After reading a recent study on some potent effects of emphasizing bar speed during the bench press (1), I thought I’d take a look at two classic studies looking at the importance of rapid contractions during training. The first of these studies (2) is definitely a classic that is often missed, but it’s such an interesting yet simple study design with some profound implications for the strength and conditioning world that I had to include it. At the end, I’ll tie them into my philosophy of bar speed during training, how this reconciles with what some of the strongest people around are doing, and after all that, I’d love to hear what you think and do in your own training in the comments section.
Ballistic movements occur at high velocities and involve extremely rapid contractions that follow a tri-phasic contraction pattern (agonist/antagonist/agonist) and are thought to be preprogrammed. Previous work observed that this ballistic motor program occurs whether the limb is free to move or not (3). If ballistic contractions are truly preprogrammed and proceed whether or not the joint actually moves through a range of motion, then potential training effects could occur with these contractions even if movement is restricted (isometric).
To test this premise the authors recruited 8 young male and female 20-year old students to complete a 16 week training program for the dorsiflexors. Each leg was randomized to receive either isometric (no movement) or isokinetic concentric (fixed velocity; movement condition) contractions performed with a rapid (maximal) rate of contraction (ballistic contraction) regardless of training mode.
For the group that trained with movement, this work confirmed the velocity specificity of strength, with the greatest improvements occurring closest to the velocity at which the leg was trained (5.23 rad per second). But when the isometric only group was considered, a similar pattern of strength changes occurred with training, suggesting that movement through the range-of-motion wasn’t necessary as long as the intent to contract as quickly as possible was used. When both groups were tested isometrically, no difference between the training modes was identified, however an overall training effect found increased peak dorsiflexor torque (9.6%), maximal rate of torque development (25.7%), and maximal rate of torque relaxation (return to relaxed from contracted state; 46.5%).
Finally the authors tested electrically-evoked contractile properties, where stimulation of the peripheral nerve causes muscle contraction, independent of activity of the central nervous system. By stimulating the nerve closest to the muscle, the researchers took the CNS out of the equation to take specific measurements of the properties of the muscle itself (although it does include neuromuscular junction function). In this case there were no differences between the training modes for all measures and when collapsed for an overall training effect, ballistic contractions reduced time to peak torque (6.2%) and half relaxation time (11.9%) when evaluating twitch contractile properties (single electrical stimulation and resulting contraction). Under tetanic stimulation (high frequency stimulation so no relaxation occurs between contractions), maximal rate of torque development increased (14.4%) while peak torque was unaffected.
All in all, these results suggest it is the intent to move quickly, using a rapid rate of contraction that is most important, even if the actual movement velocity is slow, or in this case zero (isometric). This is interesting and has implications for more dynamic conditions, such as lifting heavier weights, when lifting velocity inevitably decreases as load increases. While popular literature suggests picking a training load at the optimal power output to maximize power adaptations (that magical 30%-1RM), is it possible that heavier loads could still be used, as long as the lifter attempts to move the weight as quickly as possible, even if bar speed still remains slow?
The study from Behm and Sale (2) laid the foundation for the idea that contraction speed is an important variable and that training adaptations can occur independent of actual movement. The results of the previous study would have been strengthened if a slow or progressive contraction group was included as well. This is where our next paper comes in by Maffiuletti et al (4) who directly compared slow progressive isometric contractions against rapid, ballistic contractions similar to Behm and Sale (2).
Twenty one male participants were divided into three experimental groups, one with quick, ballistic contractions, another that performed slow progressive isometric contractions, and a control group who didn’t train at all. Similar to the Behm and Sale (2) paper above, the ballistic contractions were to be performed as quickly as possible, whereas the progressive contractions were over four seconds (increased by 25% of maximal voluntary contraction (MVC) every second). Participants trained three days per week over seven weeks, performing six sets of six isometric contractions with three minutes rest between each set.
Unexpectedly there weren’t as many differences as we’d anticipate after seeing the potent effect of rapid rates of contractions in the previous paper. Both contraction conditions increased maximal voluntary knee extensor torque (15.7% for progressive, 27.4% for ballistic), with no difference between modes. It’s surprising the difference here did come out statistically significant, given that the strength gain following the ballistic contractions was nearly double the progressive group but suggests it was probably a variable response (possibly underpowered). Just as we saw previously, isometric training resulted in strength gains in dynamic contractions as strength increased for both eccentric (15.6% progressive, 18.3% ballistic) and concentric slow and fast isokinetic contractions (velocities of 60 & 240 degrees per second respectively).
The authors also measured EMG activity of vastus laterals and medialis, however the only differences occurred in the progressive contraction condition, and not in a manner that leads to any substantial conclusions. Evoked contractile properties (electrical stimulation of the peripheral nerve) were tested as well, and found increased peak torque and decreased relaxation times, indicating that the ballistic contractions altered intrinsic properties of the muscle (or neuromuscular junction but independent of the CNS) that would be advantageous to rapid contractions.
This was definitely not the open and shut case I would have expected based on the isometric results we’ve seen previously (2), however the data does suggest that ballistic contractions do result in adaptations that would be favourable if you’re concerned with producing torque as quickly as possible. This paper also adds support to the idea that isometric training can improve strength under dynamic conditions (to a degree), although there is much more work related to this and it’s a complicated area of research. This doesn’t mean we should all go out and start training isometrically, but rather suggests that it may not be totally inappropriate to extrapolate results from these isometric conditions to dynamic conditions like high load, low velocity training.
Up until this point, these studies have relied on isometric (and isokinetic) contractions, which while interesting, can’t replace a study that uses dynamic movements with a barbell, especially since that’s the situation we’re usually applying the result to. Fortunately a recent study by Padulo et al (1) has looked at the issue of higher %RM training (85%-1RM) at different movement speeds to see if emphasizing fast bar speeds is beneficial in the gym. Traditionally, power-training revolves around an optimal mix of load and velocity to produce peak power, and this often occurs at lower %RM than we’d normally select for training (often between 30-50%RM, although higher numbers do exist in the literature). But what if we combine higher loads with the intent to contract the muscle as quickly as possible, can we have our cake and eat it too (increased strength and speed)?
Twenty male participants were divided into two groups who trained the bench press two times per week over three weeks, one group with maximal bar speed and the other at a self-selected speed to failure. Each group stopped training when they couldn’t complete another rep (self-selected), or couldn’t complete a repetition at 80% of maximum velocity (speed group) and both had two minutes rest between sets during training.
This was an interesting design, but I can’t help but be concerned about the disparity in the total volume of exercise performed between the two conditions. At the start of training the full speed group was performing roughly 14 reps against the self-selected speed group’s 48 reps, and the difference didn’t close up much with training (27 vs 72 reps). I’ve also seen some criticisms online of the strength levels of the participants despite their years of training experience (approximately 20 years). This may be a valid concern, but shouldn’t interfere heavily with the actual results of the experiment; it’s more of a concern on applicability of the results within trained populations. It would have been nice to see more statistics on what defined training, although from what is included in the paper I suspect it was just self-indentified years training. This is a simple and quick measure, but using an activity inventory would have provided more info on the group’s specific training experience.
At the end of training, only the maximum speed group increased strength (maximal load) and velocity in the bench press (10.2%, 2.22% respectively) while the self-selected group maintained their pre-training levels. On top of the elevated strength increases, EMG activity of all muscles tested (Pec Major, Biceps, Triceps, Trapezius, Deltoid) was higher in the maximum speed contractions when collapsed across all phases of the exercise (Eccentric, Pause, Concentric phase) as compared with the self-selected group.
This degree of strength increase in trained lifters in such a short period of time is definitely amazing; I would certainly kill for a 5% increase in my bench press in three weeks, let alone a 10%. Hell I’d even settle for 1-2%, I’m not picky! Couple this dramatic difference with the large discrepancy volume of exercise between the two training conditions and it’s definitely an enticing argument for favouring rapid lifting.
The concerns about strength levels and training status do weigh on my mind. If the lifters had no previous experience with rapid contractions, and were recreationally trained, perhaps the increased focus and effort during the rep could profoundly increase strength. I think it’s definitely in the realm of possibilities. But when you mention the fact that your participants are trained, given how sensitive the fitness blogosphere is about this, I’d think it’d be best to include some specifics as to their previous training.
Despite some concerns about the experimental design and the actual training status of the subjects, combined with the previous two studies this study supports that the intent to move the bar quickly, even at higher %RM loads is beneficial to strength development and at the very least isn’t detrimental to performance over a three week period.
It’s clear that between these studies emphasizing rapid contractions in training may not always allow us to increase strength per se above slower contraction speeds (paper #2). Although the paper above would disagree (paper #3), it does suggest that lifting heavier weights may still produce changes reminiscent of ‘power training’ (influence rate of force development) despite slower movement velocities as long as the intent to move as quickly as possible is used (ballistic contractions). While the final paper does demonstrate benefits of speed during training, I am left wondering about the preprogrammed nature of ballistic contractions, and even though we saw they were beneficial under isometric contractions, I wonder how the ‘motor program’ would be affected when movement is occurring albeit at a slow rate due to heavier weights.
But the benefits of emphasizing the speed of contraction may be something we’ve known all along. Based on recent surveys, it looks like the powerlifting world has been working to maximize bar speed for awhile now (5). A recent survey of training methods of elite British powerlifters suggests that even with high loads (80-100%), most (60-65% of respondents) made an effort to maximize bar speed in the big three lifts (bench, squat, deadlift). When performing lighter, repetition-based work (0-70% 1RM), bar speed was even more of a priority than with the high load (68-75% of respondents across the three lifts). Even when doing targeted explosive work, the majority of the respondents in the survey opted for loads higher than what the literature would suggest (6-9) is optimal for power output as only 3.6% of respondents reported using 31-40%-1RM while the majority preferred somewhere between 41-70% (biased to the high side). These trends have carried over to strongman training as well, with 50% of respondents performing traditional resistance training exercises at maximal speed (10).
This is by no means a conclusive case for emphasizing bar-speed, but these papers combined with practical experience has been enough to convince me to attempt to maximize contractile speed during the concentric phase regardless of the load on the bar (and actual movement velocity). But it’s also a biased account, as I didn’t include all the literature on the %RM relationship with velocity and power output and the resultant adaptations, you can only cram so much into a blog post.
Either way, the recommendation of ‘slow and controlled’ lifting may hold for beginners and those new to the lifts, but experienced lifters should just focus on being controlled and remove ‘slow’ from the equation at least during the concentric. What’s more, this extends beyond recommendations for power training where speed is usually emphasized, but into my programming for strength and hypertrophy as well.
Who would have thought that ‘grip it and rip it‘ could be an evidence-based recommendation?