was reversed in that the eccentric phase was performed first followed by the isometric and concentric phases, as in the LL exercise.
The reverse curl up (RCU) (Figure 3f ) was performed lying supine on a plinth with the legs in the air. The hips were flexed at a 90° included angle and the knees were flexed at a 45° included angle. The fingertips were placed at the base of the skull, the elbows wide and lateral to the head. The upper torso, arms, and head remained in contact with the plinth for the duration of the exercise. The subjects
were instructed to curl the hips toward the ribs (i.e., posteriorly tilt the pelvis) over 2 seconds while maintaining the 45° angle at the knees, to hold the new position for 2 seconds, and to slowly lower the hips back to the starting position over 2 seconds.
The protocol chosen for each exercise was in accordance with trends in the general fitness field. It is common to rest the head in the fingertips during many exercises in order to provide a longer lever and thereby higher intensity for the abdominal musculature during the
curl up. Using only the fingertips allows the participant to rest the head and neck, minimizing fatigue of the anterior neck musculature while performing repetitive abdominal exercises and cues participants not to strain the head and neck into a more forward position.
Using the whole hand or lacing the fingers behind the head encourages participants to use the hands and arms to forcibly pull on the head and neck. Placing the arms and hands across the chest is less intense for the abdominals, as the weight of the arms is now closer to the subjects' center of gravity. This position may also place more strain on the muscles in the anterior neck, frequently causing neck
discomfort and an early termination of abdominal exercises before fatigue of the anterior abdominal musculature is achieved.
The knee and hip joints are flexed to shorten the hip flexors and minimize their contribution to the movement in order to better isolate the anterior abdominal musculature and to reduce stress on the lumbar spine by the pulling of the psoas insertions on the lumbar vertebrae.
Exercise measurements were sectioned into 3 phases by a synch pulse that was recorded along with the EMG in order that the concentric data be easily separated from the isometric and eccentric for analysis. The pulse was operated by an assistant who turned the pulse on when the subject began to visibly move, indicating the beginning of the first phase, the concentric phase for the CU, BCU,
ATCU and RCU, and the eccentric phase for the LL and the BRO. The pulse was turned off when the subject ceased to move, indicating the beginning of the second, isometric phase, and turned on once again as the subject began to move, indicating the beginning of the third phase, the eccentric phase for the CU, BCU, ATCU, RCU, and the concentric phase for the LL and the BRO. The pulse was turned
off at the end of the third and final phase.
The RMS measures of each of the 3 test trials were sectioned in order to find the average EMG activity over the 2-second concentric phase. R was calculated across the 3 trials for each exercise's concentric phase to determine trial reliability. As a result of high R values, the 3 trials were averaged and this value used for normalizing as a percentage of each subjects' MVC for each muscle site.
These values were then averaged across the subjects for comparison (%AMVC).
Statistical Analyses
General linear model analyses of variance (ANOVA) were performed to assess the differences in %AMVC of the URA and the LRA due to the exercise conditions and phase of movement. Post hoc analysis was performed using Tukey's method for between-exercise and between-phase comparisons. The level for statistical significance was set at the 95% confidence limit (? = 0.05). All statistical analyses were performed using the Minitab statistical analysis package.
Results
High R values verified that the 4 reference contractions were reliable among the trials as well as among 3 of the 4 MVCs (Table 2 ).
During the isometric MVC at full ROM, R values dropped off considerably; however, because the MVC was not found to be significantly different from the others (p = 0.12), it was therefore included in the AMVC. A high R value (R = 0.96) also verified that the %AMVC for all exercises performed were reliable among the trials and they were therefore averaged.
The data from the concentric phase was chosen for analysis because it produced the highest EMG activity overall and is a dynamic contraction over the subjects' full ROM.
Tukey simultaneous tests performed among the phases found the average %AMVC values for all subjects during the concentric phase to be significantly higher than the eccentric phase (p < 0.05) and not to differ significantly from the isometric phase (p = 0.24).
The lower %AMVC recorded from the subjects during the eccentric phase of the CU, BCU, ATCU, and RCU (Figure 4 ) is the expected result of a decrease in the firing of motor units as the subject moves with gravity, not against it, as during the concentric and isometric phases (3). The %AMVC of the eccentric phase of the LL and the BRO (Figure 4 ) are closer in amplitude to their respective concentric phases. This is expected due to the increasing lever length as the subject lowers his or her legs to touch the plinth during the LL and rolls out into an extended position on the Sissel ball during the BRO. More force must be applied in these situations in order to control the longer lever, thereby increasing the firing of motor units (7). These results serve as another indicator that the signals
from the surface electrodes were indeed reliable.
In comparing the %AMVC of the concentric phases of the 6 exercises, the results of the ANOVA revealed there were no statistically significant effects due to muscle site (p = 0.081), interactions between subject and muscle site (p = 0.27), or exercise and muscle site (p = 0.10). Significant differences in the %AMVC were found in both the URA and the LRA due to the exercise performed (p < 0.00) and
the interaction between subject and exercise performed (p = 0.003). A 1-way ANOVA performed on the %AMVC values of the URA and the LRA during the concentric phase also revealed there was no significant difference between the 2 muscle sites (p = 0.116).
In comparing the %AMVC of the concentric phase of the 6 exercises, the BCU had the highest EMG activity for the URA and the LRA, 92.4 and 86.72%, respectively. The BCU was found to differ significantly from all other exercises (p = 0.000). The CU provoked the second-highest EMG activity of both the URA and the LRA, followed by the ATCU, the RCU, the LL, and the BRO (Figure 5 ).
Discussion
The purpose of this study was to determine and compare the behavior of the URA and the LRA while the subjects performed various abdominal exercises intended to target these specific muscle sites. No significant differences were found between the URA and the LRA during the concentric phases of any of the 6 exercises. However, there was a trend toward somewhat higher EMG amplitudes in
the URA as compared with the LRA during all exercises with the exception of the RCU, where the LRA had higher EMG amplitudes.
Though these trends were not found to be significant, findings by De Faria et al. (8), Sarti et al. (20), and Whiting et al. (24) support this observation for the CU and ATCU. In contrast, Gilleard and Brown (11) found the LRA to be more active than the URA during the LL exercise, whereas in this study, the URA has slightly higher amplitudes, though once again, the differences between the 2 muscle sites
were not significant.
Significant differences were found in the amplitudes of the action potentials of both the URA and LRA due to the exercises performed. The differences due to exercise reflect the differences inherent in the movements themselves. The CU and ATCU are essentially the same exercise, with the addition of the Ab Trainer device, and were not found to be significantly different, supporting the findings of
Whiting et al. (24). Warden et al. (23) found significantly higher amplitudes in the URA but no significant differences in the LRA during exercises performed with a similar device, the Abshaper, when compared with the conventional curl up. The BCU, with its added balance component, had much higher EMG activity than, and was significantly different from, all other exercises. Although the basic
movement is essentially still a curl up, it is very apparent that more motor units must be utilized in order to maintain balance and perform the exercise simultaneously.
The LL exercise involved the abdominal musculature as primary stabilizers of the pelvis while the subject performed perturbations of the lower limb (11). This is a different exercise from the others in that it is essentially an isometric contraction accompanied by restricted eccentric and concentric contractions throughout the abdomen as well as eccentric followed by concentric contractions of the iliacus and psoas muscles. The BRO is a very different exercise yet again. It involves strong stabilization of the pelvis, an isometric, as well as restricted eccentric contraction of the anterior abdominal musculature while rolling out onto the Sissel ball, an isometric contraction of
the abdominal musculature as well as a balancing act during the isometric phase of the exercise, and an isometric contraction coupled with restricted concentric contraction of the abdominal musculature to return to the starting position. During the BRO, there is also eccentric contraction of the latissimus dorsi, iliacus, and psoas muscles as well as their synergists, followed by an isometric and then
concentric contraction of these muscles. Although these muscles were not monitored, subjects reported ‘feeling’ these muscles throughout the exercise. The contribution of nonabdominal muscles to the movement of the BCU may have required less activation of the RA muscle for its successful completion. The BRO and LL, both primarily isometric abdominal exercises, proved to have the lowest
amplitudes of all the exercises studied. Gilleard and Brown (11) found the same of double leg lowering. Finally, the RCU involved the fixation of the insertion of the RA, the rib cage, and movement of the origin, a posterior tilt of the pelvis, toward that insertion. It is a dynamic movement, however, and had higher amplitudes than the isometric LL and BRO.
It was thought that there might have been a difference in response of the upper and lower portions of the RA to the exercises performed because of the segmental innervation of the anterolateral abdominal musculature. This metameric nerve supply from the ventral rami of the lower 6 or 7 thoracic nerves to the RA and its tendinous inscriptions (12) imply the ability to selectively isolate
different portions of the RA muscle (20).
The RA is a strap-like muscle with insertions to the costal cartilages of the 5th, 6th, and 7th ribs and the xiphoid process superiorly, and to the crest of the pubis, the pubic tubercle, and the front of the pubis symphysis inferiorly. As well, some of its fibers have been found to extend beyond the limits of the segments, crossing outside the tendinous inscriptions (12). From a mechanical viewpoint, the
latter suggests that the RA might not have the ability to contract with significantly more force at one end than the other. We found no significant difference between the URA and the LRA for any exercise, which tends to support this standpoint.
Although the subjects were physically fit and had experience with general abdominal exercises, they were not trained specifically for this study. Sarti et al. (20) divided subjects into 2 groups: those that could perform the exercises correctly, and those that could not. They found that the correct performers elicited significantly higher activity from the URA than from the LRA during the curl up, while the
incorrect performers did not. This brings in the question of specificity of training. If the subjects were trained in certain protocols, would we see training effects and subsequently differences between the URA and the LRA that may not be present in nontrained individuals?
Further study is needed on the concept of specificity and training effects in the anterolateral abdominal musculature.
Practical Applications
The RA is the most central of the anterolateral abdominal muscles. Due to its attachments and fiber orientation, the RA plays an important role in the movement and stabilization of the lumbar spine through its ability to produce and control movements of the spine relative to the pelvis and of the pelvis relative to the spine (13, 15, 19, 23). Poor pelvic position has often been cited as contributing to
low back pain and interfering in the performance of functional daily activities as well as limiting sport performance. An anterior tilt of the pelvis, in particular, contributes to lordosis and is one of the most common misalignments seen among dance-exercise participants. This pelvic tilt is generally the result of weak abdominal muscles and subsequently an incorrect standing posture (15).
All abdominal exercises tested elicited a response from both the URA and the LRA. The 3 curl up-type exercises (CU, BCU, and ATCU) had higher EMG activities on average than the LL, BRO, and RCU. Due to the balance component with the introduction of an unstable environment in which to perform the curl up, the BCU had higher activation than all other exercises. Although it is still a trunk-flexion exercise, it is apparent that more motor units must be recruited in order to maintain balance and perform the exercise simultaneously, making the Sissel ball a valuable tool for functional strengthening of the RA.
Strong abdominal musculature has been cited as contributing to improved low back health as well as improved performance in sport (13, 19) by making the transfer of forces through the trunk during perturbations of the upper and lower extremities more efficient.
Evidence suggests that specific muscle activation patterns are utilized to achieve spinal stabilization and target exercises must specifically target the stabilizing musculature. Proper conditioning of the abdominal musculature may prevent low back injuries and achieve the desired stabilization effect (13, 19). Although some of these exercises were recruiting fewer muscle fibers than others, it is also important to take into account specificity of the sport for which the athlete or fitness class participant is training. Abdominal muscle training should reflect the requirements of that particular sport. Each of the exercises tested was different in its movement pattern or its environment from the next.
Adequate muscle tone in the abdominal area is undoubtedly important in performing everyday tasks and in sport; however, one should be cautious when beginning abdominal training and start with exercises such as the CU and the ATCU, which still elicit a strong response but which also offer a stable environment in which to learn the exercise before moving on to the more difficult and higher activity BCU. There is, therefore, a place within an exercise or rehabilitation regime in which to utilize all of these exercises.
References
1. Axler, C.T., and S.M. McGill. Low back loads over a variety of abdominal exercises: Searching for the safest abdominal challenge.
Med. Sci. Sports Exerc. 29:804–810. 1997. [PubMed Citation]
2. Bankoff, A.D.P., and J. Furlani. Electromyographic study of the rectus abdominis and external oblique muscles during exercises.
Electromyogr. Clin. Neurophysiol. 24:501–510. 1984.
3. Basmajian, J.V., and C.J. De Luca. Muscles Alive: Their Functions Revealed by Electromyography (5th ed). Baltimore, MD:
Williams and Wilkins, 1985.
4. Beim, G.M., J.L. Giraldo, D.M. Pincivero, M.J. Borror, and F.H. Fu. Abdominal strengthening exercises: A comparative study. J. Sport
Rehabil. 6:11–20. 1997.
5. Brooks, D., and C. Copeland-Brooks. Uncovering the myths of abdominal exercise. Idea Today 42–49. April 1993.
6. Bullivant, A.M., C. Burchell, J. Chamberlain, E.K. Miller, and M.P. Wareham. An electromyographic and biomechanical investigation
of a variety of sit-up techniques. Br. J. Phys. Ed. 17:37–39. 1986.
7. Cram, J.R., G.S. Kazman, and J. Holtz. Introduction to Surface Electromyography. Gaithersburg, MD: Aspen Publishers, Inc., 1998.
8. De Faria Negrao Filho, R., F. Berzin, and G. Da Cunha Souza. Electromyography study of the portions of the abdominal rectus
muscle. Electromyogr. Clin. Neurophysiol. 37:491–501. 1997.
9. Flint, M.M., and J. Gudgell. Electromyographic study of abdominal muscular activity during exercise. Res. Q. 36:29–37. 1965.
10. Floyd, W.F., and P.H.S. Silver. Electromyographic study of patterns of activity of the anterior abdominal wall muscles in man. J.
Anat. 84:132–145. 1950.
11. Gilleard, W.L., and J.M.M. Brown. An electromyographic validation of an abdominal muscle test. Arch. Phys. Med. Rehabil.
75:1002–1007. 1994. [PubMed Citation]
12. Gray, H. Gray's Anatomy: The Anatomical Basis of Medicine and Surgery. London: Pearson Professional Ltd., 1995.
13. Johnson, P. Training the trunk in the athlete. Strength Cond. J. 24:52–59. 2002.
14. Juker, D., S. McGill, P. Kropf, and T. Steffen. Quantitative intramuscular myoelectric activity of lumbar portions of psoas and the
abdominal wall during a wide variety of tasks. Med. Sci. Sports Exerc. 30:301–310. 1998. [PubMed Citation]
15. Kravitz, L., and S. K*tner. Training the abdominals. Dance Exerc. Today 25–27. August 1988.
16. Lipetz, S., and B. Gutin. An electromyographic study of four abdominal exercises. Med. Sci. Sports. 2:35–38. 1970. [PubMed
Citation]
17. Mirka, G.A. The quantification of EMG normalization error. Ergonomics. 34:343–352. 1991. [PubMed Citation]
18. Ng, J.K.-F., V. Kippers, and C.A. Richardson. Muscle fibre orientation of abdominal muscles and suggested surface EMG electrode
positions. Electromyogr. Clin. Neurophysiol. 38:51–58. 1998.
19. Norris, C.M. Abdominal muscle training in sport. Br. J. Sports Med. 27:19–27. 1993. [PubMed Citation]
20. Sarti, M.A., M. Monfort, M.A. Fuster, and L.A. Villaplana. Muscle activity in upper and lower rectus abdominis during abdominal
exercises. Arch. Phys. Med. Rehabil. 77:1293–1297. 1996. [PubMed Citation]
21. Soderberg, G.L. (ed.). Selected Topics in Surface Electromyography for Use in the Occupational Setting: Expert Perspectives.
U.S. Dept. of Health and Human Services. DHHS (NIOSH). 1992.
22. Vezina, J.M.J. Electromyographic analysis of three trunk stability exercises in normal healthy subjects. Master's thesis. Dalhousie
University, Halifax, Nova Scotia, Canada, 1999.
23. Warden, S.J., H. Wajswelner, and K.L. Bennel. Comparison of abshaper and conventionally performed abdominal exercises using
surface electromyography. Med. Sci. Sports Exerc. 31:1656–1664. 1999. [PubMed Citation]
24. Whiting, W.C., S. Rugg, A. Coleman, and W.J. Vincent. Muscle activity during sit-ups using abdominal exercise devices. J. Strength
Cond. Res. 13:339–345. 1999.
Table 1. Summary of intraclass correlation coefficients for reference contractions and their trials.*
[Afbeelding niet meer beschikbaar]
Table 2. p Values for %AMVC of all subjects during the concentric phase.*
[Afbeelding niet meer beschikbaar]
ABSTRACT
The objective of this pilot study was to determine the effect of 6 different abdominal exercises on the electrical activity of the upper rectus abdominis (URA) and lower rectus abdominis (LRA). Eight healthy, adult volunteers completed 6 random abdominal exercises: curl up, Sissel ball curl up, Ab
Trainer curl up, leg lowering, Sissel ball roll out, and reverse curl up. Action potentials were recorded and analyzed from the URA and the LRA using surface electromyography (EMG) during a 2-second concentric contraction. The average normalized data were compared between the URA and the
LRA in order to determine the behavior of the different muscle sites and between exercises in order to determine which exercises elicited the highest EMG activity. There were no significant differences (p > 0.05) between the EMG activity of the URA and LRA during any exercise. There were no significant interactions between subject and muscle site or between exercise and muscle site. Significant differences were found among the 6 exercises performed, and due to the interaction between subject and exercise performed. Both the URA and the LRA recorded significantly higher mean amplitudes during the Sissel ball curl up than during all other exercises. In addition, the curl up, Sissel ball curl up, and Ab Trainer curl up had significantly higher normalized EMG activity in both muscle sites than the reverse curl up, the leg lowering exercise, and the Sissel ball roll out. The curl up and the Ab Trainer curl up exercises were not significantly different in terms of their normalized EMG activities for both the URA
and the LRA.
Reference Data:Clark, K.M., L.E. Holt, and J. Sinyard. Electromyographic comparison of the upper and lower rectus abdominis during abdominal exercises.
Key Words: motor unit, surface electromyography, normalization
Electromyographic Comparison of the Upper and Lower Rectus Abdominis During Abdominal Exercises
Kathryn M. Clark, Laurence E. Holt, and Joy Sinyard
The School of Health and Human Performance, Dalhousie University, Halifax, Nova Scotia, Canada B3H 3J5
Introduction
The abdominal musculature plays a key role in the movement and stability of the lumbar spine. When the muscles, which contribute to the stability of this inherently unstable area, are compromised and their stabilizing role is negatively influenced, there are significant implications with regard to the development of low back pain (23). For this reason, many health professionals advocate abdominal
musculature exercises for the prevention or rehabilitation from back pain (1, 11, 15, 19, 20, 24). Abdominal exercises are also popular in fitness training programs and are important for athletic training (13, 19). In a survey conducted at San Jose State University in California, aerobics participants ranked the abdominal musculature as the most important body part to train (15).
While most researchers and fitness professionals are aware that the rectus abdominis (RA) is one muscle, it is general practice to indicate to clientele that they are indeed “working the upper or lower abs” while performing certain exercises. As a result, clients are led to believe they are recruiting more motor units from one portion of the RA than the other or that the upper fibers of the RA (URA) and the
lower fibers of the RA (LRA) are actually separate muscles. Although the RA has been studied extensively using surface electromyography (EMG), most studies do not directly compare the URA to the LRA (2, 4, 6, 10, 23, 24).
Whether or not exercise type can differentially influence the upper and lower fibers of the RA has consistently been an area of debate among clinicians, researchers, and fitness professionals. Allusions have been made in the literature that the different segments of the RA offer different stabilizing potentials (22). Among the studies that do compare the 2 muscle sites (8, 11, 16, 20), the differences in the exercises used and the methods of collecting, analyzing, and reporting the data make it difficult to see any similarities between the 2 sites of interest or to draw any significant conclusions. In many of the studies in which the authors have seen differences in electrical activity between the URA and the LRA, there is no agreement as to where and how these variations occur (2, 3, 8–11, 16, 20). Many
studies have found the curl up to elicit higher EMG activity from the URA than the LRA (16, 20, 24), whereas the reverse curl up and double leg raise have been touted as primarily LRA exercises (5, 11, 20). Other studies have found no significant difference in EMG activity between the 2 muscle sites (2, 9).
It was thought there might be a difference in response to the exercises from the URA and the LRA due to the segmental innervation of the anterolateral abdominal musculature from the ventral rami of the lower 6 or 7 thoracic nerves. The segmentation of the RA by its tendinous inscriptions also implies the ability to selectively isolate different portions of this muscle (20).
In contrast, the RA is a strap-like muscle with its insertions on the coastal cartilages of the 5th, 6th, and 7th ribs and the xiphoid process superiorly and to the crest of the pubis, the pubic tubercle, and the front of the pubis symphysis inferiorly. Some fibers have also been found to extend beyond the limits of the segments, crossing outside the tendinous inscriptions (12). From a biomechanical
perspective, this evidence suggests that the RA might not have the ability to contract more forcibly at one end than the other.
Methods
Experimental Approach to the Problem
This study looked at 6 RA exercise protocols using surface EMG. Our purpose in this study was to determine the effect of 4 traditional and 2 new abdominal exercises on the URA and LRA as well as to determine which of the 6 exercises elicited the highest electrical response from the RA musculature. Four of the exercises, i.e., the curl up, the reverse curl up, the Ab Trainer Curl up and the leg
lowering exercises, replicate previous research (1, 4, 8, 11, 14, 19, 20, 23, 24). In addition, we have tested the Sissel ball curl up and Sissel ball roll out using a stability ball, a piece of equipment used frequently in both the exercise and rehabilitative industries, which has not previously been studied in this manner.
Surface EMG was chosen for its accessibility and noninvasive nature. As the abdominal muscles are organized in layers, crosstalk from nearby or underlying muscles, which may confound the validity of surface EMG signals, becomes an important issue. A thorough review of the anatomy and fiber orientation of the anterolateral abdominal musculature demonstrates areas where the abdominals may
be recorded that will minimize crosstalk. Careful placement of electrodes is also a key element in minimizing crosstalk effects (7, 18).
The bipolar electrode configuration chosen for this study is almost always used in order to eliminate movement, electrocardiogram, and 60-cycle artifacts (7, 21). In this configuration, two electrodes are placed over the site of interest 2 cm apart, and a third, common-mode reference, or ground, electrode is placed away from the testing site in an electrically neutral environment, usually on a bony prominence
and in this case, on the right anterior superior iliac spine of the pelvis. This configuration also allows the signal to be specific to the sites in question. As the 2 electrodes get farther apart, the signal has been shown to become more regional and less specific (7).
Subjects
Eight volunteer subjects, 4 men and 4 women, participated in this study. All subjects were apparently healthy adults with no history of low back pain or abdominal surgery, as determined from a verbal questionnaire given prior to testing. All subjects had extensive experience performing multiple abdominal exercise protocols and were familiar with the equipment. Subjects were participants in structured physical activity at Dalhousie University, were physically active an average of 5 times a week, and trained their abdominal musculature an average of 3 times a week. The mean age of the subjects was 31.75 &PLUSMN; 8.63 years. Their mean weight was 75.41 ± 12.77 kg
and their mean height was 174.34 ± 9.34 cm. Before participation, all subjects provided written informed consent. This study received ethical approval from a Dalhousie University committee prior to data collection.
Experimental Design
To measure the action potentials of the URA and LRA, 4 surface electrodes were placed on the abdominal musculature of the subjects.
The muscle and ground electrode sites were prepared according to Cram et al. (7) and electrodes were applied according to Ng et al. (18) (Figure 1 ).
Two sites on the RA were tested using Hewlett Packard pregelled, disposable Ag/AgCl electrodes. The URA site was 3 cm above and 2 cm to the right of the umbilicus, over the muscle belly, and parallel to the muscle fibers of the URA. The LRA site was 3 cm below and 2 cm to the right of the umbilicus on an 8° inferomedial angle, over the muscle belly, and parallel to the muscle fibers of the lower portion of the RA (Figure 1 ) (18).
The raw EMG signal was amplified 1,000 times using two 2-channel bioamps. The signal was then passed through an EMG processor, where it was again amplified 5 times and filtered to eliminate artifacts that would otherwise alter the EMG signal; the band pass was set at 20–1,000 Hz. This filtered signal was full-wave rectified and linear enveloped. The EMG signal was sampled using Lab
View, Gen99 software (National Instruments, Austin, TX). A 12-bit National Instruments Card analogue to digital (A/D) converter was used to convert the signal to a digital format. The signal was then converted to volts and the root mean square (RMS) calculated.
Reference Contractions
The utility of surface EMG has often been questioned in the field of biomechanics (8, 17). Because of differences in subcutaneous fat, muscle geometry, skin impedance, trunk flexibility, age, coïtus, and other variables, absolute EMG amplitudes are not comparable between individuals, activities, or muscles. During dynamic movements in particular, comparison of amplitude measurements across muscle groups can be quite misleading. In order to circumvent this problem and to be able to make statements about the relative intensity of an EMG signal, it is necessary to normalize the EMG data (7, 17, 21).
The most common method of normalization is to have subjects perform 1 maximal voluntary contraction (MVC) as a reference. The EMG values subsequently obtained are then presented as a percentage of this MVC (21). Mirka (17) has shown that errors occur as a result of normalization to reference values collected from only 1 position for various muscles of the trunk. The study suggests that relating EMG recorded during a dynamic task to a reference value taken at one joint angle will render inaccurate results due to changes in the portion of the muscle within the viewing area of the electrode, changes in length/tension relationship of the muscle, and changes in the moment arm of the muscle around the joint of rotation. Mirka (17) proposes that, in order to be able to quantitatively compare the relative magnitude of the EMG signals of a dynamic movement under different joint angles, the maximum EMG signal used to normalize the data should be taken at several joint angles throughout the range of motion (ROM).
In order to ensure accuracy, we had each subject perform 4 different reference contractions in randomized order and normalization of their respective EMG activity was based on the average of these 4 reference contractions.
Each subject was initially asked to perform 1 curl up (CU) (Figure 3a ) in order to determine his or her normal range of motion (ROM) during cervical and thoracic spinal flexion.
Three isometric maximal voluntary contractions (MVC) were measured while the subject pushed against the resistance of a nylon strap, placed across the torso and arms just below the clavicle, which anchored them to the plinth. The subjects' arms and hands were placed across the chest, on top of the nylon strap in order to facilitate a maximal contraction and to prevent the strap from slipping. The
MVCs were measured at (a) 0° to the horizontal in the position of initiating a CU, (b) the halfway point of the subjects' measured ROM for the CU (15–20° to the horizontal), and (c) the subjects' measured full ROM for the CU (30–40° to the horizontal). The subjects were asked to push as hard as they could against the strap for 4 seconds (Figure 2a ).
A fourth, dynamic MVC was measured using the concentric phase of the CU. The subjects, no longer anchored to the plinth, were asked to push against a force manually applied to his or her shoulders while they were allowed to move through their full ROM over 4 seconds; arms and hands were again placed across the chest (Figure 2b ).
Subjects performed 3 trials each of the 4 MVCs and were required to take a 2-minute rest between trials. Angles were measured using a goniometer attached to the plinth at the level of the superior aspect of the iliac crest.
The RMS measures of each of the 3 MVC trials were visually inspected in order to find the highest average EMG activity over a 0.5-second time interval. Intraclass correlation coefficients (R) were calculated within each reference contraction across the 3 trials to determine the reliability of the MVC values. As a result of high R and p values indicating they were not significantly different, the 4 MVCs
were then averaged and applied as a single reference contraction (AMVC) for normalization (Table 1 ).
Exercises
The subjects were asked to perform 3 trials each of 6 randomized abdominal exercises. Each trial was performed continuously over 6 seconds with a 2-second concentric phase, a 2-second isometric phase, and a 2-second eccentric phase. All of the exercises and the concentric/isometric/eccentric contraction pattern are very common to general fitness classes and muscular conditioning programs
practiced in Halifax, Nova Scotia, university athletic centers. The trials were timed using a stopwatch, by a certified fitness instructor who counted out loud to the subjects. This protocol was selected because of its similarity to the timing used during many group fitness classes. A VHS video recorder was used to videotape all test sessions in order to confirm the timing of the 3 phases of the exercises.
No warm up was permitted and no more than 2 practice trials were allowed before each set of 3 test trials. A 1-minute rest was required between every trial.
The curl up (CU) (Figure 3a ) was performed lying supine on a plinth with both feet flat on the plinth and an included knee angle of 90°. The fingertips were placed at the base of the skull to support the head, with the elbows wide and lateral to the head. The subjects were instructed to lift the head and shoulders toward the ceiling over 2 seconds, to hold the new position for 2 seconds, and to slowly
return to the starting position over 2 seconds.
The Sissel ball curl up (BCU) (Figure 3b ) was performed lying supine on an inflatable, rubber Sissel ball, 55 cm in diameter. The fingertips were placed at the base of the skull, the elbows wide and lateral to the head. The subjects were instructed to keep both feet flat on the floor and to position themselves so their lower back was supported by the ball, their torso parallel to the floor. Correct positioning was accomplished by subjective feedback from a fitness professional. Once the starting position was achieved, they were then instructed to lift the head and shoulders toward the ceiling over 2 seconds, to hold the new position for 2 seconds, and to slowly return to the starting position over 2 seconds.
The Ab Trainer curl up (ATCU) (Figure 3c ) was performed using the Ab Trainer abdominal training device. This device consists of tubular metal shaped to roll forward and backward as the person performs a CU. The device extends above the exerciser and includes armrests and a headrest. The exercise was performed lying supine on a plinth with both feet flat on the plinth and an included knee
angle of 90°. The elbows rested anterolaterally to the head on the armrests and in this case, the headrest supported the head. The subjects were instructed to lift the head and shoulders toward the ceiling over 2 seconds, to hold the new position for 2 seconds, and to slowly return to the starting position over 2 seconds. They were also asked to maintain head contact with the headrest of the device.
The leg lowering exercise (LL) (Figure 3d ) was performed lying supine on a plinth with the legs in the air. In the starting position, the hips were flexed at a 90° included angle and the knees were flexed at a 45° included angle. The fingertips were placed at the base of the skull, the elbows wide and lateral to the head. The upper torso, arms, and head remained in contact with the plinth for the duration
of the exercises. The subjects were instructed to slowly lower their legs over 2 seconds so their heels were lightly touching, but not resting on, the plinth, while maintaining both the 45° angle at the knee joint and lower back contact with the plinth. They were then instructed to hold the lowered position for 2 seconds and to slowly return to the starting position over 2 seconds. The sequence of contraction for the LL was reversed in that the eccentric phase was performed first followed by the isometric and concentric phases.
The Sissel ball roll out (BRO) (Figure 3e ) was performed kneeling on the floor with a small mat under the knees for comfort. The forearms were placed on a 55-cm Sissel ball and the subject was instructed to roll out over 2 seconds into a fully extended position (i.e., the hips were fully extended and the shoulders were in a flexed position next to the head). They were then asked to hold the extended
position for 2 seconds and to slowly return to the starting position over 2 seconds. The sequence of contraction for the BRO was reversed in that the eccentric phase was performed first followed by the isometric and concentric phases, as in the LL exercise.
The reverse curl up (RCU) (Figure 3f ) was performed lying supine on a plinth with the legs in the air. The hips were flexed at a 90° included angle and the knees were flexed at a 45° included angle. The fingertips were placed at the base of the skull, the elbows wide and lateral to the head. The upper torso, arms, and head remained in contact with the plinth for the duration of the exercise. The subjects
were instructed to curl the hips toward the ribs (i.e., posteriorly tilt the pelvis) over 2 seconds while maintaining the 45° angle at the knees, to hold the new position for 2 seconds, and to slowly lower the hips back to the starting position over 2 seconds.
The protocol chosen for each exercise was in accordance with trends in the general fitness field. It is common to rest the head in the fingertips during many exercises in order to provide a longer lever and thereby higher intensity for the abdominal musculature during the
curl up. Using only the fingertips allows the participant to rest the head and neck, minimizing fatigue of the anterior neck musculature while performing repetitive abdominal exercises and cues participants not to strain the head and neck into a more forward position.
Using the whole hand or lacing the fingers behind the head encourages participants to use the hands and arms to forcibly pull on the head and neck. Placing the arms and hands across the chest is less intense for the abdominals, as the weight of the arms is now closer to the subjects' center of gravity. This position may also place more strain on the muscles in the anterior neck, frequently causing neck
discomfort and an early termination of abdominal exercises before fatigue of the anterior abdominal musculature is achieved.
The knee and hip joints are flexed to shorten the hip flexors and minimize their contribution to the movement in order to better isolate the anterior abdominal musculature and to reduce stress on the lumbar spine by the pulling of the psoas insertions on the lumbar vertebrae.
Exercise measurements were sectioned into 3 phases by a synch pulse that was recorded along with the EMG in order that the concentric data be easily separated from the isometric and eccentric for analysis. The pulse was operated by an assistant who turned the pulse on when the subject began to visibly move, indicating the beginning of the first phase, the concentric phase for the CU, BCU,
ATCU and RCU, and the eccentric phase for the LL and the BRO. The pulse was turned off when the subject ceased to move, indicating the beginning of the second, isometric phase, and turned on once again as the subject began to move, indicating the beginning of the third phase, the eccentric phase for the CU, BCU, ATCU, RCU, and the concentric phase for the LL and the BRO. The pulse was turned
off at the end of the third and final phase.
The RMS measures of each of the 3 test trials were sectioned in order to find the average EMG activity over the 2-second concentric phase. R was calculated across the 3 trials for each exercise's concentric phase to determine trial reliability. As a result of high R values, the 3 trials were averaged and this value used for normalizing as a percentage of each subjects' MVC for each muscle site.
These values were then averaged across the subjects for comparison (%AMVC).
Statistical Analyses
General linear model analyses of variance (ANOVA) were performed to assess the differences in %AMVC of the URA and the LRA due to the exercise conditions and phase of movement. Post hoc analysis was performed using Tukey's method for between-exercise and between-phase comparisons. The level for statistical significance was set at the 95% confidence limit (? = 0.05). All statistical analyses were performed using the Minitab statistical analysis package.
Results
High R values verified that the 4 reference contractions were reliable among the trials as well as among 3 of the 4 MVCs (Table 2 ).
During the isometric MVC at full ROM, R values dropped off considerably; however, because the MVC was not found to be significantly different from the others (p = 0.12), it was therefore included in the AMVC. A high R value (R = 0.96) also verified that the %AMVC for all exercises performed were reliable among the trials and they were therefore averaged.
The data from the concentric phase was chosen for analysis because it produced the highest EMG activity overall and is a dynamic contraction over the subjects' full ROM.
Tukey simultaneous tests performed among the phases found the average %AMVC values for all subjects during the concentric phase to be significantly higher than the eccentric phase (p < 0.05) and not to differ significantly from the isometric phase (p = 0.24).
The lower %AMVC recorded from the subjects during the eccentric phase of the CU, BCU, ATCU, and RCU (Figure 4 ) is the expected result of a decrease in the firing of motor units as the subject moves with gravity, not against it, as during the concentric and isometric phases (3). The %AMVC of the eccentric phase of the LL and the BRO (Figure 4 ) are closer in amplitude to their respective concentric phases. This is expected due to the increasing lever length as the subject lowers his or her legs to touch the plinth during the LL and rolls out into an extended position on the Sissel ball during the BRO. More force must be applied in these situations in order to control the longer lever, thereby increasing the firing of motor units (7). These results serve as another indicator that the signals
from the surface electrodes were indeed reliable.
In comparing the %AMVC of the concentric phases of the 6 exercises, the results of the ANOVA revealed there were no statistically significant effects due to muscle site (p = 0.081), interactions between subject and muscle site (p = 0.27), or exercise and muscle site (p = 0.10). Significant differences in the %AMVC were found in both the URA and the LRA due to the exercise performed (p < 0.00) and
the interaction between subject and exercise performed (p = 0.003). A 1-way ANOVA performed on the %AMVC values of the URA and the LRA during the concentric phase also revealed there was no significant difference between the 2 muscle sites (p = 0.116).
In comparing the %AMVC of the concentric phase of the 6 exercises, the BCU had the highest EMG activity for the URA and the LRA, 92.4 and 86.72%, respectively. The BCU was found to differ significantly from all other exercises (p = 0.000). The CU provoked the second-highest EMG activity of both the URA and the LRA, followed by the ATCU, the RCU, the LL, and the BRO (Figure 5 ).
Discussion
The purpose of this study was to determine and compare the behavior of the URA and the LRA while the subjects performed various abdominal exercises intended to target these specific muscle sites. No significant differences were found between the URA and the LRA during the concentric phases of any of the 6 exercises. However, there was a trend toward somewhat higher EMG amplitudes in
the URA as compared with the LRA during all exercises with the exception of the RCU, where the LRA had higher EMG amplitudes.
Though these trends were not found to be significant, findings by De Faria et al. (8), Sarti et al. (20), and Whiting et al. (24) support this observation for the CU and ATCU. In contrast, Gilleard and Brown (11) found the LRA to be more active than the URA during the LL exercise, whereas in this study, the URA has slightly higher amplitudes, though once again, the differences between the 2 muscle sites
were not significant.
Significant differences were found in the amplitudes of the action potentials of both the URA and LRA due to the exercises performed. The differences due to exercise reflect the differences inherent in the movements themselves. The CU and ATCU are essentially the same exercise, with the addition of the Ab Trainer device, and were not found to be significantly different, supporting the findings of
Whiting et al. (24). Warden et al. (23) found significantly higher amplitudes in the URA but no significant differences in the LRA during exercises performed with a similar device, the Abshaper, when compared with the conventional curl up. The BCU, with its added balance component, had much higher EMG activity than, and was significantly different from, all other exercises. Although the basic
movement is essentially still a curl up, it is very apparent that more motor units must be utilized in order to maintain balance and perform the exercise simultaneously.
The LL exercise involved the abdominal musculature as primary stabilizers of the pelvis while the subject performed perturbations of the lower limb (11). This is a different exercise from the others in that it is essentially an isometric contraction accompanied by restricted eccentric and concentric contractions throughout the abdomen as well as eccentric followed by concentric contractions of the iliacus and psoas muscles. The BRO is a very different exercise yet again. It involves strong stabilization of the pelvis, an isometric, as well as restricted eccentric contraction of the anterior abdominal musculature while rolling out onto the Sissel ball, an isometric contraction of
the abdominal musculature as well as a balancing act during the isometric phase of the exercise, and an isometric contraction coupled with restricted concentric contraction of the abdominal musculature to return to the starting position. During the BRO, there is also eccentric contraction of the latissimus dorsi, iliacus, and psoas muscles as well as their synergists, followed by an isometric and then
concentric contraction of these muscles. Although these muscles were not monitored, subjects reported ‘feeling’ these muscles throughout the exercise. The contribution of nonabdominal muscles to the movement of the BCU may have required less activation of the RA muscle for its successful completion. The BRO and LL, both primarily isometric abdominal exercises, proved to have the lowest
amplitudes of all the exercises studied. Gilleard and Brown (11) found the same of double leg lowering. Finally, the RCU involved the fixation of the insertion of the RA, the rib cage, and movement of the origin, a posterior tilt of the pelvis, toward that insertion. It is a dynamic movement, however, and had higher amplitudes than the isometric LL and BRO.
It was thought that there might have been a difference in response of the upper and lower portions of the RA to the exercises performed because of the segmental innervation of the anterolateral abdominal musculature. This metameric nerve supply from the ventral rami of the lower 6 or 7 thoracic nerves to the RA and its tendinous inscriptions (12) imply the ability to selectively isolate
different portions of the RA muscle (20).
The RA is a strap-like muscle with insertions to the costal cartilages of the 5th, 6th, and 7th ribs and the xiphoid process superiorly, and to the crest of the pubis, the pubic tubercle, and the front of the pubis symphysis inferiorly. As well, some of its fibers have been found to extend beyond the limits of the segments, crossing outside the tendinous inscriptions (12). From a mechanical viewpoint, the
latter suggests that the RA might not have the ability to contract with significantly more force at one end than the other. We found no significant difference between the URA and the LRA for any exercise, which tends to support this standpoint.
Although the subjects were physically fit and had experience with general abdominal exercises, they were not trained specifically for this study. Sarti et al. (20) divided subjects into 2 groups: those that could perform the exercises correctly, and those that could not. They found that the correct performers elicited significantly higher activity from the URA than from the LRA during the curl up, while the
incorrect performers did not. This brings in the question of specificity of training. If the subjects were trained in certain protocols, would we see training effects and subsequently differences between the URA and the LRA that may not be present in nontrained individuals?
Further study is needed on the concept of specificity and training effects in the anterolateral abdominal musculature.
Practical Applications
The RA is the most central of the anterolateral abdominal muscles. Due to its attachments and fiber orientation, the RA plays an important role in the movement and stabilization of the lumbar spine through its ability to produce and control movements of the spine relative to the pelvis and of the pelvis relative to the spine (13, 15, 19, 23). Poor pelvic position has often been cited as contributing to
low back pain and interfering in the performance of functional daily activities as well as limiting sport performance. An anterior tilt of the pelvis, in particular, contributes to lordosis and is one of the most common misalignments seen among dance-exercise participants. This pelvic tilt is generally the result of weak abdominal muscles and subsequently an incorrect standing posture (15).
All abdominal exercises tested elicited a response from both the URA and the LRA. The 3 curl up-type exercises (CU, BCU, and ATCU) had higher EMG activities on average than the LL, BRO, and RCU. Due to the balance component with the introduction of an unstable environment in which to perform the curl up, the BCU had higher activation than all other exercises. Although it is still a trunk-flexion exercise, it is apparent that more motor units must be recruited in order to maintain balance and perform the exercise simultaneously, making the Sissel ball a valuable tool for functional strengthening of the RA.
Strong abdominal musculature has been cited as contributing to improved low back health as well as improved performance in sport (13, 19) by making the transfer of forces through the trunk during perturbations of the upper and lower extremities more efficient.
Evidence suggests that specific muscle activation patterns are utilized to achieve spinal stabilization and target exercises must specifically target the stabilizing musculature. Proper conditioning of the abdominal musculature may prevent low back injuries and achieve the desired stabilization effect (13, 19). Although some of these exercises were recruiting fewer muscle fibers than others, it is also important to take into account specificity of the sport for which the athlete or fitness class participant is training. Abdominal muscle training should reflect the requirements of that particular sport. Each of the exercises tested was different in its movement pattern or its environment from the next.
Adequate muscle tone in the abdominal area is undoubtedly important in performing everyday tasks and in sport; however, one should be cautious when beginning abdominal training and start with exercises such as the CU and the ATCU, which still elicit a strong response but which also offer a stable environment in which to learn the exercise before moving on to the more difficult and higher activity BCU. There is, therefore, a place within an exercise or rehabilitation regime in which to utilize all of these exercises.
References
1. Axler, C.T., and S.M. McGill. Low back loads over a variety of abdominal exercises: Searching for the safest abdominal challenge.
Med. Sci. Sports Exerc. 29:804–810. 1997. [PubMed Citation]
2. Bankoff, A.D.P., and J. Furlani. Electromyographic study of the rectus abdominis and external oblique muscles during exercises.
Electromyogr. Clin. Neurophysiol. 24:501–510. 1984.
3. Basmajian, J.V., and C.J. De Luca. Muscles Alive: Their Functions Revealed by Electromyography (5th ed). Baltimore, MD:
Williams and Wilkins, 1985.
4. Beim, G.M., J.L. Giraldo, D.M. Pincivero, M.J. Borror, and F.H. Fu. Abdominal strengthening exercises: A comparative study. J. Sport
Rehabil. 6:11–20. 1997.
5. Brooks, D., and C. Copeland-Brooks. Uncovering the myths of abdominal exercise. Idea Today 42–49. April 1993.
6. Bullivant, A.M., C. Burchell, J. Chamberlain, E.K. Miller, and M.P. Wareham. An electromyographic and biomechanical investigation
of a variety of sit-up techniques. Br. J. Phys. Ed. 17:37–39. 1986.
7. Cram, J.R., G.S. Kazman, and J. Holtz. Introduction to Surface Electromyography. Gaithersburg, MD: Aspen Publishers, Inc., 1998.
8. De Faria Negrao Filho, R., F. Berzin, and G. Da Cunha Souza. Electromyography study of the portions of the abdominal rectus
muscle. Electromyogr. Clin. Neurophysiol. 37:491–501. 1997.
9. Flint, M.M., and J. Gudgell. Electromyographic study of abdominal muscular activity during exercise. Res. Q. 36:29–37. 1965.
10. Floyd, W.F., and P.H.S. Silver. Electromyographic study of patterns of activity of the anterior abdominal wall muscles in man. J.
Anat. 84:132–145. 1950.
11. Gilleard, W.L., and J.M.M. Brown. An electromyographic validation of an abdominal muscle test. Arch. Phys. Med. Rehabil.
75:1002–1007. 1994. [PubMed Citation]
12. Gray, H. Gray's Anatomy: The Anatomical Basis of Medicine and Surgery. London: Pearson Professional Ltd., 1995.
13. Johnson, P. Training the trunk in the athlete. Strength Cond. J. 24:52–59. 2002.
14. Juker, D., S. McGill, P. Kropf, and T. Steffen. Quantitative intramuscular myoelectric activity of lumbar portions of psoas and the
abdominal wall during a wide variety of tasks. Med. Sci. Sports Exerc. 30:301–310. 1998. [PubMed Citation]
15. Kravitz, L., and S. K*tner. Training the abdominals. Dance Exerc. Today 25–27. August 1988.
16. Lipetz, S., and B. Gutin. An electromyographic study of four abdominal exercises. Med. Sci. Sports. 2:35–38. 1970. [PubMed
Citation]
17. Mirka, G.A. The quantification of EMG normalization error. Ergonomics. 34:343–352. 1991. [PubMed Citation]
18. Ng, J.K.-F., V. Kippers, and C.A. Richardson. Muscle fibre orientation of abdominal muscles and suggested surface EMG electrode
positions. Electromyogr. Clin. Neurophysiol. 38:51–58. 1998.
19. Norris, C.M. Abdominal muscle training in sport. Br. J. Sports Med. 27:19–27. 1993. [PubMed Citation]
20. Sarti, M.A., M. Monfort, M.A. Fuster, and L.A. Villaplana. Muscle activity in upper and lower rectus abdominis during abdominal
exercises. Arch. Phys. Med. Rehabil. 77:1293–1297. 1996. [PubMed Citation]
21. Soderberg, G.L. (ed.). Selected Topics in Surface Electromyography for Use in the Occupational Setting: Expert Perspectives.
U.S. Dept. of Health and Human Services. DHHS (NIOSH). 1992.
22. Vezina, J.M.J. Electromyographic analysis of three trunk stability exercises in normal healthy subjects. Master's thesis. Dalhousie
University, Halifax, Nova Scotia, Canada, 1999.
23. Warden, S.J., H. Wajswelner, and K.L. Bennel. Comparison of abshaper and conventionally performed abdominal exercises using
surface electromyography. Med. Sci. Sports Exerc. 31:1656–1664. 1999. [PubMed Citation]
24. Whiting, W.C., S. Rugg, A. Coleman, and W.J. Vincent. Muscle activity during sit-ups using abdominal exercise devices. J. Strength
Cond. Res. 13:339–345. 1999.
Table 1. Summary of intraclass correlation coefficients for reference contractions and their trials.*
[Afbeelding niet meer beschikbaar]
Table 2. p Values for %AMVC of all subjects during the concentric phase.*
[Afbeelding niet meer beschikbaar]
ABSTRACT
The objective of this pilot study was to determine the effect of 6 different abdominal exercises on the electrical activity of the upper rectus abdominis (URA) and lower rectus abdominis (LRA). Eight healthy, adult volunteers completed 6 random abdominal exercises: curl up, Sissel ball curl up, Ab
Trainer curl up, leg lowering, Sissel ball roll out, and reverse curl up. Action potentials were recorded and analyzed from the URA and the LRA using surface electromyography (EMG) during a 2-second concentric contraction. The average normalized data were compared between the URA and the
LRA in order to determine the behavior of the different muscle sites and between exercises in order to determine which exercises elicited the highest EMG activity. There were no significant differences (p > 0.05) between the EMG activity of the URA and LRA during any exercise. There were no significant interactions between subject and muscle site or between exercise and muscle site. Significant differences were found among the 6 exercises performed, and due to the interaction between subject and exercise performed. Both the URA and the LRA recorded significantly higher mean amplitudes during the Sissel ball curl up than during all other exercises. In addition, the curl up, Sissel ball curl up, and Ab Trainer curl up had significantly higher normalized EMG activity in both muscle sites than the reverse curl up, the leg lowering exercise, and the Sissel ball roll out. The curl up and the Ab Trainer curl up exercises were not significantly different in terms of their normalized EMG activities for both the URA
and the LRA.
Reference Data:Clark, K.M., L.E. Holt, and J. Sinyard. Electromyographic comparison of the upper and lower rectus abdominis during abdominal exercises.
Key Words: motor unit, surface electromyography, normalization
