Unstable Surface Training: More Activation, Less Force?

Stand on a BOSU ball with a pair of dumbbells and everything shakes. Your legs burn, your trunk works overtime, and the set feels brutally hard. The question serious lifters and clinicians keep asking is whether that feeling translates into more strength and more muscle, or whether it is simply a harder way to lift less.

Why this matters

Instability devices are everywhere: Swiss balls, BOSU domes, foam pads, wobble boards, and more recently "unstable loads" such as kettlebells hung from bands or flexible bars. The sales pitch is intuitive. If the surface moves, more muscles have to switch on to control it, so the exercise must be more effective. The Canadian Society for Exercise Physiology examined exactly this claim in a formal position stand and reached a more careful conclusion: instability does raise core and limb muscle activation in many exercises, but it also reduces force, power, velocity and range of motion, so it should not be the primary mode of training for athletic strength.¹

That tension, more activation but less output, is the whole story. Understanding it helps you decide when an unstable variation earns a place in a program and when it quietly steals the load that drives adaptation.

The force penalty is large and consistent

The most reproducible finding in this literature is that you simply cannot push as hard on an unstable base. In an early study from Memorial University, seated knee extension force fell by 70.5% when participants sat on a Swiss ball instead of a bench, and plantar flexor force fell by 20.2%.² The difference between the two came down to contact points: the more the body is free to move, the larger the penalty.

The same group found that maximal isometric chest press force was 59.6% lower on a Swiss ball than on a bench.³ An isometric squat performed on inflatable balls reduced peak force by 45.6% and rate of force development by 40.5%.⁴ These are not small differences. Force and rate of force development are the outputs strength training is trying to improve, and instability cuts them sharply.

The size of the penalty depends heavily on the device. When trained men performed maximal isometric squats on four surfaces, a power board produced force similar to the floor (7% lower, not significant), while a BOSU ball and a balance cone reduced force by 19% and 24% respectively.⁵ In dynamic lifting, the penalty is smaller but persistent. Six-repetition maximum bench press loads fell to about 93% of stable values on a balance cushion and about 92% on a Swiss ball,⁶ and a foam cushion reduced squat and Bulgarian split squat loads by 7% and 10%.⁷ In practice, the more chaotic the surface, the less weight you can move.

What happens to muscle activation

Here the picture becomes more nuanced, and it is where surface EMG earns its place. The claim that instability "activates more muscle" holds for some muscles and fails for others.

For the prime movers, activation is usually maintained or reduced, not increased. In the Swiss ball chest press study, overall EMG activity of the pectoralis major, anterior deltoid, triceps and latissimus dorsi did not differ between stable and unstable conditions, even though force was almost 60% lower.³ During the isometric squat on inflatable balls, vastus lateralis and vastus medialis activity was 37.3% and 34.4% lower than on the floor.⁴ In trained men bench pressing their six-repetition maximum, pectoralis major EMG fell to roughly 81% of the stable value on a Swiss ball, and triceps EMG fell to about 69%.⁶ On a BOSU ball or balance cone squat, most lower-limb muscles kept similar activation, but rectus femoris activity dropped.⁵

For the stabilisers, the trend runs the other way. Squatting on two balance discs produced the highest activity in the soleus, abdominal stabilisers and both upper and lower erector spinae, with the Smith machine producing the lowest.⁸ Bench pressing a light load with the shoulders, feet or both on unstable surfaces progressively increased trunk stabiliser EMG, with dual instability producing the greatest activation.⁹ Rectus abdominis EMG during a six-repetition maximum bench press was also higher on a Swiss ball.⁶

Even this has limits. A meta-analysis of 33 EMG studies on periscapular muscles found that adding an unstable surface produced only a trivial increase in upper trapezius activity, no meaningful change in middle or lower trapezius, and a small decrease in serratus anterior activity.¹⁰ Instability is not a reliable way to switch on every stabiliser you care about.

Unstable loads, rather than unstable surfaces, behave somewhat differently. Pressing a flexible bar with oscillating weights raised activity in the biceps and middle deltoid while using 15% less load than a standard barbell, with prime mover activation broadly preserved.¹¹ Because the lifter's base remains stable, the force penalty is smaller.

Activation is not the same as stimulus

It is tempting to read "same EMG, less weight" as a free lunch: equal muscle activity with lighter loads. That interpretation needs care. Surface EMG amplitude reflects how strongly the nervous system is driving a muscle, but it is not a direct measure of mechanical tension, and it does not predict growth on its own. When EMG stays the same while force falls by half, the more accurate description is that neuromuscular efficiency has dropped. The muscle is working just as hard to produce much less force, much of it spent on control rather than output.

Sparkes and Behm measured this directly. Stable isometric chest press produced 42.2% more force than unstable, alongside 43.2% and 33.2% greater neuromuscular efficiency in the triceps and pectoralis.¹² There is also a co-contraction cost. On the Swiss ball, antagonist to agonist activation ratios rose by 40.2% in the knee extensors and 30.7% in the plantar flexors.² More antagonist activity means more internal braking, which is one reason output falls.

What training studies actually show

Acute studies tell you what happens during a set. Training studies tell you whether it matters over weeks. Here the evidence is more reassuring for instability than the acute force data would suggest, at least within limits.

In an eight-week program with recreationally active adults, stable and unstable resistance training produced similar improvements in isometric chest press force, bench press, squat and jump performance, with no significant group differences.¹² The unstable group tended to improve its unstable to stable force ratio more, which is unsurprising: you get better at what you practise.

Task specificity was the clearest finding of a ten-week study in resistance-trained men who trained the chest press on a Swiss ball, a Smith machine or with dumbbells.¹³ The Swiss ball and dumbbell groups improved most in their own exercise, and much of that improvement appeared in the first seven sessions, which points to skill and coordination rather than muscle. EMG activity of the prime movers did not differ between groups or across time.

The most complete comparison followed 59 men through three weeks of familiarisation and seven weeks of squat training with low (Smith machine), medium (free weight) or high (wobble board) stability.¹⁴ All three groups improved muscle thickness, rate of force development and maximal isometric force to a similar degree, and the free weight and wobble board groups improved ten-repetition maximum similarly across conditions. Only the wobble board group improved jump height on an unstable surface. In other words, moderate instability did not stop people getting stronger or bigger over a short program, but it did not outperform free weights either, and the main unique benefit was specific to unstable tasks.

For trunk-focused work in athletes, a 2025 systematic review and meta-analysis of 12 studies reported that instability core training improved trunk strength and sprint performance more than traditional stable core training, with moderate to large effects.¹⁵ That is a narrower claim than "unstable lifting builds more strength", and study quality was moderate, but it suggests instability has a defensible role in core and athletic preparation.

What this means in practice

For a lifter whose goal is maximal strength or hypertrophy of the prime movers, the evidence points one way. Do your heavy, progressive work on a stable base. The position stand is explicit that ground-based free weight lifts such as squats and deadlifts produce core activation similar to or higher than unstable exercises while still allowing heavy loads.¹ Standing on a BOSU ball to squat trades away load and force for a stabilisation challenge your trunk can largely get from the barbell itself.

Instability is more useful as a deliberate, secondary tool. It fits well in rehabilitation and return-to-training phases where high loads are not appropriate, in core and balance work, in warm-ups, and for athletes whose sport demands force production on a moving base. Moderate devices, such as a balance cushion rather than a Swiss ball, keep more of the load.⁵,⁶ Unstable loads on a stable base keep more of the force while adding a control challenge.¹¹

The practical problem is that you cannot feel the difference between "harder" and "more productive". A wobbling set feels intense whether or not your target muscle is doing more work. This is where a muscle-worn wearable adds interpretive value. With a ZELOS sensor on the target muscle, you can compare the activation of a stable variation against an unstable one relative to your own baseline, and the motion sensor can show how much of the set was spent correcting sway rather than moving the load. If the unstable version drops your pectoralis or quadriceps signal and slows your reps, that is a clear sign it belongs in your accessory or rehab work, not in your main lift. What the signal cannot tell you is how much a single set will make you grow. Activation is one input, read alongside load, effort and progression over time.

Key takeaways

• Unstable surfaces consistently reduce force output, from roughly 7–10% on foam pads to 45–70% on Swiss balls and inflatable bases.²,³,⁴,⁷

• Prime mover EMG is usually maintained or reduced, while trunk and postural stabiliser EMG tends to rise.³,⁶,⁸,⁹

• Equal EMG with less force means lower neuromuscular efficiency and more co-contraction, not a hidden hypertrophy bonus.²,¹²

• Over weeks, moderate instability produces strength and size gains similar to free weights, but largely task-specific advantages.¹³,¹⁴

• Keep heavy progressive work stable and use instability deliberately for core, balance, rehab and sport-specific preparation.¹,¹⁵

References

1. Behm, D. G., Drinkwater, E. J., Willardson, J. M., & Cowley, P. M. (2010). Canadian Society for Exercise Physiology position stand: The use of instability to train the core in athletic and nonathletic conditioning. Applied Physiology, Nutrition, and Metabolism, 35(1), 109–112. https://doi.org/10.1139/H09-128

2. Behm, D. G., Anderson, K., & Curnew, R. S. (2002). Muscle force and activation under stable and unstable conditions. Journal of Strength and Conditioning Research, 16(3), 416–422.

3. Anderson, K. G., & Behm, D. G. (2004). Maintenance of EMG activity and loss of force output with instability. Journal of Strength and Conditioning Research, 18(3), 637–640.

4. McBride, J. M., Cormie, P., & Deane, R. (2006). Isometric squat force output and muscle activity in stable and unstable conditions. Journal of Strength and Conditioning Research, 20(4), 915–918. https://doi.org/10.1519/R-19305.1

5. Saeterbakken, A. H., & Fimland, M. S. (2013). Muscle force output and electromyographic activity in squats with various unstable surfaces. Journal of Strength and Conditioning Research, 27(1), 130–136. https://doi.org/10.1519/JSC.0b013e3182541d43

6. Saeterbakken, A. H., & Fimland, M. S. (2013). Electromyographic activity and 6RM strength in bench press on stable and unstable surfaces. Journal of Strength and Conditioning Research, 27(4), 1101–1107. https://doi.org/10.1519/JSC.0b013e3182606d3d

7. Andersen, V., Fimland, M. S., Brennset, O., Haslestad, L. R., Lundteigen, M. S., Skalleberg, K., & Saeterbakken, A. H. (2014). Muscle activation and strength in squat and Bulgarian squat on stable and unstable surface. International Journal of Sports Medicine, 35(14), 1196–1202. https://doi.org/10.1055/s-0034-1382016

8. Anderson, K., & Behm, D. G. (2005). Trunk muscle activity increases with unstable squat movements. Canadian Journal of Applied Physiology, 30(1), 33–45. https://doi.org/10.1139/h05-103

9. Norwood, J. T., Anderson, G. S., Gaetz, M. B., & Twist, P. W. (2007). Electromyographic activity of the trunk stabilizers during stable and unstable bench press. Journal of Strength and Conditioning Research, 21(2), 343–347. https://doi.org/10.1519/R-17435.1

10. Cappato de Araújo, R., Andrade da Silva, H., Pereira dos Passos, M. H., Alves de Oliveira, V. M., & Rodarti Pitangui, A. C. (2021). Use of unstable exercises in periscapular muscle activity: A systematic review and meta-analysis of electromyographic studies. Journal of Bodywork and Movement Therapies, 26, 318–328. https://doi.org/10.1016/j.jbmt.2020.12.010

11. Ostrowski, S. J., Carlson, L. A., & Lawrence, M. A. (2017). Effect of an unstable load on primary and stabilizing muscles during the bench press. Journal of Strength and Conditioning Research, 31(2), 430–434. https://doi.org/10.1519/JSC.0000000000001497

12. Sparkes, R., & Behm, D. G. (2010). Training adaptations associated with an 8-week instability resistance training program with recreationally active individuals. Journal of Strength and Conditioning Research, 24(7), 1931–1941. https://doi.org/10.1519/JSC.0b013e3181df7fe4

13. Saeterbakken, A. H., Andersen, V., Behm, D. G., Krohn-Hansen, E. K., Smaamo, M., & Fimland, M. S. (2016). Resistance-training exercises with different stability requirements: Time course of task specificity. European Journal of Applied Physiology, 116(11–12), 2247–2256. https://doi.org/10.1007/s00421-016-3470-3

14. Saeterbakken, A. H., Olsen, A., Behm, D. G., Bardstu, H. B., & Andersen, V. (2019). The short- and long-term effects of resistance training with different stability requirements. PLoS ONE, 14(4), e0214302. https://doi.org/10.1371/journal.pone.0214302

15. Gao, J., Liu, D., Zhu, J., Guo, Q., & Wang, X. (2025). Instability core training vs traditional core training on trunk strength and sprint performance among athletes: A systematic review and meta-analysis. PeerJ, 13, e20212. https://doi.org/10.7717/peerj.20212