Motor unit recruitment

[JonA]

After reading a few threads here, I've been trying to read up more on motor unit recruitment, how rep/set schemes and energy systems interact, and where bar speed fits in. Basically trying to find some sort of comprehensive model that puts it all together for me. I came across this article:

https://www.t-nation.com/training/secre ... ecruitment

I'm curious what what other people think of it. Is this a good summary?

Some points that I took from the article:

• Sets should last < 15 seconds total, any longer and your body will have exhausted it's phosphagen energy system and won't be able recruit as many motor units. This naturally puts an upper limit on the number of reps in a set, especially for heavier loads.
• Bar speed reduction in a set is also an indicator of phosphagen energy system exhaustion. Using lower loads with higher bar speed makes this easier to see than high loads, because high loads always move slower.
• Grinding is counter-productive, your forcing your smaller Type I motor units to do the work because your Type II's are taking a break.

[Murelli]

Grinding is counter-productive, your forcing your smaller Type I motor units to do the work because your Type II's are taking a break.

Agreed. Don't know about the whole energy system 15s deal. Sounds sketchy.

Edit: the Tren nation article is full of horse manure. The guy blends correct bits of information with a lot of incorrect assumptions. I remember Hanley writing about how "speed sets" are bullshit because if you go full ballistic with a 60% load the barbell would go flying through the ceiling, but it doesn't now, does it?

[JonA]

I remember Hanley writing about how "speed sets" are bullshit because if you go full ballistic with a 60% load the barbell would go flying through the ceiling, but it doesn't now, does it?

Maybe for a light olympic lifts like the snatch. I think that's where the trend of hex-bar jump squats and high clean pulls without the catch come into play.

For a squat, though. 60% load for me (which is the lightest the article talks about) puts me at ~250lbs. I can make that bounce on my shoulders at the top of a rep, but not much more. Certainly not through the ceiling. I still consider myself to be a pretty high ballistic/Type IIx guy. (Just measured my SVJ at 32" @ 42 years old)

[Hanley]

Well, Goddamn. I can't believe he wrote that in 2007.

That is by far the best explanation of motor unit recruitment I've ever come across.

I've come to a pretty much identical conclusion after trying to reconcile bar-speed data with recruitment models.

[Hanley]

Grinding is counter-productive, your forcing your smaller Type I motor units to do the work because your Type II's are taking a break.

Agreed. Don't know about the whole energy system 15s deal. Sounds sketchy.

Edit: the Tren nation article is full of horse manure. The guy blends correct bits of information with a lot of incorrect assumptions. I remember Hanley writing about how "speed sets" are bullshit because if you go full ballistic with a 60% load the barbell would go flying through the ceiling, but it doesn't now, does it?

I think "speed sets" are bullshit for "strength", but not hypertrophy.

There's pretty broad consensus that ~60-65% and 80% 1rm mark two useful programming thresholds.

60%

I think 60% represents a useful threshold because that's the intensity where I'm recruiting a good proportion of IIx fibers on a single rep performed at max sane* volitional velocity (* absolute full recruitment would indeed be ballistic...but the bar won't go through the roof...it might just sail an inch or two beyond my ROM).

- Given the restriction of segment-length determined ROM, I can't recruit all IIx fibers at 60% (I moderate velocity). BUT, I am indeed recruiting a good proportion. I will cycle through my pool of IIx fibers in the first few reps. This jives with bar speed data. The first 5-6 reps at 60% are usually the same speed, then -- as IIx fibers fatigue -- velocity slows. Basically, I've "touched"/recruited all IIx fibers in those first 5-6 reps. After that, I'm recruiting less IIx on every rep (these are the smaller IIx fibers that were already recruited on reps 5-6, though). As the set goes on, Type I and II- slog through the rep with ever less contribution from the biggest/strongest/fastest-to-fatigue IIx fibers.

Below 60% 1rm, recruitment of IIx fibers will indeed only happen as I incur ridiculous fatigue in I and IIa fibers. This is basically jogging with a barbell. Stupid. Stupid. Stupid. Do not fuck with IIx fibers.

80%

80% represents (roughly) the transition at which I'm recruiting nearly all IIx fibers from the very first rep. I don't have to moderate velocity at all 80%. Bar speed decay indicates this. First rep is fast; bar speed falls from there. 80% represents the transition to "strength" zone. At loads of 80% and up, I'm training the neural aspects of strength.

Edits: I kept writing “type IIb when I meant IIa”

John Crox

[SeanHerbison]

John Crox

Lies. He died last year.

[damufunman]

• Grinding is counter-productive, your forcing your smaller Type I motor units to do the work because your Type II's are taking a break.

Doesn't depend in whether you're grinding due to fatigue vs a 1RM (or maybe even a submaximal but heavy single)? Above is correct if you're doing reps, but for a single heavy rep you're Type II's are coming online last, and the higher intensity is required for the to fire, if my understanding of the size principle is correct.

[PatrickDB]

@Hanley I'm still waiting for your "why you shouldn't do sets greater than 5 reps" article like a kid still waiting for his dad who went out for cigarettes and never came back.

[Hanley]

@Hanley I'm still waiting for your "why you shouldn't do sets greater than 5 reps" article like a kid still waiting for his dad who went out for cigarettes and never came back.

I want to use my own bar-speed data to support my conjecture. My position is based entirely on bar speed decay patterns at various relative intensities. And using those patterns as [hand wave**] proxies for MU recruitment patterns. So, I really need actual data. I was going to use Manveer's bar speed data (he gave me a fuckton of it), but I'm never clear on what intensities he's using (he's using RPEs not %s).

** it is hand-wavy, but the henneman-size principle of recruitment is about as accepted as accepted science gets in ex phys, I think.

###

Also, I'm lazy. Send those amphetimines. I'm a motherfucking monster of productivity on that shit.

[Manveer]

I was going to use Manveer's bar speed data (he gave me a fuckton of it), but I'm never clear on what intensities he's using (RPE).

Huh? There's a column for RPE.

[Hanley]

I was going to use Manveer's bar speed data (he gave me a fuckton of it), but I'm never clear on what intensities he's using (RPE).

Huh? There's a column for RPE.

I know. But shit gets murky around @6-7. Right? @ 6 could be anything from 60% to low 70%, yeah? It would be easier to interpret if I knew your percentage-intensities.

###

Sorry, inside the parentheses shouldn't read "RPE"; it should be "he notes RPEs, but percentages would make it easier to interpret the data".

Your data is what got me started on this MU recruitment path in the first place. Everyone yaps on about "linear regression". That's horsepooh. I've never seen this in a) bar speed decay patterns, b) plots of velocity at various intensities The curvy quality of curvlinear regression is really fucking interesting. And -- perhaps -- very useful.

[PatrickDB]

@Hanley Oh come on, the RPE excuse for not using Manveer's data is lame. Just take his RPE at face value, convert all the unfatigued top singles to e1RMs, and average. Compare to his most recent PL comp to make sure you're in the right ballpark.

There's going to be measurement error your way too (daily fluctuations in performance and fatigue) so you've got to deal with error somehow. I'm not exactly sure what your conjecture is, but if it's robust enough to support designing an entire program around, it should be robust enough that you should be able to verify it even with a little noise.

I would be very skeptical of using polynomial regression for this shit. Don't worry about being fancy. Just plot the data with seaborn and run the best statistical test: the eye test. If there's an effect you'll see it.

I hate whoever made this thread, because now I'll be spending today reading Waterbury's Huge in a Hurry instead of working on research. I'm watching myself getting distracted like a cat finding a shiny object and yet I can't stop myself.

[Hanley]

@Hanley Oh come on, the RPE excuse for not using Manveer's data is lame. Just take his RPE at face value, convert all the unfatigued top singles to e1RMs, and average. Compare to his most recent PL comp to make sure you're in the right ballpark.

There's going to be measurement error your way too (daily fluctuations in performance and fatigue) so you've got to deal with error somehow. I'm not exactly sure what your conjecture is, but if it's robust enough to support designing an entire program around, it should be robust enough that you should be able to verify it even with a little noise.

At the very least, I want to test a second organism (me).

[PatrickDB]

At the very least, I want to test a second organism (me).

So do the data viz for Manveer now and you'll have the code ready for when your nerd unit arrives.

My guess is that either you won't be able to see the effect or it will be very obvious. No need to hang-wring about "is this a RPE 6 or really a RPE 6.5?" and similar questions.

Maybe an interesting first thing to do is go through each set normalize the velocities to percentage of the first rep velocity (maybe first sorting by number of reps). What we're hoping for is something like a bunch of reps at about the speed of the first rep, then some phase transition, then decaying speed until failure, right?

[Hanley]

I'm not exactly sure what your conjecture is,

It's pretty similar to what that Waterbury dude said in his T-Nation piece: I can best accumulate tonnage using IIx fibers by moderating interset fatigue.

In terms of training IIx fibers within a session: 40 high-velocity reps at 70% is a very different thing than 20 high velocity reps and 20 "slow" reps @ 70%.

[PatrickDB]

I'm not exactly sure what your conjecture is,

It's pretty similar to what that Waterbury dude said in his T-Nation piece: I can best accumulate tonnage using IIx fibers by moderating interset fatigue.

How does bar speed data help answer this question? Just do 40 high velocity singles at 70%, done.

Or are you trying to find how large the sets can be while still keeping the velocity up?

[Murelli]

I'm not exactly sure what your conjecture is,

It's pretty similar to what that Waterbury dude said in his T-Nation piece: I can best accumulate tonnage using IIx fibers by moderating interset fatigue.

In terms of training IIx fibers within a session: 40 high-velocity reps at 70% is a very different thing than 20 high velocity reps and 20 "slow" reps @ 70%.

What about the whole biochemical pathway to growth (more damage = more metabolytes = more growth)?

[PatrickDB]

What about the whole biochemical pathway to growth (more damage = more metabolytes = more growth)?

It appears not to matter compared to total tonnage.

Look on sci-hub.hk for "Effects of different volume-equated resistance training loading strategies on muscular adaptions in well-trained men," by Schoenfeld et al. and "Greater Gains in Strength and Power with Intraset Rest Intervals in Hypertrophy Training."

[Hanley]

Or are you trying to find how large the sets can be while still keeping the velocity up?

Exactly this.

High velocity singles don't work great. I (and others) have tried. On a given single at 70%, I'm not recruiting all IIx MUs (or I'd throw the bar an inch or two). I need a few reps (about 4-5 at 70%) to cycle through my pool of IIx MUs. After those reps there is a very obvious decay in bar speed (drop in IIx recruitment per rep).

I'm guessing if I do a single, rest, do a single...I'm not (over the entire session) recruiting all IIx MUs [in fact I'm probably doing minimal work with my biggest MUs...not good].

[PatrickDB]

Ok @Hanley, so what's wrong with:

1) Pick the lift for which you have the most data,
2) Eyeball Manveer's e1RM from his top singles using percentage/RPE chart (don't even need to code here),
3) Find all the sets that use 70-75% of his e1RM,
4) Normalize bar velocities relative to the speed of the first rep,
5) Scatterplot them with reps on the x-axis and relative speed on the y-axis,
6) Use your big ol' brain to figure out where the right cutoff is?

We can do a formal statistical test to buttress your conclusion from 6, if you want, but who cares.

(Speaking of phase transitions and amphetamines: http://aleph.se/andart2/math/did-amphet ... elp-erdos/)