Secondary Logo

Journal Logo

Original Research

Effects of BOSU Ball(s) During Sit-Ups With Body Weight and Added Resistance on Core Muscle Activation

Saeterbakken, Atle H.1; Andersen, Vidar1; Jansson, June1; Kvellestad, Ann C.1; Fimland, Marius S.2,3

Author Information
Journal of Strength and Conditioning Research: December 2014 - Volume 28 - Issue 12 - p 3515-3522
doi: 10.1519/JSC.0000000000000565
  • Free



A variety of core exercises are used to develop core stability and core strength, which are believed to be important for athletic performance, general health, and preventing low back pain (3,5,13,18,20). Curl-ups and sit-ups are the most common abdominal exercises. During the past decades, unstable surfaces, such as a Swiss ball, wobbler board, and the BOSU ball, have been used to increase the neuromuscular stress in the core muscles compared with a stable surface (7,14,15,20). Performing exercises on an unstable surface has been proposed to increase proprioceptive demands and stress the muscles to a greater extent than performing exercises on a stable surface (3,14,20), but the reports are mixed (7,9,15,19,20).

In rehabilitation, BOSU or Swiss balls may be beneficial, because the lower back is supported while performing sit-ups compared with performing on a flat surface. In addition, there are several reports of greater core muscle activation when performing exercises on an unstable surface compared with performing exercises on a stable surface (10,14,20). However, these investigations compared only the same absolute loads. This may be problematic because there is ample evidence that inducing instability reduces absolute strength and force production during single-joint and multijoint resistance exercises (1,2,12,16,17). Thus, using the same absolute load means that higher relative loads (percentage of 1RM) were used in the unstable conditions, which may explain the increased muscle activity compared with that for a stable surface (12).

Studies examining muscle activation in isolated core exercises (e.g., sit-ups or curl-ups) on stable and unstable surfaces are limited. Sundstrup et al. (19) examined crunches (10RM loading) on a Swiss ball and a training machine and demonstrated greater rectus abdominis activation using the Swiss ball. Further, Vera-Garcia et al. (20) compared 4 conditions with different stability requirements during isometric curl-ups and demonstrated a greater rectus abdominis and external oblique electromyographic (EMG) activity on a Swiss ball than on a stable surface. Finally, Lehman et al. (9) examined 4 repetitions of curl-ups on a stable bench and Swiss ball and demonstrated a similar core muscle activation comparing the surfaces. Hence, only 1 previous study compared matched resistance (19), and to our knowledge, there have not been any investigations of isolated abdominal exercises performed on BOSU ball(s).

The purpose of this study was to assess core muscle activation during sit-ups with various stability requirements induced by BOSU balls with body weight and with added, matched resistance. We hypothesized that (a) BOSU balls would increase abdominal activation during sit-ups with body weight, that is, with the same absolute resistance; and (b) abdominal activation would be similar when using added, matched resistance.


Experimental Approach to the Problem

A within-subject, repeated-measures study design was used to assess the neuromuscular activity of superficial core muscles during sit-ups exercise performed with 10 repetitions with body weight and with 10 repetition maximum (10RM) under 4 different conditions in a randomized and counterbalanced order: (a) on a stable surface, (b) with the BOSU ball under the feet (dome side down, lower-body instability), (c) BOSU ball under the low back (dome side up, upper-body instability), and (d) with BOSU balls under the feet and the low back (dual instability). Before the experimental test, the 10RM in the sit-up exercises was tested in a familiarization session. An elastic tube ( provided added resistance (19).


Twenty-four healthy men (age: 23 ± 2 years, mass: 77.5 ± 8.2 kg, height: 1.79 ± 0.06 m) with at least 2 years of resistance training experience volunteered to participate in the study. The exclusion criteria were musculoskeletal pain that affected performance (e.g., low back pain) or not being able to perform the exercises with proper technique (see Procedures).

Ethics approval was obtained from the local research ethics committee before the testing. All the participants were informed of testing procedures and possible risks involved, and provided written consent to participate. None of the participants were competitive weightlifters or power lifters, but had 5.2 (±2.5) years of previous resistance-training experience. The participants were instructed to refrain from any additional resistance exercise in the 72 hours before testing.


Ten-minute warm-ups were performed before sessions on a treadmill or cycle ergometer at an intensity at which the participants could talk comfortably. In the familiarization session, the order of the conditions was randomized and counterbalanced. The same order was used in the experimental session. The sit-ups with body weight were tested before the 10RM tests. A 4-minute rest period was given between the 2 test intensities. Before an exercise, the subjects placed their feet shoulder width apart with a 90° flexion in the knees and the arms held across the chest. From a supine position, the upper body was lifted 45° (controlled using a protractor). To prevent the participants from performing greater or smaller lifting angles, a horizontal band was adjusted so that the arms touched the band when the upper body was lifted 45°. During the sit-ups, the participants had to touch the band while keeping their arms in contact with their chest before returning to the supine position (Figure 1A). The participants were instructed to perform the sit-ups in a controlled and similar tempo for each exercise without pauses in the upper or lower position. A natural sway in the lower back was maintained in both lifting phases.

Figure 1
Figure 1:
A) Sit-ups on a stable surface, B) lower-body instability, C) upper-body instability, and D) dual instability.

Sit-ups on a stable surface were performed on a 2-cm-thick gymnastic mat (Figure 1A). The participants were not allowed to start the concentric phase by lifting their hip. The eccentric phase ended when the participants touched the gymnastic mat with their shoulder blades. During sit-ups with lower-body instability, a BOSU ball was placed underneath the feet (Figure 1B). To ensure the same hip and knee angles, the upper body was placed on a step box with the gymnastic mat on top. The upper and lower body was then at the same height. The same procedure was used for the final 2 exercises. During sit-ups with upper-body instability, the center of the BOSU ball was placed under the low back (Figure 1C) while the feet were placed on the step box. During sit-ups with dual instability, BOSU balls were placed underneath the both the upper and lower body (Figure 1D). The air pressure of the BOSU balls was adjusted so that the horizontal height between the BOSU balls and BOSU ball and step box was the same with an 80-kg participant lying supine on them. A band was used for the upper-body instability and dual instability trials to control that the participants ended the eccentric phase when the upper body reached horizontal position.

During sit-ups with 10RM, an elastic tube ( was attached to the floor and used to increase the resistance until the 10RM load was established for each exercise (19). The elastic tube provided a variable resistance that increased as the participants progressed in the range of motion. Using added weights on the chest would provide a constant resistance that would make it difficult to start the movement when the feet are not attached to the floor, because of the greater lever arm in this position. However, throughout the movement, the lever arm decreases in sit-ups, which make elastic bands a feasible resistance modality. The resistance was increased by reducing the distance from the participants and the attachment point of the elastic tube. Greater resistance was provided by increasing the distance executing the sit-ups and the attachment point to the elastic band. There was a linear relationship between stretching length and resistance from the elastic band (R2 = 0.972) measured by a force cell (Ergotest Technology A/S, Langesund, Norway). A 4-minute rest period was given between each 10RM attempt (6). The 10RM resistance was established within 1–3 attempts. The 10RM resistance from the elastic band was measured at the upper position of the sit-ups.


The positions and lifting time during the sit-ups were measured by a linear encoder (Ergotest Technology A/S) attached to the forefinger of the participants. The linear encoder was synchronized with the EMG recordings using a 4020e Musclelab (Ergotest Technology A/S) and was used to identify the beginning and ending of the first and last repetitions. Commercial software (Musclelab V8.13; Ergotest Technology AS) was used to calculate the mean root-mean-square (RMS) EMG activities. The EMG activity was calculated as the mean area of the beginning of first repetition to the end of the 10th repetition.

Before the electrode placement, the skin was prepared (shaved, washed with alcohol, abraded). Self-adhesive gel-coated electrodes (Dri-Stick Silver circular sEMG Electrodes AE-131; NeuroDyne Medical, Cambridge, MA, USA) were used in the experimental test. The electrodes (11-mm contact diameter) were placed in the presumed direction of the underlying muscle fibers with a center-to-center distance of 2.0 cm, according to the recommendations of the Surface ElectroMyoGraphy for the Non-Invasive Assessment of Muscles (8). The surface EMG electrodes were positioned on the lower region of the rectus abdominis (3 cm lateral and 1 cm inferior to the umbilicus) and the upper region (3 cm lateral and 5 cm superior to the umbilicus), and on the external oblique (∼15 cm lateral to the umbilicus) (20). To minimize noise induced by external sources, the raw EMG signal was amplified and filtered using a preamplifier located as near the pickup point as possible. The preamplifier had a common mode rejection ratio of 100 dB. The signals were high pass and low pass (600, 8 Hz) (fourth-order Butterworth filter). The filtered EMG signals were RMS converted using a hardware circuit network (frequency response 0–600 kHz, averaging constant 100 milliseconds, total error ±0.5%). The RMSs were sampled at a rate of 100 Hz using a 16-bit analog-to-digital converter. The stored data were analyzed using commercial software (Musclelab V8.13; Ergotest Technology AS).

Statistical Analyses

To assess differences in the EMG activity in sit-up exercises, we used 2 separate (body weight; 10RM conditions) 2-way (4 exercises × 3 muscles) analysis of variance (ANOVA). To assess differences in EMG activity between repetitions 1 and 2 compared with repetitions 9 and 10, we used a 2-way (4 exercises × 2 test times) ANOVA for the 2 testing intensities (10 repetitions and 10RM). When a significant interaction was found, paired t-tests with Bonferroni post hoc corrections was used to locate the differences. To assess differences in the time to perform the exercises, we conducted a repeated-measures 1-way ANOVA with Bonferroni post hoc corrections tests. All results are presented as mean ± SD and Cohen's d effect size (ES) unless otherwise noted. An ES of 0.2 was considered small, 0.5 was considered medium, and 0.8 was considered large. SPSS (v 20.0; SPSS Inc., Chicago, IL, USA) was used. Statistical significance was accepted at p ≤ 0.05.


When comparing the muscle activities obtained with body weight, an interaction between exercise and muscle was observed (p ≤ 0.001, F = 6.779). Post hoc analysis demonstrated similar EMG activities in the lower and upper parts of the rectus abdominis between the exercises (p = 0.370–1.000, Table 1). However, for the external oblique, the EMG activity for the stable surface was 131 and 128% compared with that for upper-body and dual instabilities and the lower-body instability was 129 and 127% compared with that for upper-body and dual instabilities (p = 0.002–0.006, ES = 0.49–0.56, Table 1, Figure 2A). The average EMG activities of repetitions 1–2 and 9–10 are presented in Figure 3A. There was no main effect of time or interactions between muscle and time (p = 0.542–0.970).

Table 1
Table 1:
Electromyographic activity in millivolts for 10 repetition loading comparisons.
Figure 2
Figure 2:
Electromyographic activity expressed as a percentage of the stable surface examining (A) 10 repetitions and (B) 10RM loading.
Figure 3
Figure 3:
The average electromyographic activity during repetitions 1–2 and 9–10 examining (A) 10 repetitions and (B) 10RM loading.

For the 10RM loading, a significant interaction was observed between exercise and muscle (p = 0.020, F = 2.606). Post hoc corrections demonstrated a similar EMG activity in the external oblique between the exercises (p = 0.054–1.000), with a strong tendency of a greater EMG activity in dual instability compared with that in lower-body instability (p = 0.054, ES = 0.31, Table 2, Figure 2B). In the lower region of the rectus abdominis, the EMG activity for the stable surface was 80–81% compared with that for both upper-body instability and dual instability, and for lower-body instability, it was 79–80% compared with that for both upper-body instability and dual instability (p < 0.001–0.001, ES = 0.49–0.56, Table 2, Figure 2B). In the upper region of the rectus abdominis, the EMG activities for the stable surface and lower-body instability were 82 and 81% compared with that for the upper-body instability (p = 0.016–0.036, ES = 0.39–0.49, Table 2, Figure 2B). The average EMG activities of repetitions 1–2 and 9–10 are presented in Figure 3B. There was no main effect of time or interactions between muscle and time (p = 0.055–0.507).

Table 2
Table 2:
Electromyographic activity in millivolts for 10RM loading comparisons.

When performing sit-ups with 10RM loading, resistance was added by 19 ± 10 N on the stable surface, by 18 ± 10 N with lower-body instability, by 51 ± 20 N with upper-body instability, and by 50 ± 10 N with dual instability. The total lifting time was similar for performing the sit-ups with body weight (19.1 ± 4.3 seconds–22.5 ± 4.5 seconds, p = 0.101–1.000) or 10RM (18.4 ± 4.5 seconds–21.4 ± 3.6 seconds, p = 0.116–1.000).


The main findings of this study are that, with body weight, external oblique activation was decreased by upper-body instability and dual instability, whereas the rectus abdominis was not affected by the surface. Using 10RM loads, the upper and lower rectus abdominis activities were increased by upper body and dual instability compared with that for stable surface. Further, lower-body instability did not significantly affect muscle activities with either load.

These results are partly in contrast to our expectations, because several studies have earlier reported an increased neuromuscular activity on unstable surfaces when using the same absolute load (10,14,20), which usually is attenuated when the intensity is matched (11,12,19). This is not surprising considering that the force output typically decreases with increasing instability (1,2,16,17). Therefore, it was opposite to our expectations that we observed reduced activity of external oblique with upper-body and dual instability with body weight (i.e., same absolute load), but not with 10RM loading (i.e., same relative load). Similarly, rectus abdominis activation was only increased for upper-body instability with the 10RM loading and not with body weight. The reason for these seemingly counterintuitive findings seems to be that placing a BOSU-ball dome side up under the low back makes sit-ups exercise easier to perform, at least under the present experimental conditions, as evidenced by the higher load that could be added in these 10RM tests. When sit-ups are performed on a stable surface, a depression of the pelvis is impossible, whereas on a BOSU ball, the pelvis is increasingly depressed toward the end of the concentric phase providing greater support of the lower back, which decreases the lever arm—making the exercise easier to perform (Figure 1A, D).

Although a higher load could be added for sit-ups on BOSU balls under the low back than on the stable surface (50 vs. 20 N), both loads were quite modest additions. The reason for modest differences in the 10RM load is largely because the feet were not attached to the surface. Therefore, the limiting factor in the 10RM conditions was not abdominal or hip flexor strength but simply that this position did not allow the generation of very high forces because generation of high forces would cause the feet to be elevated from the ground. The 10RM EMG results and the fact that there were no changes in the EMG activity from the beginning to the end of the set for either condition must be interpreted in light of this. Additionally, placing a BOSU ball under the feet with the dome side down had little impact on the EMG activity and strength demands of the exercise. This could be explained by the low forces being applied on the surface.

With body weight, the external oblique activation was reduced in the less strength demanding upper-body and dual instability conditions compared with that for a stable surface, whereas this reduction was not observed in the 10RM condition. It could be that with submaximal loads, the primary lumbar spine flexor muscle, rectus abdominis, require less assistance from the assisting spine flexor external oblique muscle.

The finding in our study of similar rectus abdominis activation with body weight on different surfaces is in line with the results reported for the Swiss ball by Vera-Garcia et al. (20) and Lehman et al. (9). Vera-Garcia et al. (20) demonstrated a similar EMG activity in the lower and upper parts of the rectus abdominis among the stable surface, wobble board, and Swiss ball. Similar to that done in this study, Lehman et al. (9) demonstrated similar rectus abdominis activation between curl-ups on a stable surface and on a Swiss ball. Further, our finding of higher activation of the rectus abdominis with a BOSU ball under the low back using 10RM loads but similar for external oblique, compared with a stable surface, is similar to the findings reported by Sundstrup et al. (19). These authors demonstrated the same activation patterns for the Swiss ball compared with a training machine. Arguably, a BOSU ball with the dome side down is quite similar to a Swiss ball, whereas performing sit-ups on the floor bears more resemblance to exercising on training machines.

There are some strengths and limitations to this study. Dynamic sit-ups were performed with both matched absolute (body weight) and relative intensity (10RM), in contrast to those performed in previous studies that only chose either of the methods (9,19,20). Further, for EMG measurements, there is always an inherent risk of crosstalk between surrounding muscles even if a small interelectrode distance was used (8,21). In addition, there are methodological concerns with dynamic EMG measurements (4). Importantly, we made an attempt to standardize movement distance and movement time, and these were similar for the different tests. Also, it is an advantage that all EMG measurements were made in the same session, so electrodes were identically placed in all experimental tests. Further, the participants were healthy recreationally trained men, and the results may not necessarily be generalized to other populations, such as persons with low back pain.

In conclusion, placing a BOSU ball under the low back increased or decreased abdominal muscle activation, depending on the loading, whereas placing a BOSU ball under the feet with the dome side down had little impact. When the feet are not attached to the surface, placing a BOSU ball under the low back makes it easier to perform sit-ups. With added resistance, placing a BOSU ball under the low back increased rectus abdominis activation, whereas this was not observed for body weight.

Practical Applications

It has been proposed that performing exercises on unstable surfaces increases the neuromuscular activation in the core muscles to a greater extent than performing exercises on a stable surface, but the reports are mixed. This study demonstrated that performing sit-ups on a BOSU ball is easier than performing sit-ups on the floor, at least when the feet are not attached to the surface. Therefore, in novice training, BOSU balls could be introduced before progressing to a stable surface. Placing a BOSU ball under the feet with the dome side down had little influence on the strength demands of sit-ups and concomitant abdominal activation and therefore has little practical relevance. When resistance was applied via an elastic band, so that the same training intensity was being compared (10RM), the rectus abdominis activation was increased by approximately 20% with the BOSU ball placed under the low back compared with exercising on a stable surface.


The authors thank the participants for their enthusiastic participation. This study was conducted without any funding from companies, manufacturers, or outside organizations.


1. Anderson KG, Behm DG. Maintenance of EMG activity and loss of force output with instability. J Strength Cond Res 18: 637–640, 2004.
2. Behm DG, Anderson K, Curnew RS. Muscle force and activation under stable and unstable conditions. J Strength Cond Res 16: 416–422, 2002.
3. Cosio-Lima LM, Reynolds KL, Winter C, Paolone V, Jones MT. Effects of physioball and conventional floor exercises on early phase adaptations in back and abdominal core stability and balance in women. J Strength Cond Res 17: 721–725, 2003.
4. Farina D. Interpretation of the surface electromyogram in dynamic contractions. Exerc Sport Sci Rev 34: 121–127, 2006.
5. Garcia-Vaquero MP, Moreside JM, Brontons-Gil E, Peco-Gonzalez N, Vera-Garcia FJ. Trunk muscle activation during stabilization exercises with single and double leg support. J Electromyogr Kinesiol 22: 398–406, 2012.
6. Goodman CA, Pearce AJ, Nicholes CJ, Gatt BM, Fairweather IH. No difference in 1RM strength and muscle activation during the barbell chest press on a stable and unstable surface. J Strength Cond Res 22: 88–94, 2008.
7. Hamlyn N, Behm DG, Young WB. Trunk muscle activation during dynamic weight-training exercises and isometric instability activities. J Strength Cond Res 21: 1108–1112, 2007.
8. Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol 10: 361–374, 2000.
9. Lehman GJ, Gordon T, Langley J, Pemrose P, Tregaskis S. Replacing a Swiss ball for an exercise bench causes variable changes in trunk muscle activity during upper limb strength exercises. Dyn Med 4: 6, 2005.
10. Marshall PW, Murphy BA. Increased deltoid and abdominal muscle activity during Swiss ball bench press. J Strength Cond Res 20: 745–750, 2006.
11. McBride JM, Cormie P, Deane R. Isometric squat force output and muscle activity in stable and unstable conditions. J Strength Cond Res 20: 915–918, 2006.
12. McBride JM, Larkin TR, Dayne AM, Haines TL, Kirby TJ. Effect of absolute and relative loading on muscle activity during stable and unstable squatting. Int J Sports Physiol Perform 5: 177–183, 2010.
13. McGill SM. Low back exercises: Evidence for improving exercise regimens. Phys Ther 78: 754–765, 1998.
14. Norwood JT, Anderson GS, Gaetz MB, Twist PW. Electromyographic activity of the trunk stabilizers during stable and unstable bench press. J Strength Cond Res 21: 343–347, 2007.
15. Nuzzo JL, McCaulley GO, Cormie P, Cavill MJ, McBride JM. Trunk muscle activity during stability ball and free weight exercises. J Strength Cond Res 22: 95–102, 2008.
16. Saeterbakken AH, Fimland MS. Electromyographic activity and 6RM strength in bench press on stable and unstable surfaces. J Strength Cond Res 27: 1101–1107, 2013.
17. Saeterbakken AH, Fimland MS. Muscle force output and electromyographic activity in squats with various unstable surfaces. J Strength Cond Res 27: 130–136, 2013.
18. Saeterbakken AH, van den Tillaar R, Seiler S. Effect of core stability training on throwing velocity in female handball players. J Strength Cond Res 25: 712–718, 2011.
19. Sundstrup E, Jakobsen MD, Andersen CH, Jay K, Andersen LL. Swiss ball abdominal crunch with added elastic resistance is an effective alternative to training machines. Int J Sports Phys Ther 7: 372–380, 2012.
20. Vera-Garcia FJ, Grenier SG, McGill SM. Abdominal muscle response during curl-ups on both stable and labile surfaces. Phys Ther 80: 564–569, 2000.
21. Winter DA, Fuglevand AJ, Archer SE. Crosstalk in surface electromyography: Theoretical and practical estimates. J Electromyogr Kinesiol 4: 15–26, 1994.

trunk; unstable surfaces; instability; EMG

Copyright © 2014 by the National Strength & Conditioning Association.