Wednesday, July 20, 2011

Fast & Slow, Heavy & Light, Eccentric & Concentric: Do All These Fancy Training Variables Really Matter? Probably For Power. Not So Much for Size, Though.

Image 1: According to Sakamoto, 2011, EMG
activation during the bench press increases
with rep-speed & weight (pic from mylot.com)
"Go heavy or go home!" You probably have heard this advice time and again and, after all, two recent studies appear to suggest that, when all is set and done, ah... I mean all sets are done, the thing your muscles seem to care most about is workload. While the study by More et al. (Moore. 2011) does not tell us anything about the effectiveness of training with different rep speeds, it goes to show that concentric and eccentric training are similarly effective, when it comes to building sleeve bursting biceps. Sakamoto et al. (Sakamato. 2011), on the other hand, found that EMG activity of the pectoralis major increases with rep speed and (readers of the SuppVersity EMG Series know that already) weight. Now, before we jump to any preliminary conclusions, let's tackle the studies in some more detail...

The nine healthy, but previously not weight-training average Joes (mean age: 22yr; height: 1.75m; weight:78.3kg) from the Moore study (Moor. 2011) performed single arm biceps curls on a dynamometer twice per week. The volume increased from week one to week five from 2 to 6 sets and was cut back again in the last (ninth) week before the final testing session. Other than in similar studies on the effectiveness of eccentric vs. concentric training, the subjects did not perform their dynamometer curls either concentrically or eccentrically, but were instructed to perform concentric curls with one arm and eccentric curls with the other. Right and left arm had previously been randomly assigned to either the maximal lengthening (eccentric) or shortening (concentric) condition, so that limb dominance (n=5 dominant; n=4 non-dominant) was adequately counterbalanced. Moreover, the subjects had to perform ~40% more repetitions on the concentrically trained arm, to ensure total work was equal, or, put differently, to make up for the greater muscle force generation (+60% total work per repetition in eccentric vs. concentric) during eccentric dynamometer curls. Thusly, the participants performed the same 51.8MJ of work with each of their arms in the course of the 9-week training program.

Under these equalizing conditions, workoutput for both conditions rose similarly over the 9-week training program:
Total work per repetition increased from week 1 to week 9 for both LC and SC (main effect for time, P = 0.001) with no difference between conditions (time by condition interaction, P = 0.63). The average increase in work per repetition was similar between LC and SC (17.2 ± 6.3 vs. 22.1 ± 8.9%, respectively; P = 0.69). There were increases (at least P<0.05) in peak torque for all velocities tested (*8–20%) with no signi´Čücant difference between conditions.
The scientists also found similar results for the muscle crossectional area (CSA), which had been "virtually identical (P = 0.99) before training" (48.5mm² vs. 48.4mm², for the ecc. and con. trained arm) and "increased similarly between conditions" (ecc. 6.5 ± 0.6% vs. con. 4.6 ± 0.4%, respectively; interaction, P = 0.37). What may initially sound like one of those statistically induced geeky underestimations of real world effects, i.e. calling 6.5% vs. 4.6% increases in muscle CSA "similar", turns out to be actually negligible if you calculate the respective absolute difference in CSA increase which is less than 1mm², or an area with the size of a pinhead.
Figure 1: Other than total work per repetition, the respective peak torque development did vary significantly (+8.9% vs. +13.5% for con vs. ecc) between the concentrically and the eccentrically trained arm (data adapted from Moor. 2011)
If, however, you plot the peak torque data from table 1 from the Moore study (I did that for you in figure 1), you will realize that, after all, there is more of a difference between the two training regimens than Moore and his colleagues dissertations would make you think. In fact, their assertion that "there was a main effect for condition for peak torque measured at 0.79 rad/s [slow concentric] in that LC [eccentrical training] was *8% greater than SC [concentric training]" is simply not consistent with the data they provide.
Note: A comment by "anoymous" (guys give me at least a pseudonym!) reminded me that in yesterday's hurry I forgot to mention a major caveat to the study. The latter is directly related to the unilateral training protocol which could potentially (or rather certainly) lead to cross-over effects from one arm (probably the eccentrically trained one) to the other. Similar effects have been observed in e.g. Adamson et al., 2008, where rate of force development and maximal isometric contraction (37% vs. 35%) in 10 adult females increased similarly in both arms, although the ladies had trained only one arm. It is yet notable that the 1RM increased almost exclusively in the trained arm and that the strength increases in the Adamson study occurred in the absence of muscular hypertrophy and are thus attributed by the authors to neurological addaptions of which obviously both arms benefited to a similar degree.
As far as peak torques are concerned the available data (with reservations that the data the authors provide in table 1 of their paper is correct) would suggest that the peak torque increments in the concentrically trained arm for different repetition tempos were on average 4.6% greater than those for the eccentrically trained arm, or, in other words, the higher rep lower weight concentric training resulted in greater strength improvements than the higher weight, lower rep eccentric training, which is so contrary to what you see in similar studies that I would assume that the authors just got the captions wrong and the data in figure 1 would have been reversed, i.e. what now is red should be blue and what now is blue should be red... but who cares, anyway? Focus on getting a good contraction on both the con- and the eccentric phase of your curls, do the exercises described in the SuppVersity EMG series and grow ;-)!
Figure 2: The time [in s] to "speed failure" (i.e. not being able to complete another rep at the given tempo / slow: 5.6s; medium: 2.8s; fast: 1.9s) increases with increasing tempo and load expressed in % of 1 repetition max, 1RM (data adapted from Sakamato. 2011)
Fortunately, the Sakamato study does not contain similarly confusing results. In essence the study, which investigated muscle activations under varying speeds and intensities during bench press using surface electromyography (EMG) found that in the 13 weight-trained men (21.7 ± 3.6-year-old) who performed bench press until fatigue under five intensities (40–80% 1RM), and four speeds (slow 5.6-s/repetition, medium 2.8-s/repetition, fast 1.9-s/repetition, and ballistic maximum speed), found that ...
...faster conditions [...] produced a significant fall in amplitude during the final concentric phase compared to slower movements [while at the same time] after fatigue, EMG amplitude increased, with the speed effect being maintained. 
This means that in the rested state at the beginning of the training you still have the explosiveness to really "pump" the weight up and thus pump out more reps. On the othrt hand, maximum muscle stimulation does not occur before your pectoralis major brgins to fatigue later in the exercise session (cf. figure 3).
Figure 3: Normalized EMG activity at five time points for a given rep tempo / slow: 5.6s; medium: 2.8s; fast: 1.9s / in the rested and fatigued state (data adapted from Sakamato. 2011)
According to the EMG data in figure 3, it does make sense to start (after an appropriate warm up) with heavy and explosive sets / movements and to switch to medium weights and tempo later in a training session. But wait, isn't that exactly what generations of successful trainees have been doing already? Well, I guess this is then another case, where practical training experience beat exercise science by decades and thus further evidence that much more than in the case of nutrition & supplements most of the research that is put into specific exercise programs does little more than reproduce pieces of the knowledge that has accumulated in the heads of trainers and trainees all around the globe ever since the earliest days of physical culture.