This new mind-muscle study changed my mind

Chapters:

00:00 Intro

03:12 The results

06:39 My online PT Certification course

07:03 Observations from practice

08:01 EMG Research

13:39 Conclusion

15:22 Outro

Transcript:

Most people in the gym heavily base their exercise selection based on how an exercise feels. You try an exercise, see if you feel it in the target musculature, and if so, you keep it in the program. If you don’t feel it, you might switch it out. If something gives a good pump or you really feel like a nice contraction that’s probably an exercise you’re going to keep doing. This mind muscle connection, the idea that we can feel which muscles are activating and which muscles are experiencing high tension and will therefore grow more is seen as almost axiomatically true, undeniable, unquestionable. But I’m a big proponent of facts over feelings. So can we actually feel which muscles are active? We did a study to find out.

We had a group of strength trained athletes perform a pull day workout consisting of the following exercises: pulldowns, barbell and dumbbell rows, pullovers, reverse flys and biceps curls. We standardized the exercise intensity and we ask all of the athletes how well they felt the exercise in certain muscle groups. And then we also measured the objective level of muscle activity using electromyography, EMG. EMG is a technique that measures the electrical impulse to a muscle and therefore provides an objective level of muscle activity, the amount of electrical muscle activity that is taking place there. As we’ll go into later in more detail, EMG research has come under scrutiny for not directly reflecting muscle hypertrophy.

While it doesn’t directly reflect muscle hypertrophy, it is a lot more validated as a measure of muscle activity. By correlating the subjective and the objective level of muscle activity we could see if people can subjectively or intuitively sense how well a muscle is activating during a certain exercise. So how accurate do you think the athletes were at ascertaining how much a muscle was activating during the exercises? I expected moderate correlations. Funny enough, while tons of people base their exercise selection and their program in general heavily on how their exercises feel and which muscle groups they feel being active, this is actually the first study that has properly investigated if these feelings correspond to an objective reality.

The reason we investigated this is because I was skeptical of people’s ability to feel this because, as a coach, I know that many people, when they try an exercise for the first time in particular, they always ask which muscle group does this train? Especially when the exercise is more uncommon, like a hanging leg curl, a lot of people will intuitively think it trains the abs or something, a completely unrelated muscle group. (it’s for the hamstrings) And then later people tend to get better at, at least knowing which muscle groups are active, but is it the case that they actually feel which muscle groups are active or is it the case that we just know that a row trains the lats and therefore we start focusing on the lats and then maybe we feel the lats more, or we just kind of tell ourselves that we feel the lats more?

I think when people are actually blinded and made to rely purely on what they actually introspectively feel they are not as good as they think they are and most of the things that we think we feel are more related to what we know, which muscle groups should be active based on biomechanics and what we know from science rather than actual introspective feelings. But the results were actually much worse even than I thought. There were no significant correlations at all between subjective and objective muscle activity. The correlations that were there were small, inconsistent, and not even always positive. Statistically all of these correlations were insignificant. We could see in particular that athletes significantly underestimated the differences between the levels of muscle activity of different muscle groups. So they underestimated how much difference there was between the primary and the secondary targets of an exercise and there was also very wide variance in how much muscle activity people felt during an exercise. And that muscle activity did not reflect in objective levels of muscle activity as measured by EMG.

On average, on the group level people were not too bad. Here you can see the levels of muscle activity as rated by the athletes themselves. So this is the intuitive, the subjective level, the perceived muscle activation levels for all the exercises and the targeted muscle groups and here you can see the objective level of muscle activity as measured by electromyography during the exercises. On average, at the group level these charts roughly correspond. There are a lot of significant differences. For example, if you look at perceived activation, for example, during barbell rows, the athletes did not feel their rear delts. They did not feel them as much as their lats or their traps and they also didn’t feel them as much as during dumbbell rows. And that’s not reflected in the EMG data at all and also doesn’t make any biomechanical sense. And if you add all of these inconsistencies together, especially at the individual level, the correlations again were statistically insignificant. So if you ask a whole lot of people at a group and you average out the responses, you probably get some sense of which muscle groups and exercise trains, but at the individual level the data are essentially useless.

Look at, for example here the correlation chart. This is the correlations between subjective and objective muscle activity for all the exercises and all the muscle groups that we measured. If you look at the lat pulldowns, if we zoom in on the lat pulldown here, you can see that, for example for the last there was a positive correlation. Statistically insignificant, but generally there was a rough trend for people to experience actually higher levels of muscle activity when they also reported higher levels of muscle activity. However, for the posterior deltoids and for the traps there was no correlation at all, the line was flat. So there was wide variance in how much muscle activity the athletes reported, but this was not reflected in objective levels of muscle activity. So for example, here, especially for the posterior deltoids, during pull downs, you could see that the range spans from about three out of ten to almost ten out of ten. So about 3 to 9 out of ten. So a very wide range of how well people felt their posterior deltoids during pull downs. But this range did not correspond with any differences in actual electromyography results.

So people experienced a wide difference in how much they felt their rear deltoids during the pull downs, but there was actually not that much difference in objective muscle activity for these athletes. And then for some exercises the trends were even negative. So overall the conclusion was very clearly that at an individual level asking someone how much they feel a certain muscle group being activated during an exercise does not correspond well with objectively how much muscle activity there is in that muscle group. You can see in the correlation chart that while some of the lines are somewhat positive, many of the lines actually run downward, they are negative and many of the lines are almost flat, meaning there’s just no correlation between subjective and objective muscle activity levels.

The major variance in the levels of reported muscle activity, so how well people felt an exercise in a certain muscle group corresponds with what I at least observed in practice. Many people don’t feel certain muscle groups at all. Many people have trouble with their lats, for example, the delts, the glutes. To the extent even that some people think they have gluteal amnesia because they don’t feel their glutes when they’re squatting, for example. The glutes are certainly activating and you can see this with EMG research and when you measure muscle growth, they will grow.

Myself, for example, I never feel my glutes during, well, almost any exercise, but especially not an exercise like squats and they’ve certainly grown. My delts are probably one of my best body parts. Actually glutes and delts both and neither of them I feel much. My delts never get sore, they don’t get much of a pump and I almost never feel my delts during an exercise. A lot of people have this, certain muscle groups that they just don’t feel very well, and it doesn’t seem to correspond to any differences in how well these muscles are activating or growing from their training. It is worth noting that we used 4 to 6 RM loads. This was to standardize the training intensity and make sure people were lifting heavily enough. Most EMG research uses very sub maximal loads which means the results are basically useless because if you’re doing something like three repetitions with 50% of your 1RM as many studies use as a silly protocol, it doesn’t necessarily correspond to levels of muscle activity seen with maximal strength training.

Now, you might argue that in general people are better at sensing the level of muscle activity with lower loads like 6 to 12 rep max, maybe 8 to 12 rep max. People have an easier time feeling which muscles are active. But even if this is true it only further highlights that we cannot sense muscle tension because heavier loads necessarily require higher levels of muscle tension. Our muscles activate more, they are recruited more when we are using heavier weights because they have to produce more force. Therefore, even if it is the case that our feelings correspond more with muscle activity levels at lower loads, this still means that it is not actual muscle activity or mechanical tension that we are feeling, rather, we are probably feeling something like metabolic stress, the pump, the burn and maybe we extrapolate that to muscle growth or muscle activity, but that is also not necessarily always true. There are a lot of other factors that influence metabolic stress such as whether there is constant tension, the resistance curve of the exercise, that don’t necessarily correspond with differences in actual mechanical tension in the muscle.

It could also be that our subjective sense of muscle activity is actually correct and it is EMG that is wrong. We have a section on this in detail in the paper and I think many people on social media may be partial to this argument, especially if they got their fitness education on social media because EMG has come under fire in recent years as being kind of debunked. And the reality is that EMG is very far from debunked. EMG is very noisy. EMG can not directly predict muscle growth and there are a lot of limitations with many EMG studies. Many EMG studies are very poorly conducted. As I said before, they use very submaximal training intensities and many studies don’t even equate exercise intensity. So they compare, for example, pushups with a certain grip with pushups versus another grip and they don’t take into account that that also changes how difficult the exercise is. EMG is actually very sensitive to how difficult an exercise is. And that’s a good thing about EMG.

EMG has been validated against a number of measures of muscle activity. For example, if you train harder and you report also training harder, we see higher levels of muscle activity as per EMG. EMG levels are also higher when you use higher training intensities. We also see that when you get closer to failure throughout the set, the level of muscle activity increases. So as you get closer to failure all of these measures of effort increase, we see that EMG readings also increase. EMG has also been validated against ultrasound and MRI in certain studies. So we see, for example, that when EMG research predicts high levels of muscle activity we also see more muscle swelling in those muscle groups afterwards and these measures correlate. And we also see that measures of MRI muscle activity correspond with levels of EMG muscle activity. So EMG is certainly not debunked. It is very noisy and it has many limitations, but it would be very unscientific to simply dismiss a properly conducted EMG study.

We also have many practical examples where EMG research turned out to be better than our intuitive sense of which muscles are active and which aren’t. For example, for squats. EMG research has long shown that the rectus femoris, the middle head of the quads, is not activated well during squats, but it is activated well during leg extensions. Indeed, we now know that leg extensions grow the rectus femoris better than squats do. We also see in EMG research that the hamstrings are not activated well during squats. And indeed, we now know that squats don’t grow the hamstrings very well. Many, many people for a long time insisted that squats still train the hamstrings, even though biomechanically this doesn’t make sense. There was a huge argument in the Starting Strength community with Rippetoe defending that squats are a complete lower body exercise and they do train the hamstrings, whereas others argued that this is not the case. And EMG research won out in the end.

As I mentioned earlier, when it comes to gluteal amnesia and the glutes not contracting during squats, EMG research very clearly shows that the glutes are very highly activated during squats and indeed they grow a lot from squats. And we also see that in research, when we look at gluteal activation drills, this does not increase muscle activity levels and it also doesn’t increase people’s gains. So people might feel their glutes more, but it’s not objectively doing anything in EMG research and it’s not objectively increasing people’s gains. So again, EMG research has proven useful in that context. Another example is calf raises.

I see many times in clients when they have a standing or a straight leg calf raise in their program and a machine is broken, or for some reason they cannot do the exercise they will do a seated calf raise with their knees bent and tell me that’s what they did instead. And many people intuitively think that what does it matter – when your seat is where you’re standing? It’s a calf raise, it’s the same exercise. It’s not. A seated calf raise only trains the soleus, the long part underneath. A standing calf raise also trains the gastrocnemius, because the gastrocnemius is a biarticulate muscle and when you flex the knee the gastrocnemius is shortened too much to contribute and we see this in EMG research. EMG research quite clearly shows that gastrocnemius is not activated very well during a seated calf raise and we also know now that they don’t grow much from a seated calf raise. So people don’t seem to feel the difference between those seemingly similar exercises very well, whereas EMG research clearly differentiates them in a way that corresponds to objective muscle growth and corresponds to biomechanics.

So overall, I think these results make sense. They fit with practice and they make biomechanical sense. It’s unlikely that EMG research is just completely wrong and it’s much more likely that our intuitive sense of how much a muscle group is activating is just not based on any objective reality. People have always just assumed that what they feel must be true and it’s very tempting to do this psychologically, but just because we feel something in a muscle group doesn’t actually mean that that is mechanical tension or muscle activity. That said, I am of course biased because it’s my own study and I look forward to seeing replication of these results via other research groups and seeing maybe how these results translate for other muscle groups and other sample sizes. However, overall, for now, what I teach my students in my online PT course is that they should base their exercise selection principally on biomechanics and scientific data. That includes MRI data, ultrasound data, and yes, EMG data as well.

We can look at which muscle groups grow more from certain exercises. All of these measurements have at least some data to support that they are a measure of muscle activity and muscular development. There are no data supporting that our subjective sense of muscle activity corresponds to the actual training stimulus for muscle growth. So feelings should be strictly secondary in consideration, in my view, compared to all of these data for which we have reliable evidence that they actually do mean something. The best data are, of course, long term muscle growth data, where we simply see this muscle group grows a lot more than this muscle group from this exercise. Therefore, this exercise trains this muscle group the most in this muscle group less. Until we have more of those data, though, I would rely on objective proxies for that level of muscular development and take feelings with a big grain of salt. If you’re also a proponent of facts over feelings and you like this type of evidence based content, I’d be honored if you like and subscribe.