core stability exercises on and off a swiss ball

8/6/2019 Core Stability Exercises on and Off a Swiss Ball

http://slidepdf.com/reader/full/core-stability-exercises-on-and-off-a-swiss-ball 1/8

Core Stability Exercises On and Off a Swiss Ball

Paul W. Marshall, PG Dip Sci, Bernadette A. Murphy, PhD

ABSTRACT. Marshall PW, Murphy BA. Core stability ex-ercises on and off a Swiss ball. Arch Phys Med Rehabil 2005;

86:242-9.Objectives: To assess lumbopelvic muscle activity during

different core stability exercises on and off a Swiss ball.Design: Prospective comparison study.Setting: Research laboratory.Participants: Eight healthy volunteers from a university

population.Intervention: Subjects performed 4 exercises on and off a

Swiss ball: inclined press-up, upper body roll-out, single-leghold, and quadruped exercise.

Main Outcome Measures: Surface electromyography fromselected lumbopelvic muscles, normalized to maximum volun-tary isometric contraction, and median frequency analysis of electromyography power spectrum. Visual analog scale for

perception of task difficulty.Results: There was a significant increase in the activation of the rectus abdominus with performance of the single-leg holdand at the top of the press-up on the Swiss ball. This led tochanges in the relation between the activation levels of thelumbopelvic muscles measured.

Conclusions: Although there was evidence to suggest thatthe Swiss ball provides a training stimulus for the rectusabdominus, the relevance of this change to core stability train-ing requires further research because the focus of stabilizationtraining is on minimizing rectus abdominus activity. Furthersupport has also been provided about the quality of the quad-ruped exercise for core stability.

Key Words: Abdominal muscles; Electromyography; Exer-cise; Rehabilitation.

© 2005 by American Congress of Rehabilitation Medicineand the American Academy of Physical Medicine and Rehabilitation

THE SWISS BALL (or gym ball) is widely reported in therecreational training environment to be a training device

for core stability exercises.1 However, there is little scientificevidence to support its use.2,3 It is also not clear whetherperforming an exercise on a Swiss ball has greater benefit thanperforming the same exercise on a stable surface.

The term core stability is a generic description for thetraining of the abdominal and lumbopelvic region. To definecore stability, the combination of a global and local stabilitysystem has been used. The global stability system refers to the

larger, superficial muscles around the abdominal and lumbarregion, such as the rectus abdominus, paraspinals, and external

obliques.4,5 These muscles are the prime movers for trunk orhip flexion, extension, and rotation. Local stability refers to the

deep, intrinsic muscles of the abdominal wall, such as thetransverse abdominus and multifidus. These muscles are asso-ciated with the segmental stability of the lumbar spine duringgross whole body movements and where postural adjustmentsare required.4,6-8

The validity of both the concept of core stability andthe optimal training protocols for core stability requires inves-tigation. For example, an exercise such as abdominal hollowing(eg, the drawing-in technique) attempts to emphasize local overglobal stability.9,10 For long-term core stability exercise pro-grams, this type of exercise neglects the synergistic relationbetween the muscles of the global and local stability systems.For any movement task that involves the trunk region, it wouldbe wrong to believe that only 1 specific muscle systemis actively involved. It is known that 1 muscle cannot be

identified as being more important for lumbar stability thananother.11 A more appropriate approach to core stability train-ing is to find exercises that incorporate the synergistic relationbetween the global and local stability systems, but still elicit asatisfactory training effect.

Our purpose in this study was to compare the activationpatterns of muscles associated with the global and local stabil-ity systems during different core stability tasks on and off aSwiss ball. The exercises did not involve prime movementtasks for the trunk region but permitted us to investigate thesynergistic relation between muscles when the overall stabilityof the lumbopelvic region is challenged by the weight force of the body segments. The hypotheses of this study were (1) theexercises performed on the Swiss ball would have greaterlevels of muscle activation compared with the stable surface,

and (2) the synergistic relationship between the ventrolateralabdominals and erector spinae expressed relative to the activityof the rectus abdominus would not be influenced by the exer-cise surface.

METHODS

Participants

Eight healthy subjects (4 men, 4 women) from our universityvolunteered for this study. The mean anthropometric charac-teristics standard deviation (SD) of the men were age,23.52.65y; height, 1.85.04m; and weight, 81.53.42kg; forthe women, they were age, 23.52.65y; height, 1.64.07m;and weight, 61.52.89kg. No subject was experiencing pain in

his/her body when tested, and no subject had experienced asignificant episode of low back pain (LBP) within the last 5years. Informed written consent was received from the subjectsbefore their participation. This study was approved by theAuckland Human Subjects Research Ethics Committee.

Data Recording

All testing was performed in the somatosensory physiologylaboratory at the University of Auckland. Skin impedance tothe electric signal was reduced to below 5k by (1) shavingexcess body hair if necessary, (2) gently abrading the skin withfine grade sandpaper, and (3) wiping the skin with isopropylalcohol swabs. If the measured impedance was greater than

From the Department of Sport and Exercise Science, University of Auckland,Auckland, New Zealand.

No commercial party having a direct financial interest in the results of the researchsupporting this article has or will confer a benefit upon the authors(s) or upon anyorganization with which the author(s) is/are associated.

Reprint requests to Paul Marshall, Dept of Sport and Exercise Science, Universityof Auckland, Tamaki Campus, Private Bag 92019, Auckland, New Zealand, e-mail:

[email protected].

0003-9993/05/8602-8830$30.00/0doi:10.1016/j.apmr.2004.05.004

242

Arch Phys Med Rehabil Vol 86, February 2005



5k , the surface electrodes were removed and the skin prep-aration was repeated.

Pairs of electrodes (3M Red Dot, Ag/AgCl electrodesa) witha contact diameter of 2cm and a center-to-center distance of 3cm were applied to the following locations on the right side of the body only: the rectus abdominus, 3cm lateral and superiorto the umbilicus, arranged along the longitudinal axis; theexternal obliques, the first electrode was placed at the intersec-tion of a line lateral to the umbilicus and superior to the anteriorsuperior iliac spine (ASIS), with the second electrode arrangedso that the bipolar configuration was approximately 45° to thehorizontal; the transversus abdominus and internal obliques,approximately 2cm inferior and medial to the ASIS (the musclefibers of the transversus abdominus and internal obliques areblended at this site,12 so a distinction between the musclesignals cannot be made in this location); and the erector spinae,located at the level of L4-5, approximately 3cm lateral to thespinous process and arranged along the longitudinal axis. Thereference electrode was placed over the superior aspect of theleft iliac crest.

Exercise Procedures

Upper-body roll out. In the prone roll out position, the

subject lay with the lower leg and feet only in contact with thesurface of the ball (fig 1). The hands were positioned directlyunderneath the shoulders, with the fingers facing forward. Thesurface test height (55cm or 65cm) was chosen so that the angleof the shoulder joint and the trunk was approximately 90° (asmanually measured with a flexible goniometer). The samesurface height was used for both test conditions.

Inclined press-up. The top and bottom positions of aninclined press-up on a 65-cm high surface were recorded. Thetop position was the initial starting point, with the hands placedon the surface directly beneath the shoulder joint, with armsfully extended, and the trunk positioned as far back as possibleso that upper-body position could be maintained (fig 1). Theposition of each subject’s feet was marked and held consistentduring all press-up trials. The bottom of the press-up wasrecorded after the subject had flexed the elbow joint to approx-imately 90°, lowering the trunk toward the ball but withoutmaking contact. The bottom of the press-up was moved intoimmediately after the collection period from the top of thepress-up.

Contralateral single-leg hold. The subject lay on a 65-cmhigh surface with the sacroiliac joint being the most distal partof the trunk supported. The right foot was positioned flat on thefloor throughout this task. The left leg was manually assisted toapproximately 90° of hip and knee flexion. From this position,the subject was instructed to extend the knee, then extend thehip until the thigh was parallel to the prone trunk position . Thisposition was the isometric test position for this exercise (fig 1).

Quadruped exercise. This isometric task was performed ina 2-point stance with a contralateral arm and leg raise (fig 1).

The subject was initially positioned in a 4-point stance withknees and hands on the floor (hips flexed to 90° and handsbeneath shoulder joint). On a verbal command, the subjectflexed the arm and extended the contralateral hip until bothupper- and lower-body segments were parallel to the trunk.This position was then held for the 4-second contraction. Thecommand for the alternate limbs to move was given after a

Fig 1. Digital photographs ofthe exercises performed dur-ing this experiment: (A) roll-out: performed on Swiss ball,inclined press-up in (B) topposition and (C) bottom posi-tion, (D) single-leg hold, and(E) quadruped exercise withright arm and left leg move-ment.

243CORE STABILITY EXERCISES ON AND OFF A SWISS BALL, Marshall




1-minute rest between trials. Three trials were performed for

each movement combination. For the unstable condition, aSwiss ball was placed beneath the subject’s abdomen so thatthere was contact between the torso and a labile surface. Eithera 55- or 65-cm Swiss ball was used, depending on the initialheight of the subject in the 4-point stance, to ensure that thetrunk position was consistent in comparison with the stablecondition.

All test positions were held isometrically for 4 seconds, withthe final 3 seconds providing the data to be analyzed. The taskswere always administered in a randomized order. For all tasks,3 repetitions were performed with a 1-minute rest between eachtrial. All subjects were familiarized with the tasks before datawere recorded.

Data Analysis

All data signals were recorded via a MACLABb interfaceunit connected to a Pentium II computer at a sampling fre-quency of 2000Hz with 16-bit analog-to-digital conversion, acommon mode rejection ratio of greater than 96dB at 50Hz,and an input impedance of 100M. The data was digitallyfiltered (20500Hz), and the root mean square (RMS) wascalculated for the 3 seconds collected for each muscle signal.13

The mean RMS activity over the 3 seconds was expressed asa percentage of a maximum voluntary contraction (MVC)performed for each muscle signal before the experiment. Themaximum trunk flexor activation (rectus abdominus) was per-formed by a resisted sit-up task, while resisted trunk rotation(external obliques) and extension tasks (erector spinae) werealso performed. The abdominal hollowing task was specifically

performed for the transversus abdominus–internal obliquessite, although the maximum activation obtained from either thiscontraction or the resisted rotation was used to define the MVCfor this signal. Two trials were performed for each MVC task,with 2 minutes rest allowed between each trial. The average of the 2 trials provided the value for normalization.

Frequency Spectrum

The median frequency (MF) of the electromyographic powerspectrum was calculated for each muscle signal for each trialwith a fast Fourier transform (FFT; 512-point Hamming win-dow). The MF was calculated as the point where the area of theFFT-derived spectrum was halved.

Ratios of Activity

Optimal stabilization has been considered to be increasedmuscle activation of the ventrolateral abdominals when com-pared with the rectus abdominus.2,7,14,15 To determine thesynergistic relation between the muscles in this experiment, wecalculated the ratio of the ventrolateral abdominal and erectorspinae muscle activity expressed relative to the rectus abdomi-nus for all trials, based on the percentage of MVC.

Task Difficulty

To evaluate the physical difficulty of each task, a 100-mmvisual analog scale (VAS; left anchor, very easy; right anchor,very hard) was administered after each task. Subjects and theexperimenters were blinded to the responses for each task

throughout the experiment.

Statistical Analysis

SPSS, version 11.5,c was used for data analysis. The intra-class correlation coefficient (ICC1,1)16 was calculated to assessthe reliability of the measurement between the 3 trials for eachtask. We used a repeated-measures analysis of variance(ANOVA; task by surface) for muscle activation, MF values,and VAS scores. Paired t tests were used to compare the ratioof activity between the rectus abdominus and the other musclesfor the stable and unstable conditions. The Bonferroni adjust-ment was applied to a priori pairwise comparisons, and Scheffépost-hoc analysis was used to determine where the differenceswere in the ANOVA if the main effect was significant. The

significance level of this study was set at P less than .05.

RESULTS

Reliability Between Trials

Table 1 shows the reliability data among the 3 trials for eachtest position. The ICC represents the relative variability be-tween trials, and the standard error of the mean the absolutevariability. All tasks and positions had strong ICC reliabilitybetween trials, apart from 2 tasks for the rectus abdominus(unstable roll-out; stable press-up bottom position).

Table 1: Reliability Analysis Among the 3 Trials Performed for Each Task, With the ICC and SEM Presented for the Relative Amplitude ofElectromyographic Activity for Each Muscle

Exercise

Muscle

condition

TA/IO RA EO ES

ICC SEM ICC SEM ICC SEM ICC SEM

Roll out Stable .90 2.94 .87 2.62 .99 5.48 .95 2.36

Unstable .95 4.54 .37 1.42 .99 6.39 .96 2.04

Press-up top Stable .98 2.38 .92 1.36 .99 3.86 .99 3.39

Unstable .99 6.47 .99 8.65 .94 2.67 .99 2.33

Press-up bottom Stable .96 2.75 .45 .75 .91 3.82 .98 3.52

Unstable .98 2.96 .94 2.63 .97 3.33 .99 2.45

Single-leg hold Stable .88 3.84 .84 1.59 .99 5.31 .99 2.39

Unstable .93 3.90 .68 5.18 .97 4.93 .99 2.26

Quadruped left arm/right leg Stable .99 2.03 .98 2.67 .99 6.71 .99 6.70

Unstable .97 3.21 .99 .18 .97 6.29 .84 2.71

Quadruped right arm/left leg Stable .99 1.52 .97 2.56 .98 6.66 .81 4.93

Unstable .99 1.24 .99 .40 .97 6.24 .92 3.07

Abbreviations: EO, external obliques; ES, erector spinae; RA, rectus abdominus; SEM, standard error of the mean; TA/IO, transversusabdominus/internal obliques.

244 CORE STABILITY EXERCISES ON AND OFF A SWISS BALL, Marshall




Electromyographic Amplitude Comparison Between

Surfaces and TasksTable 2 shows the RMS amplitude results expressed as a

percentage of MVC. For the transversus abdominus and inter-nal obliques, the activation at the top of the press-up on theunstable surface had the greatest activation. This activity dif-fered significantly from the same position on the stable surface(P.05). For comparison of the tasks for the transversus ab-dominus and internal obliques on the Swiss ball, the activity atthe top of the press-up was significantly greater than the activ-ity for the transversus abdominus and internal obliques for bothpositions in the quadruped exercise (P.05). There were nodifferences between the tasks for the activity of the transversusabdominus and internal obliques on the stable surface.

For the activity of the rectus abdominus, there were signif-

icant differences between the surfaces for both the press-up topposition and the single-leg hold, with the higher activity re-corded on the unstable surface (P.05). The activity of therectus abdominus during the aforementioned unstable surfacetasks was significantly greater than the activity for the rectusabdominus in any of the other test positions (P.05). Therewere no differences between the tasks on the stable surface forrectus abdominus activation.

There were no differences between the surfaces for theactivity of the external obliques and erector spinae during anytask. There were no significant differences between the tasksfor the external obliques activity. For the erector spinae, theactivity recorded during the quadruped exercise with left armand right leg raise differed significantly from the activity mea-sured during all other tasks (P.05). The activity during the

right arm/left leg quadruped exercise was significantly differentfrom the remaining tasks also (P.05). This pattern was con-sistent for both test surface conditions.

Task Difficulty

The unstable press-up was rated as the most difficult task performed in this experiment (82.754.43), and this ratingdiffered significantly from the rating of the press-up performedon the stable surface (51.1316.98, P.05) (fig 2). The onlyother exercise that showed a difference between the surfaceswas the roll-out task, with the unstable surface being rated asthe more difficult task to perform (unstable, 43.889.26; sta-ble, 31.759.47; P.05).

Ratio of Muscle Activity Compared With the Rectus

AbdominusThe ratio of the transversus abdominus and internal obliques

to the rectus abdominus activity did not change between thesurfaces for any of the tasks (fig 3). The ratio of activity of theexternal obliques compared with the rectus abdominus changedbetween the test surfaces for the press-up at the top position(stable, 5.581.6; unstable, 1.870.6; P.05) and for thesingle-leg hold (stable, 3.341.15; unstable, 1.610.90;P.05). The ratio of activity between the external obliques andrectus abdominus was significantly lower on the unstable sur-face for these tasks, indicating a greater relative activity levelof the rectus abdominus. In the erector spinae–rectus abdomi-nus comparison, there was reduced relative activity of theerector spinae compared with the rectus abdominus on the

unstable surface for the top of the press-up position (stable,1.480.4; unstable, 0.370.14; P.05) and for the single-leghold (stable, 1.16.36; unstable, 0.440.27; P.05).

MF Analysis

The significant results from the MF analysis of the powerspectrum are presented in table 3. There were no other signif-icant differences between tasks or surfaces for any muscle or

Table 2: Mean SD Average Normalized Surface Electromyographic Amplitudes (%MVC) for Each Muscle During the Tasks Evaluated

Exercise Muscle TA/IO RA EO ES

Roll out Stable 19.098.33 7.432.62 43.2115.50 11.986.69

Unstable 22.3618.85 4.021.42 40.9818.09 11.145.77

Press-up top Stable 12.636.74 8.383.85 42.910.92 9.623.40

Unstable 32.8818.31* 34.3824.48* 51.947.56 6.602.33

Press-up bottom Stable 17.317.78 7.752.12 42.1610.80 13.389.97

Unstable 19.698.38 9.257.44 47.539.41 13.636.93

Single-leg hold Stable 22.6610.87 14.034.52 41.6415.02 12.256.78Unstable 23.1511.01 31.5314.65* 40.9313.95 11.786.38

Quadruped left arm/right leg Stable 12.635.76 5.387.56 33.3818.98 33.9918.97

Unstable 14.509.07 2.630.52 35.8817.80 31.657.67

Quadruped right arm/left leg Stable 12.254.30 5.137.24 31.2518.25 21.7513.96

Unstable 13.433.50 3.031.12 34.6317.66 23.638.68

F value for interaction between

surface and exercise

2.37 (P .05) 7.26 (P .001) 0.29 (P .92) 0.09 (P .99)

NOTE. Significant differences are shown between the surfaces for the activation of that muscle during the particular task.*P .05.

Fig 2. Mean VAS results for the physical difficulty of each taskcomparing between the test surfaces. *P <.05.





between the first and third trials to indicate that there was nofatigue influence on the results of this experiment. There was asignificant decrease in the MF for the rectus abdominus fromthe unstable press-up at the top compared with the unstablepress-up bottom position (P.05). The MF also differed sig-nificantly for the 2 different movements performed for thequadruped exercise for the signal obtained from the erectorspinae. The left arm and right leg MF for both test surfaceconditions were significantly higher as compared with the rightarm and left leg (P.05).

DISCUSSION

In this study, we compared the activation levels of musclesof the lumbopelvic region during the performance of tasks onand off a Swiss ball. We also examined the relation betweenthe external obliques, transversus abdominus and internal ob-liques, erector spinae, and rectus abdominus by comparing therelative activity levels. Our results provide evidence supportingour hypothesis that the performance of tasks on the Swiss ballwould lead to greater activation levels when compared with the

Fig 3. Mean ratio of muscleactivity: the rectus abdominusfor each exercise, comparingthe relationship between testsurfaces for each task: (A) rollouts, (B) single leg hold, (C)press-up top position, (D)press-up bottom position, (E)quadruped with right arm andleft leg, and (F) quadrupedwith left arm and right leg.*P <.05. The ratio of 1:1 indi-cates equal relative activity ofthe comparison muscle: rec-tus abdominus.





stable surface. There was also evidence to suggest that specificexercises involve different synergistic relationships betweenthe muscles and that the Swiss ball can directly influence thoserelationships. This suggests that there should be a variety of exercises for a core stability training program.

Surface Comparison

Exercising on the Swiss ball increased the activity for therectus abdominus and transversus abdominus and internal ob-liques at the top of the press-up. There were no differencesbetween the surfaces for either muscle at the bottom of thepress-up. This suggests that the Swiss ball increased the per-turbation to the trunk when the body’s center of mass (COM)was further away from the labile surface. However, in therollout position, in which the COM is also away from the labilesurface, there was no difference between the surfaces formuscle activity. This may be because of the greater contactarea between subject and surface for the rollout, with the entireshank remaining in contact with the surface, compared with

just the palms of the hands in contrast for the press-up. If therollout position was held with only the feet in contact with theball, the reduction in contact area may be enough to cause

increases in muscle activity. Subjects rated both the rollout andpress-up as being more physically difficult when performed onthe unstable surface. This suggests that there are other musclesassociated with the rollout (eg, muscles of the shoulder girdle)that may have increased activity over that elicited on a stablesurface. However, the only muscles of interest in this studywere those of the lumbopelvic region that are associated withcore stability.

The result of the MF analysis suggests that the difference inthe rectus abdominus activation levels between the top andbottom positions of the unstable press-up may result frommuscle fatigue. This is because of the lower MF in the bottomposition compared with that at the top of the unstable press-up.A shift in MF toward the lower end of the frequency spectrumhas been associated with neuromuscular and physiologic mea-

sures of fatigue such as decreases in pH and decreases in motorunit conduction velocity and firing rates.13,17 The greater per-ceived difficulty of the unstable press-up in comparison withthe stable press-up may be attributed to the influence of mus-cular fatigue.

The activity of the rectus abdominus was also greater duringthe single-leg hold performed on the Swiss ball. This supportsour previous research in which we investigated a double-leghold that found increased rectus abdominus activity on theunstable surface (unstable, 54.916.23; stable, 42.6314.37;unpublished data, 2003). The increased activation of the rectusabdominus could be attributed to the greater hip flexion torquerequired to maintain the static equilibrium of the body on the

Swiss ball. The weight force of the leg causes torque about thehip that challenges the stability of the body, and this is coun-terbalanced by the activation of the hip flexors. An increase inhip flexor activation (rectus abdominus) is required to preventthe reactive movement of the ball to the weight force of the leg.From this it may be concluded that the Swiss ball causesinstability when a body segment is away from the center of theball sufficient to increase the activity of a prime mover asso-ciated with the task. The single-leg hold was used in this studyto cause a reactive rotation of the ball about the longitudinalaxis of the body that may have increased the activity of theventrolateral abdominals. Previous research found no differ-ence between the surfaces for the activation of the ventrolateralabdominal muscles, because these muscles cannot produce ahip flexion or extension torque (unpublished data, 2003). In thepresent study, there was no change in activation of the obliquemuscles with performance of the single-leg hold on the unsta-ble surface. This indicates that the weight force of the single-leg hold was insufficient to elicit an increase in oblique activityon the Swiss ball.

Clinical Relevance

Previous research has emphasized in previous research thatthe motor control and rehabilitation training of the ventrolateralabdominals is successfully achieved with exercises that mini-mize activation of the rectus abdominus.9,10,15,18 Activation of the ventrolateral abdominals has been associated with sacroil-iac joint laxity.19 Performance of the drawing-in technique hasalso been associated with the feed-forward activation of thetransversus abdominus before rapid limb movement.7 It hasbeen proposed that attempting to train lumbar stability byplacing importance on 1 set or group of muscles is not viable.Research has shown that no single muscle can be identified asbeing more important for spinal stability than another during arange of trunk movement tasks.11 The exercises evaluated inthis study provided no clear evidence for an obvious pattern of muscle recruitment associated with the performance of lumbar

stabilization exercises.The quadruped exercise with contralateral arm and leg raisereplicated the pattern of activity suggested for appropriatemotor control training of the ventrolateral abdominals.9,20 Theratio of muscle activity expressed relative to the rectus abdo-minus was the highest for each muscle for this task. Researchhas shown that the quadruped exercise had the highest mea-sured stability index as compared with several other corestability exercises.21 It has also been shown that the activity of the obliques is consistently greater than that of the rectusabdominus when extra resistance is added to the limbs for thistask.22 Therefore, the quadruped exercise fulfills the require-ment for a stabilization exercise with minimal rectus abdomi-

Table 3: Mean SD of MF Results Comparing Different Positions Within Set Tasks

Excercise

TA/IO RA EO ES

Stable Unstable Stable Unstable Stable Unstable Stable Unstable

Press up top 38.8110.14 37.432.91 51.778.68 62.127.83 47.447.19 42.529.98 51.5310.27 61.5118.72

Press up bottom 45.2320.54 40.919.88 44.6410.92 39.756.76* 42.6412.51 41.306.86 39.259.76 43.138.23

Quadruped left

arm/right leg 47.3119.12 56.938.75 52.8115.12 48.4427.64 46.808.91 45.733.58 97.2111.56 86.4510.52

Quadruped right

arm/left leg 68.2641.75 48.6319.24 68.3522.41 55.7135.83 55.317.20 56.1618.17 55.3812.12* 59.5324.73*

NOTE. The tasks compared above are the top and bottom positions of the inclined press-up, and the alternate limb movements for thequadruped exercise.*P .05.





nus activity in comparison with other muscles of the lumbopel-vic region. The abdominal drawing-in task and quadrupedexercise may be an effective combination of exercises fortraining the local stability system. Use of the Swiss ball wasinsufficient to change the activity patterns associated with thistask.

The activity level of the external obliques was unchangedregardless of the task or surface. This supports previous find-

ings that external oblique activity is unaffected by the task performed.23 Muscle activity up to 30% of MVC is required foran aerobic training effect to be achieved for the abdominalmuscles when the task is repeated.24 The activity level of theexternal obliques was approximately 40% for all tasks, sug-gesting that these exercises provide a training effect for thismuscle that is not enhanced by use of a Swiss ball.

The main effect of the Swiss ball was to increase the activityof the rectus abdominus to greater than 30% of MVC at the topof the press-up and during the single-leg hold. This suggeststhat the Swiss ball is a sufficient stimulus to provide a trainingeffect for the rectus abdominus. The relation of rectus abdo-minus activity to external oblique and erector spinae activitywas influenced by this increase in activity, meaning that the

synergistic relationship between these muscles has been al-tered. Our study found that the relative activity of the rectusabdominus increased in comparison with the external obliquesand erector spinae on the unstable surface for the aforemen-tioned tasks (see table 2, fig 3). As previously stated, Richard-son et al2,14,20 emphasize minimal activation of the rectusabdominus in comparison with other lumbopelvic muscles forstability exercises. If this is true, then an intervention thatincreases the activity of the rectus abdominus and changes thesynergistic activation patterns between the muscles may not beappropriate as a lumbar stability exercise.

The quadruped exercise showed a difference in the MFbetween the opposite movement directions for the right erectorspinae muscle signal, with the right leg and left arm beingsignificantly higher than the opposite movement. The right legand left arm movement also elicited a greater activation levelfor this muscle in comparison with the opposite side of the task.This difference was found for both test surface conditions. TheMF difference may be attributed to the different muscle lengthof the erector spinae with the different movements. Previousresearch25 has found that when the erector spinae are length-ened, there is a decrease in the MF. The shortening of thismuscle associated with the isometric hip extension of the rightleg is probably why there is a difference between the oppositesides for the MF.

A strength of the methodology of this experiment is that theelectromyography normalization procedures were done in aprone position similar to that used during the exercises. There-fore, the relative activity levels measured during the Swiss ballexercises reflects the maximal activity obtained in a similarposition. A potential limitation of the normalization procedureswe used is that they were based on MVC. It is recognized thatfor patients with LBP, the same exercise may lead to a greaterrelative activity level as a result of a distorted MVC. This canbe dealt with in a rehabilitation setting by beginning with alower number of repetitions of exercises that a patient cansuccessfully perform and that elicit a greater relative intensity.The utility of the MVC, as used in this study, was to allownormalization of muscle activity levels for comparison betweensurfaces and to show that the exercises studied can provide aneffective training stimulus.

CONCLUSIONS

The exercises presented here address issues regarding corestability training. The quadruped exercise replicates a patternof activity deemed appropriate for training the local stabilitysystem, with minimal activity of the rectus abdominus ascompared with other lumbopelvic muscles. In comparison, theSwiss ball increased rectus abdominus activity for the single-leg hold and at the top of the press-up. The unstable press-up

was also deemed to be the most physically difficult task. Aquestion to be addressed in future research is whether theincrease in the rectus abdominus activity caused by the Swissball is beneficial or whether minimizing rectus abdominusactivity is the priority for a core stability training program.

References1. Curtis C. Get your bounce on: use an exercise ball to enhance your

upper-body workouts. Muscle Fitness 2002;63:64.2. Richardson C, Jull G, Hodges P, Hides J. Therapeutic exercise for

spinal segmental stabilisation in low back pain. Sydney: ChurchillLivingstone; 1999. p 191.

3. Vera-Garcia FJ, Grenier SG, McGill SM. Abdominal muscleresponse during curl-ups on both stable and labile surfaces. PhysTher 2000;80:564-9.

4. Behm DG, Anderson K, Curnew RS. Muscle force and activationunder stable and unstable conditions. J Strength Cond Res 2002;16:416-22.

5. Bergmark A. Stability of the lumbar spine. A study in mechanicalengineering. Acta Orthop Scand 1989;230:20-4.

6. Hodges P, Richardson C. Inefficient muscular stabilization of thelumbar spine associated with low back pain: a motor controlevaluation of transversus abdominus. Spine 1996;21:2640-50.

7. Hodges P, Richardson C, Jull G. Evaluation of the relationshipbetween laboratory and clinical tests of transversus abdominisfunction. Physiother Res Int 1996;1:30-40.

8. McGill SM. Low back stability: from formal description to issuesfor performance and rehabilitation. Exerc Sport Sci Rev 2001;29:26-31.

9. Jull G, Richardson C, Toppenberg R, Comerford M, Bui B.Towards a measurement of active muscle control for lumbarstabilization. Aust J Physiother 1993;39:187-93.

10. Jull G, Richardson C, Hamilton C, Hodges P, Ng J. Towards thevalidation of a clinical test for the deep abdominal muscles in back pain patients. Paper presented at: The 9th Biennial Conference of the Manipulative Physiotherapists Association of Australia; 1995;Gold Coast (Aust).

11. Cholewicki J, Van Vliet JJ 4th. Relative contribution of trunk muscles to the stability of the lumbar spine during isometricexertions. Clin Biomech (Bristol, Avon) 2002;17:99-105.

12. Marshall PW, Murphy BA. The validity and reliability of surfaceEMG to assess the neuromuscular response of the abdominalmuscles to rapid limb movement. J Electromyogr Kinesiol 2003;13:477-89.

13. Farina D, Gazzoni M, Merletti R. Assessment of low back musclefatigue by surface EMG signal analysis: methodological aspects.J Electromyogr Kinesiol 2003;13:319-32.

14. Richardson C, Toppenberg R, Jull G. An initial evaluation of eightabdominal exercises for their ability to provide stabilisation for thelumbar spine. Aust J Physiother 1990;36:6-11.

15. O’Sullivan PB, Twomey L, Allison GT. Altered abdominal mus-cle recruitment in patients with chronic back pain following aspecific exercise intervention. J Orthop Sports Phys Ther 1998;27:114-24.

16. Rankin G, Stokes M. Reliability of assessment tools in rehabili-tation: an illustration of appropriate statistical analyses. Clin Re-habil 1998;12:187-99.





17. Allison GT, Fujiwara T. The relationship between EMG medianfrequency and low frequency band amplitude changes at differentlevels of muscle capacity. Clin Biomech (Bristol, Avon) 2002;17:464-9.

18. O’Sullivan PB, Twomey L, Allison GT. Evaluation of specificstabilizing exercise in the treatment of chronic low back pain withradiologic diagnosis of spondylolysis or spondylolisthesis. Spine1997;22:2959-67.

19. Richardson CA, Snijders CJ, Hides JA, Damen L, Pas MS, Storm

J. The relation between the transversus abdominus muscles, sac-roiliac joint mechanics, and low back pain. Spine 2002;27:399-405.

20. Richardson C, Jull G, Toppenberg R, Comerford M. Techniquesfor active lumbar stabilisation for spinal protection: a pilot study.Aust J Physiother 1992;38:105-12.

21. McGill SM, Grenier S, Kavcic N, Cholewicki J. Coordination of muscle activity to assure stability of the lumbar spine. J Electro-myogr Kinesiol 2003;13:353-9.

22. Souza GM, Baker LL, Powers CM. Electromyographic activity of selected trunk muscles during dynamic spine stabilization exer-cises. Arch Phys Med Rehabil 2001;82:1551-7.

23. Beith ID, Synnott RE, Newman SA. Abdominal muscle activityduring the abdominal hollowing manoeuvre in the four pointkneeling and prone positions. Man Ther 2001;6:82-7.

24. Arokoski JP, Kankaanpaa M, Valta T, et al. Back and hip extensormuscle function during therapeutic exercises. Arch Phys MedRehabil 1999;80:842-50.

25. Mannion AF, Dolan P. The effects of muscle length and forceoutput on the EMG power spectrum of the erector spinae. JElectromyogr Kinesiol 1996;6:159-68.

Suppliersa. 3M, 3M Ctr, St Paul, MN 55144.b. Analog Digital Instruments, Unit 6, 4 Gladstone Rd, Castle Hill,

NSW, 2154, Australia.c. SPSS Inc, 233 S Wacker Dr, 11th Dr, Chicago, IL 60606.



core stability exercises on and off a swiss ball

Documents