Dear Julie: In defense of the crunch

Dear Julie, You wrote an interesting letter on your beautiful website ( to the coach of your girl's gymnastic program.  You wished to

share with you and your staff how a decade or so of research is transforming our understanding of how we create a strong sturdy center that anchors all of our movements

As a father of two young girls (5 and 3) I really appreciated your views on healthy sports participation, concerns about body issues and the importance of fun in physical activity.  Like you I am also a physiotherapist with a special interest in spine function. I am also a chiropractor, was a spine biomechanics researcher, I completed a MSc in Spine Biomechanics with one of the authors of the references you cited (Stu McGill), I have published a few papers on trunk muscle function (here, here and here) during a variety tasks and was initially very interested in doing research on the lowly and often derided abdominal crunch (here and here).   I love talking about spine stability and how much of this actually old research (I don't think it's emerging, most has been around since the 90s) is applied to clinic or sport in ways that the research does not actually support.  I am also a former recreational gymastics coach and regularly "threw back tucks" after two beers at parties well into my twenties.

The big issue you had with your coach was his use of the crunch in his conditioning program. This was the area where our views seemed to differ. I think at best the jury is still out on much of what you have said and at worst the scientific data does not actually support your beliefs.

NO! Don't make me do crunches!
NO! Don't make me do crunches!

NO! Don't make me do crunches!

I personally don't think any exercise is bad.  It may be inappropriate at different times in a training cycle or there may be other exercises that are better for the specific goals of an athlete. But I rarely throw any movement under the bus.  The data just isn't there for that. In this letter I was hoping to give my rationale to defend the crunch and to discuss some of arguments you and many others (I think I'm in the minority) had against it. In the next few paragraphs I will cite the most common arguments against the crunch and give a rebuttal.

Argument #1:  "Overused abdominals result in bums tucking under and the lower spine flattening"

I noticed that there was actually no reference for this statement.  I would suggest that is because the research that has investigated the idea that strength training a muscle causes it to plastically deform and become shorter at its resting length and then pull bones into different position is actually lacking. I've written on this previously and for a review you can look here showing how strength exercises do not significantly change posture.  While this is a common idea it just doesn't happen.  Anecdotally, you will probably see the opposite postures in gymnasts. The tend to anteriorly pelvic tilt despite all those crunches.

gymnast apt
gymnast apt

A dominant motion in gymnastics is extension hence the concern by some for gymnasts to develop Spondys. Further, by what mechanism would even 15 minutes of crunches daily overcome all of the other neutral spine or extended postures that a gymnast undergoes?  While I recognize muscle "shortening" as a common belief, it is certainly not supported in the "10 years of emerging research".  Rather, it dates back to outmoded views of function perpetuated in the 1950s by Kendal and Kendal that somehow persist to this day.

Argument #2: Flexing the spine compromises spinal function

Gymnasts flex their spines.  The do this under load and they do it repeatedly.  Take a look at the gymnast doing the front tuck and the one doing the release on the parallel bars. Kind of looks like a crunch.

OMG, I'm flexing my spine.
OMG, I'm flexing my spine.
crunch with a vsit
crunch with a vsit

Most athletic sports don't have rigid spines all the time.  Gymnasts certainly train to have a "tight body" but they also move.  The spine generates and prevents movement.  In gymnastics we perform front tucks, back tucks, back handsprings, back fulls etc. There is an approximation between the ribcage and the pelvis in these movements, under high loading.  It happens.  Thus we should train it. I love planks (I've had to defend planking in the past) but they don't mimic all of the demands of gymnastics.  That is the purpose of conditioning - an overload stimulus to prepare for the demands of the sport.  And remember, a curl up is not just the rectus abominis.  Both obliques can be activated more than 50% MVC and the transverse abdominis also turns on substantially more than any indrawing exercise.

You mentioned that the ideal spine stabilty occurs with the co-ordinated action of all muscles.  This is true but in order to progressively overload a muscle group to ask it to adapt we have to train in a somewhat isolated manner.  To achieve strength adapations we probably need the MVC values to exceed 60-70%.  Curl up variations can do this. Imagine if we trained all of the of the muscles of the trunk together as a team exceeding 60-70% MVC.  The compressive penalty would be massive as would the increase in IAP.  The spine would be in neutral but you would still be loading it incredibly - far greater than what would occur in the simple crunch. I would even guess that our neuromuscular system would not even allow such high levels of loading to achieve a strength response (for example see here).

Bottom line: gymnasts flex their spines thus we should prepare them to flex their spines under load

Argument #3: There are better exercises than the crunch

Yup, that is a good argument.  You suggest planks (just as the coach did) and go on to suggest cartwheels and handstands as being even better than the plank.  I think arguing that there are better exercises for gymnastics based on specificity or effort level is a good argument but it does not mean crunches are evil.  Gymnasts could do V-snaps, leg lowers, rockers or other things that mimic their demands.  However, the lowly crunch is a good progression.  As for the cartwheels and handstands, aren't the girls already doing this during the session?  The point of a conditioning component is to overload and stress the system to create adaptations.  We want to choose exercises that are often a greater challenge than the goal task (cartwheels and handstands).

Argument #4: Crunches are dangerous because of spinal compression and disc strain

Again, a gymnast will flex their spine during the sport.  There is probably more compression and shear during their gymnastic activities than the crunch.  The crunch prepares the body to tolerate this load.  Crunches are rather innocous relative to other exercises and tasks.  What is  3-4000 newtons of compression and some shear?  Thats nothing.  If they do an L-sit on the parallel bars that demand will exceed the crunch. Or a kip-up on the uneven bars.

If we use compression and shear loading as the arbiter of safe activity we should get our kids out of gymnastics.  Most of those activities will exceed NIOSH's limits.   You quote Stu's book but we should go back to his original papers in the 1990's and we can see how low the compressive and shear penalties a crunch has in respect to all the other tasks we routinely ask athletes to perform.  I don't see spine flexion as all that evil.  Bret Contreras wrote a nice review questioning whether spine flexion is all that bad.  At the very least, it is a grey area.  Hence, I tend to have an issue with absolutist advice knocking one exercise.

Argument #5: Crunches create abnormal intra-abdominal pressure

Most spine exercises that will train capacity or stability of the trunk under high loads will create an increase in pressure.  This increased IAP creates stability.  Again, if we train with smart progressions and protocols the trunk and the whole musculoskeletal system will adapt as it should.  You wrote:

For the diaphragm this results in a change in breathing patterns, including breath holding, to meet stability challenges, and reduced respiratory capacity.These scenarios will create an issue for sustained respiratory support for endurance in athletics. The pelvic floor may not be able to match the excessive pressure from above, which can lead to incontinence

I'm not sure what the issue is here.  If you are concerned with breath holding, have the kids breathe during a crunch. Breath holding is normal and certainly even beneficial in my opinion if you want to increase stability especially during a landing or a back tuck. And while respiratory capacity might decrease while you are doing a crunch it does not carry over into prolonged dysfunction. How would it?  I don't even know what "sustained respiratory support for endurance in athletics" means or how it can possibly be compromised by a crunch.

As for the pelvic floor, maybe the crunch should be the least of our worries. I would hazard that the IAP will increase to a greater extent in all of the other more challenging activities that occur in the performance of gymastics. Is it the flexion coupled with a little increase of IAP that is so bad? No papers you cited and I doubt any research shows this. But again, I think it is a far reaching conclusion to suggest that crunches lead to urinary incontinence.  The studies you cited certainly don't support this.  The Sapsford study you cited tells me that the pelvic floor muscles have greater muscle activity during upright sitting. Super.  So they have less activity when your spine flexed.  Why is this a concern? The Biceps muscle functions better at midrange yet we still train the muscle through its full range.  Same with every other movement in the body.  Why is just the spine that we avoid ranges of movement that have less than optimal force production?

Random Argument #6: The crunch isn't functional.

I think this depends on your goal task and your beliefs behind exercise prescription.  I think in some sports we flex our spines and that flexion creates movement and is necessary.  Take a look at my golf mechanics e-book.  You will see in the kinematics graphs that trunk flexion is one of the first movements to start the swing.  Think of paddlers that flex their trunk under huge loads and force production.


We are not always in neutral during sport and activities of daily living.  Should we not then train out of neutral?  Maybe the crunch isn't the best way, maybe you can use cables, or throw medicine balls at the ground but you are still training flexion.

And now for some random thoughts...

The creation of Nocebo with all this catastrophizing

I get worried when we use a biomechanical rationale to warn against the normal movement of simply bending our spines under load.  If you tell the coach the danger of spine flexing and he passes this onto the kids what are you putting in their head?  Don't flex the spine?  Again, the do this all the time.  They learn to think their spine's are delicate rather than this robust, beautiful structure with an immense ability to adapt to the strains we place upon it.  Now when told how dangerous the crunch is, they become hypervigilant and their threat detectors are ready to go off.  Our biomechanical models of teaching people not to bend in the workplace has failed and we have more persistent pain and disability days than ever.  Lets not do this in our impressionable young athletes.


I'm on board if people hate the crunch and think they can create a program that avoids crunching.  We could maybe critique the coach for not having some variety or some exercises that better mimic the functional demands of the sport.  But, where I think we can go too far is by  saying is such absolutist terms that some exercise is completely off limits and that is supported by recent science.  Its not supported - the evidence is not there for that absolute statement.  Right now, its an opinion.  I think coaches can still have science and evidence on their side for justifying the use of crunches.  This, like most coaching, comes down to preferences and experience - there are many paths to achieving a performance goal.

I end this letter much as you ended yours.  Please don't view this as an attack on you.   And to quote "I would be more than happy to answer any question you or your staff (i.e. like minded colleagues) may have as you try to assimilate this information".  If there is some information out there that I have not heard of please send it my way.  I am open to changing my view.  I don't get a nickel every time a kid does a crunch so they could be gone tomorrow if there is some actual strong evidence against their use.



Related Links

1. Jeff Cubos wrote a piece on gymnastics and movement which I remember liking at the time and then completely forgot when I wrote this.  Its worth a read:

Quick EMG Review: Training the rotator cuff trains the scapulothoracic muscles

I had a discussion with a Physio friend of mine about a blog he wrote championing performing scapular stability exercises before rotator cuff exercises.  Because I am bit of a picky bitch I immediately thought that while I can see the clinical rationale for it I don't think the muscles actually do this in practice and thus we had a respectable difference of opinion.  From some old EMG reviews I knew that some of the best exercises to train the lower traps (with out upper trap activity) were actually lame old rotator cuff exercises.  A couple of  years ago I made a few graphics that illustrated this (prompted by a similar discussion on Mike Reinold's blog). Below you will see that shoulder External Rotation (both at 90 degrees and at 0 degrees) are great scapulothoracic exercises.  So while, Jesse might have great success with training scapula before he trains the rotator cuff the explanation for this can probably be explained million other ways then the mechanistic idea that there is some deficit in scapulothoracic function that needs correcting.

Anyways, here are the pictures.  Please feel free to download them and use them in any patient education or other blog posts as you see fit.


Core stability and pain: Is it time to stop using the word stability to explain pain?

sidebridge feet
sidebridge feet

Purpose: To cherry pick a few research articles to suggest that even though our knowledge of core stability is very impressive its link to pain is poor. Nutshell summary: People in pain have spines that function differently than those not in pain.  Many treatments can influence pain.  The spine stability model of low back pain does not explain how people have pain and takes an overly mechanical view of the pain experience.  No test has ever shown that a spine is unstable or how "increasing stability" would lead to a decrease in pain.  Thinking that our spines need more stability or control may be the completely wrong path in explaining how people have pain or how our exercises help them.   Our treatment "corrections" occur not via one specific "corrective" mechanism (e.g. improving stability) but rather through global non-specific mechanisms that our better explained by our understanding of pain neuroscience.  Making the shift from believing that "stability" is the issue with pain can thus free up to choose completely different exercise programs.  Exercise and treatment prescription thus become simpler.  We have preliminary evidence to support this view with the clinical studies that show benefits with the various exercise conditioning programs that train different schools of thought on stability or the just as effective programs that completely ignore any concepts of stability.

Caveat of Ignorance

The purpose of this post is to question things that we say about exercise, low back pain and of course "spine stability".  This is an informed opinion piece and everything I say can be challenged strongly...that's why I write it.  I am also going to put out some "notions" on how I think exercises can help with pain and function.  These are certainly subject to debate and will probably change with time.  I have also ended the piece with a general overview of what I do in Part Two.

A BRIEF and not complete background on spine stability across two hemispheres

Two schools of thought regarding spine stability and low back pain emerged in the 1990s.

The Australians and their inner muscles - Train the local muscles first

The first was based on Bergmark's classification of muscles into "segmental" stabilizers and others into "global" movers.  Segmental stabilizer muscles were often considered to be tonic (constantly on) while the others were phasic (on intermittently to create movement).  This idea of muscles having different roles was suggested decades earlier by Janda.

Low back pain was assumed to occur when the segmental stabilizer muscles were inhibited and the global muscles took over.  The research supporting this idea came from the great work of Paul Hodges (A nice review of Paul Hodges and Motor control can be seen a Todd Hargroves site  In early studies, Paul showed that in healthy subjects the transverse abdominis and the multifidus muscle (two local muscles) should fire in a feedforward manner when someone is asked to lift their arm.  Lifting the arm is a perturbation to the body and muscles in the trunk and legs must turn on for us to keep our balance (some call this "stability').  Dr. Hodges showed that the "Tranny" and MFD turn on before or within 50 milliseconds of the deltoid muscle.  Since the muscles become active  before the deltoid we can assume that the brain did some motor planning to prepare the body for the arm raising - muscle activation was NOT a reactive response to the movement of the arm.

With low back pain Dr. Hodges showed that this feedforward (or motor control planning) was delayed in the Tranny and the MFD.  And BINGO a whole  industry was born and the misapplication of science ran hogwild over common sense.  So that's it.  All Paul showed was that in those with pain you got a DELAY in firing.  No one showed that the tranny was weak, no one showed that the muscle was turned off and no one involved in the research said that the Tranny was the most important muscle on the planet.

But somehow physiotherapists, chiros and personal trainers started telling everyone to suck in their stomach when they did squats because the muscle was erroneously deemed to be super important for spine stability.  This was never what the research suggested and caused fits in the North American Spine researchers who really railed against this simple idea.

The other school of thought - Train general core stability (a brief simple version)


Fortunately, I was innoculated against this because of my MSc with Dr Stu McGill in the late 1990s.  Dr. McGill and Dr. Sylvain Grenier were excellent in challenging the supremacy of the Tranny.  I view their research as less a repudiation of Paul Hodges' ideas and more of an attack of the misuse of Paul Hodges' research.  What McGill and his colleagues had always advocated and also modeled with their biological fidelitous spine model was that spine stability (aka the ability of a system to return to its normal position after a perturbation) was most robust when all of the muscles worked together in the trunk - all muscles were important for stability.  This was again nothing new and we knew this from other joints.  Muscles co-activate, create joint compression and the cost of compression is assumed to be offset by the benefit of stability.   This North American model of stability assumes that all muscles of the trunk work together to balance the stability demands of the spine.  Hence rehabilitation from low back pain should train all the muscles of the trunk in a manner that creates stability but does not do so at a huge compressive cost or adverse tissue loading cost.

Dr. McGill was a leader and pioneer in this.  He was the only one actually evaluating exercises and measuring stability and measuring the compressive/shear loading on the spine to determine which exercises might be "safe".   Dr. McGill was able to classify exercises in to ones which were "safeish" (lower compressive or shear loading on the spine) and others which might have a high compressive penalty but an individual got a good workout (i.e. lots of muscle activity).

The clinical relevance of both the North American and the Australian views are founded on a number of assumptions and unknowns.

What both views assume is that exercise training will make the spine more robust in terms of stability (not more stable, as we know a system is either stable or unstable - you don't make it more stable) and this will lead to less pain and perhaps decrease your injury risk.

Faulty research extrapolations to people in pain and other random stability issues

Below are a number of points regarding the limitations of the relationship between spine stabiliy and pain

1. We do not know why people have low back pain. We do not know what tissue is actually cranky/irritated, fires off a volley of nociception that may ultimately result in the production of pain in the brain (if it even is coming from some cranky/irritated nerve embedded in tissue and is not wholly a production of pain from the brain in response to some perceived threat).  We can not say that a disc is pissed off, a muscle is cranky, a facet joint is upset or if some ligament wants a vacation.  Damage in the spine has a poor correlation to pain. So if you can't identify what tissue is the source of nociception (and we can't) what is the mechanical basis for the prescription of any stability exercise?  How would changing the stability of the spine decrease nociception? If you think spine stability exercises actually change stability parameters by what mechanical means does this change nociception? If you think spine stability exercises help your patients and clients but you can't explain it via a mechanical explanation (but you know it works) do you think there might be  something else going on besides stability issues that you are affecting to influence the perception of pain?

2. Who cares if a muscle is delayed 50 milliseconds?  Really, what relevance does this have.  The muscle turns on eventually and does its job during a task.  Why is a delay of 50 ms relevant in terms of biomechanics.  Is this delay a defense or a defect?  Is the problem in the spine (unlikely) or more a symptom of "something is up" with the brain (more likely, and this is where Dr. Hodges is doing most of his work now yet in popular clinical culture we are stuck at the level of spine). I will go into Hodges work later in another post because I think his work on motor control and the brain may be extremely relevant.  Big point here, Hodges never measured stability.  Just muscle activation in all the muscles that make up the trunk cylinder (side note: he did a wonderful job here, I think his research is excellent, he is an excellent researcher and his contributions to our understanding in the area of motor control are without par.  I would also prognosticate that his future research might bridge the gap from mechanical views of spine and pain neuroscience).  Everyone just jumped on the stability wagon and assumed that it was compromised.  Maybe there is something else going on here besides stability.

3. The argument for the motor control camp against bracing and planking - "Don't brace or do planks because your spine becomes rigid" is a wee bit weak.  This is the argument against the North American model of spine stability and is used to justify"motor control" or low level exercise. It suggests that if you do a bunch of planks you will become rigid and activate your muscles too much. I disagree with this puppetry view of the body. Doing planks will not somehow carry over to rigidity in our activities of daily living.  We aren't puppets where we can tighten and loosen the strings of our spine. This is catastrophizing against a therapy rather by the patient.  These exercises aren't that powerful both in a negative or a positive way.  However, if you actively brace and assume a rigid posture as a choice during all of your normal activities then you can make this argument.  Don't blame the exercise blame the conscious choice of movement.

4. Do you think your patients are really "unstable"? Patients are in pain.  They move differently, you might perceive them to have "tight" muscles.  But is their spine really unstable? Is there a vertebrae in there sloshing around, sliding this way and that, pinching on stuff.  Is the spine really buckling?  We can have patients with high levels of spondylolithesis and their spines are not unstable.  I think we might want to reconsider telling our patients that their backs are unstable and they need stability exercises.  How much fear do you think this creates?  No one has ever shown that a patient with persistent low back pain has funny uncontrolled movements at a segmental level in the spine.

5. But my SI joint needs force closure, I need to train my Tranny or MFD or some bloody fascial sling.

How  is your SI joint unstable?  What wonky movement do you really think is happening in there?   I believe that there is less than 2 degrees of movement and a few millimetres of slide in that SI joint but how is having a delay of 60 milliseconds in one muscle changing this movement?  If it does change that movement why does this cause pain? And so what if the joint slides too much.  Other joints slide around and they don't create nociception.  And if you have a delay in the tranny won't the big, bad global muscles be on at the same time and thus increase force closure and shut down the movement.

These global muscles certainly have the architectural requirements to create force closure.  None of this makes sense. Oh wait, those global muscles are on too much and that causes too much compression in the joint and that causes pain.  Oh, gotcha that makes perfect sense.  But guess what, no consistent research actually suggesting that this happens.   The studies showing increases show increases that are extremely subtle and again how this would cause pain is never laid out in any logical or supported manner.   Well what if that joint is fused?  That seems like a lot of compression.  Should that not be painful yet its not? And why would compression from muscles be painful? Would someone not be better lying down and not lifting weight, walking, running if compression was so nasty for the SI.  More compression on a joint is not necessarily bad and does not lead to pain.  There is something else going on here.

childs pose
childs pose

5. Is it really that bad to get away from the neutral spine? I agree that a neutral spine is generally stronger when the spine is undergoing maximal compressive and shear loading.  Maintaining a neutral spine when deadlifting, doing kettlebell swings, squats and picking up your sofa makes some sense to me.  But do I really need to never bend or twist my spine.  It has a certain amount of movement built into it.  Why would I not use it?  Motion is lotion.  We would never tell another body part to not move.  Taking away movement is how we torture in Guantanomo.  The majority of spine pain does not occur because of we have overloaded it to an extent where it reaches the limits of tissue injury capacity.  This may be one of those issues where we can confuse injury with pain.  Neutral spine bracing can probably help with injury and performance when under high loads but is it necessary to decrease pain in someone getting up from a chair with low back pain?  I will grant that sometimes when you brace and move with a neutral spine and get out of a chair you have less pain.  In other people it gets worse.  Maybe there is something else that explains this besides stability.

6. Patients get better with all types of spine exercise programs.

We have clinical efficacy trials showing that a motor control program (e.g sucking in your belly and then progressing with more global exercises) and a global exercise program helps for low back pain.  So do general exercise programs.   We know that exercise for the spine can help but perhaps it does not matter which exercises we do.  When we get similar results from two different theoretically supported exercise regimes perhaps there is something about the two different programs that is similar.  Perhaps it is that similarity that leads to improvements in pain.  A recent paper by Mannion et al (2012) championed a similar idea.  In other words, we get results but not for the reasons that we think we get results.

7. I think we scare the shit out of people when you tell their spine needs stability

This the default word that many of us tell our clients.  "You're unstable, you can't "control" your movement and that is why you are in pain".  Its so defeatist and catastrophizing and really has little support.  I say we stay away from these words...See my previous post here on this same topic (The words we use can harm)


You can rehab a patient using the two different schools of thought on spine stability.  You will probably have similar results.  Conversely you could just have patients exercise their entire body and they will also show improvements. You will also have good results if you just teach people about pain and give them the confidence to keep moving and not get worried about their "bloody lack of stability" that some therapist told them they once had.

Stability is probably the most inappropriate word we can use to describe our patient's spines that are in pain.  No one has documented that patients in pain have unstable spines nor is there any reliable clinical test for it...yet we have been using this word for twenty years.  That is crazy yet so many of us think that we have to "increase the stability of the spine" in those with low back pain.  No one has shown how any dysfunctions related to "stability" actually cause pain. Again, crazy.  Yet we tell patients they need stability exercises to correct some mysterious bogeyman.   When we get results with completely different movements or exercises that totally conflict in terms of spine stability theory this tells me that the reason our treatment is effective probably has nothing to do with stability.

In part two, I will layout how the spine function is different in people in pain and also give some theories on what treatment does to help our patients.

Fascial NeuroBiology: An explanation for possible manual therapy treatment effects


Below is a guest post from Chris Beardsley in response to a recent post of mine that questioned the possibility of any manual intervention (this includes foam rolling) influencing the physical properties of fascia.  It is also questioned how relevant is to pain and dysfunction.

Research Review: An analysis of Robert Scheip's paper on Fascial Plasticity

By: Chris Beardsley

more from Chris at his website:

Recently, Greg wrote about the problem of understanding if and how manual therapists modify fascia during their work.  At the end of his post, he asked several very pertinent questions, two of which were: “by what means can your hands actually mold, shape or cause some change in fascia?” and  “why can’t they do this in muscle, which is the far more responsive tissue to stress?”

In 2003, Robert Schleip wrote a substantial review article on fascia, in which he detailed how the nervous system might be involved in the process of manual therapy, the main points of which are summarized below.  At the end, we can also see to what extent Schleip’s proposals shed light on Greg’s key questions.

The study: Fascial plasticity – a new neurobiological explanation: part 1, by Schleip, in Journal of Bodywork and Movement Therapies, 2003

Introducing fascia and manual therapy

Fascia surrounds every muscle and even each myofibril within the muscles, as well as wrapping around each internal organ of the body as visceral fascia. It is arranged in sheets that slide over the top of each other and also forms interconnecting lines that join various muscles together.

As part of their work, manual therapists seek to alter the qualities of the fascia in respect of the way it responds to tension or pressure.  They do this in order to bring about reductions in pain or improvements in movement quality.  How these changes and improvements occur is not certain.

In this review article, the Robert Schleip explains that fascia has a neural component as well as a mechanical component.  Schleip proposes that it is this neural component that will help us understand what is happening during manual therapy.  But first, he considers some early theories of what manual therapists are doing.

The classical model

Schleip explains that many of the current manual therapies stem from the practices of Rolf in the 1970s.  Rolf applied considerable manual pressure to fascia with the express purpose of changing its density and arrangement. She proposed that the density of fascia could be altered to make it less dense and therefore more fluid.

This process of making a substance more fluid through mechanical pressure, technically called thixotropy, was later found by researchers to occur following long-term studies on fascia. However, Schleip notes that these thixotropic changes were found to reverse following removal of the pressure, which is not what Rolf was referring to, as manual therapy appears to involve sudden releases of tension as the result of fascial rearrangement under the hands and not slow, gradual alterations.

The piezoelectric model

Schleip explains that another proposal for the mechanism by which fascial plasticity could be explained is piezoelectricity. He notes that fascia researchers have proposed that fibroblasts and fibroclasts, which create and digest collagen fibers, might be responsive to the electric charges that are created through pressure. Many researchers are already aware of similar responses in bone. However, such changes again appear to occur over long periods of time, which is not the mechanism that manual therapists were looking for.

The need for a more rapid self-regulatory system

Because these two basic models cannot be used to explain the phenomenon of release that manual therapists observe in practice, Schleip therefore suggests that the nervous system must be included in any model of fascial plasticity.

Close but not quite: the Golgi reflex arc

Schleip explains that Golgi receptors are an important first step in understanding the neurophysiology of fascia. He explains that these receptors are found in all connective tissues, where they are called Golgi end organs in ligaments and Golgi tendon organs (GTOs) at the junction between muscle and tendon. Importantly, they are arranged in series with fascial fibers (unlike muscle spindles, which are arranged in parallel).

In a stretch, GTOs provide afferent feedback to the spinal cord, which causes the motor neurons to lower their firing rate and therefore reduce the tension produced by the muscle fibers. Schleip explains that researchers originally suggested that manual therapy might lead to a stimulation of GTOs, which could then result in a lower firing rate of certain motor neurons and consequently a reduction in the tension in the muscles. However, later research showed that the passive stretching of a muscle does not in fact stimulate the GTOs. This is because GTOs require active muscular tension to be stimulated because in a passive stretch the elasticity of the muscle absorbs the majority of the deformation, while in an active contraction, it is the tendon that elongates.

However, despite this setback, Schleip notes that this may well still be the right track to follow, as there are many other Golgi receptors in various places.  It is possible that these less-well understood Golgi receptors could be the ones that are stimulated by manual therapists.

Ruffini and Pacini corpuscles

Schleip explains that researchers have found that the thoracolumbar fascia contains many mechanoreceptors.  These mechanoreceptors take the form of three types: (1) large Pacini corpuscles and slightly smaller Paciniform corpuscles, (2) smaller and more longitudinal Ruffini organ, and (3) interstitial muscle receptors.

Schleip comments that the Pacini and Paciniform corpuscles respond to rapid changes in pressure but not to unchanging levels of high pressure.   On the other hand, the Ruffini organs appear to respond in the opposite way. Schleip therefore suggests that it is possible that manual therapists may be stimulating Ruffini endings, which may then cause the nervous system to reduce the tension in some of the motor units within the muscle. (Greg's note: Ruffinis respond to gentle lateral skin stretch - you don't need to dig in deep to activate these receptors)

Schleip comments that only 20% of all the sensory nerves in the body belong to type I and II afferents.  These are the ones that are activated by the muscle spindles, Golgi organs, Pacini corpuscles and Ruffini endings.  On the other hand, he explains that 80% are type III and IV afferents and these have their origin in interstitial muscle receptors. In fact, Schleip explains that many of these receptors are actually within the fascia rather than inside the muscle and many of them are mechanoreceptors, with two main groups: low-threshold pressure units and high-threshold units.  These mechanoreceptors could therefore be the ones responsible for the changes induced by manual therapy.

Motor units not muscles

Schleip reminds us that the central nervous system does not activate single muscles but rather activates groups of motor units. This is an important fact and it may help to explain the release of parts of tissue rather than whole muscles. Exactly what occurs in order for the tissue to release is unclear.  Schleip suggests that it is possible that the release might be because of the lowered firing rate of a few motor units.

So what are manual therapists really doing?

So after all of this analysis, are we any closer to understanding what manual therapist are actually doing when they carry out their work?  Can we answer two of Greg’s key questions, which are “by what means can your hands actually mold, shape or cause some change in fascia?” and  “why can’t they do this in muscle, which is the far more responsive tissue to stress?”

Well, as detailed above, we can see that Schleip proposes a model in which manual therapy or fascial manipulation involves stimulating intra-fascial mechanoreceptors, which then causes changes in the afferent input to the central nervous system, which in turn causes an reduction in the level of muscle activation of specific groups of motor units.

So the answer to Greg’s two questions in the Schleip model would be that the hands are stimulating specific mechanoreceptors in the fascia, that cause afferent feedback to the brain, which then signals the muscles to release.  The effect does not occur in muscle in the same way because the receptors are of different kinds in each tissue.

Of course, there remains a great deal of work to be done to understand the limitations and benefits of this model but it certainly helps to address many of the issues that Greg raised with the idea of the fascia being physically altered.


Chris Beardsley is a co-founder of Strength and Conditioning Research, a monthly publication that summarizes the latest fitness research for strength and sports coaches, personal trainers, and athletes.  The above study review was taken from his Background product, which is collection of over fifty reviews of the most comprehensive review articles and meta-analyses of the fundamental concepts underlying exercise science.

Greg's final note: My question regarding fascia versus muscle as a more responsive tissue to stress only asked about the mechanical properties of these two tissues.  I purposefully left out any discussion of the nervous system which I put at the top of the heap for responsiveness.  When you treat a patient or foam roll yourself and feel better it is the nervous system that you have influenced.  It is unlikely that any changes in the mechanical properties of tissues have occurred.  You aren't more flexible because you broke up some scar tissue.  You may have influenced the viscoelastic properties of tissue (a short term change) but more dominantly you have convinced the nervous system to let you move farther, with greater ease or with greater strength.   An excellent model to understand this is the long term changes that occur in neuromuscular system with stretching.  Following a stretching regime you don't change the stiffness of the muscle and it is unlikely that you change the mechanical architecture of muscle.  Rather, you change the stretch tolerance.  Your nervous system lets you bend farther.  A brief review of stretching science was reviewed previously.

Related questions to think about: If you foamroll do you think that you are changing "tissue quality"?  If so, how are you changing quality and what exactly is tissue quality? Could any changes to "tissue quality"  merely be the nervous systems response to your rolling rather than changes in the structure properties (e.g. ground substance, extracellular matrix, collagen etc).

The relationship between functional tests and athletic performance: Part I - The single leg balance test.

Background: Testing and assessing an individual is popular.  There is an old saying that if you aren't assessing than you are guessing.  The assumption here is that the tests and assessments you do are somehow relevant and meaningful yet I would suggest that the majority of tests and the information gleaned from them hardly change (10% ish) a therapeutic approach once you have heard your patients history. I can have a patient with knee pain and run them through 30 different tests and the results of those tests may hardly change my treatment.  Tests have to provide us meaningful information that we can do something with.What the hell am I talking about?

This blog post is the first of what I hope to be many posts that review the research on our physical tests of purported function.  We make a lot of assumptions about the tests that we commonly use in the clinic.  For example, someone might tell you that your hamstrings or psoas muscles is tight and this will cause you to be hyperlordotic (pelvic anterior tilt) during running.  Or you might hear that you suck at doing a squat and without being able to squat your form during other activities will suffer.  Ultimately, "failing" these tests can lead to you spending time doing lots of "corrective exercise" to fix a test.  This fix is assumed to help with function in other realms of activity.  I want to explore the assumptions behind the tests we use and then the purported fixes that follow.   Some of the topics I might explore:

- assessing tightness in a runner and whether or interventions influence running form

-  how strength training influences joint biomechanics/dynamic form

- how tests of static function relate to dynamic form (e.g. running)

- how simple dynamic tests (e.g. a squat) relate to dynamic form/joint biomechanics

- how measures of joint strength or ROM relate to dynamic form/joint biomechanics

- whether dynamic tests (e.g. a single leg squat) actually give insight into what they are supposed to be giving insight into (e.g. pelvic stability, hip abductor strength and a surrogate for hip control when running etc).

I have touched on these ideas in a number of other posts below:

- specificity of movement training (looks at our ability to change posture) (link here)

- the inability of tests to identify functional hallus limitis  (link here)

- weakness of the prone hip extension (Janda) test. (link here)

So here goes with a look at a common test for runners...

The Standing on one leg balance test is not a relevant test for Runners

Whoa, quick disclaimer.  This first post is an opinion as there is no evidence that specifically looks at this.  However, there is plausibility and I want to make the argument that being able to balance on one leg has nothing to do with the act of running.  This is a common test advocated by rehab professionals (many of whom I greatly respect and one has good book out on running biomechanics and injury here) looking to screen for deficits in function that are assumed to affect ideal running.  But I just don't see it.  It makes no sense to me so if you are reading this and think I am wrong please let me know.

The one leg balance test has you simply balancing on one leg with your eyes open or closed.   If you can't do it the assumption holds that you will be prone to injuries during running or you  may end up with some sort of issues with your running form.  The test is based on the idea that running is merely a succession of leaps through the air interspersed with balancing on one leg until the next leap.  I would respectfully suggest not.  Here is my reasoning:

1. The skill of running is completely different from the skill of one legged balancing. Running is not balancing on one leg as you are not static like you are during the balance test.  These two tasks are completely different. And we should not expect that one can be a surrogate for the other and reflects the same demands.  Our ability to balance is activity and context dependent.  I don't think we can assume that balancing skill in one task will carryover to another completely different task.  The same holds true for that crazy dead fad of doing squats on BOSUs or balance balls with the assumption that training this form of balance will carryover to the balance needs of a golfer or hockey player.  The latter is from Mars and the former is from Uranus.

                                                                   2. Static versus Dynamic Analogy

Running is dynamic.  You do not pause and balance on one leg.  Comparing one leg balance to running is like comparing the ability of a stationary top to balance on its point with the ability of a SPINNING top to balance on its point.  Suggesting that we need to be able to balance on one leg in order to run safely is like arguing that in order to ride a bike properly you have to be able to hop on the bike, clip in and then balance in place without moving.  I have been having my five year old do this for months before I will let her ride her bike and surprisingly she still can't ride a bike...and that bruised little monkey hates me for it.  (Obvious disclaimer: I am kidding).  But this example illustrates my point.  If it seems absurd on a bike then it is just as absurd on your feet.

Caveat of Ignorance

I reserve the right to change my opinion with new research or with fresh ideas.  I can also be convinced to think otherwise.  My mind is not as closed as my writing might indicate.  As a further hedge I also recognize that it is possible that some cross sectional study might come out and suggest that one leg balance is compromised in runners with knee/ankle/back/hip pain.  I don't doubt a correlation might exist BUT this does not imply that the decrease in balance ability is the cause of the pain.  This can merely be correlation.

Last, this does not mean that one leg balance training can not be beneficial for runners in preventing injuries or improving performance.  The act of doing one leg squats, one leg hops, hip airplanes, front scales and other great exercises could certainly improve your balance and would also change other aspects of muscle function.  In this case the improvements in balance could be secondary to other aspects of improved function that are the true drivers of improvement in performance and injury resilience.

Last point, honest- perhaps we over assess and our assessments change nothing in practice

What I think can be argued with this assessment test (along with many other tests) is that it just isn't necessary and is plausibly not valid.  If I am working with a runner or an athlete their conditioning program would be all encompassing and comprehensive.  This would include exercises that ultimately improve their balance along with other variables.  I don't need a rudimentary test to tell me to add these exercises.  I just need that human in front of me that tells me that they want to run and be an athlete.  I don't want to freak them out because of their shitty balance or advise against doing these wonderful exercises because they have great balance during a static, albeit not relevant to running test.

Exceptions to the exceptional joint by joint approach

Exceptions to the joint-by-joint approach – by Greg Lehman with commentary from Bret Contreras. by Greg Lehman and Bret Contreras

Quick Background: The joint-by-joint (JBJ) approach, popularized by Mike Boyle and Gray Cook (link here), is a method of categorizing how each joint should ideally function and what tendencies a joint might have toward dysfunction. It also suggests how joints interact with each other and might provide shortcuts to identifying shortcomings in a joint’s or system's functioning in the cause or persistence of pain, injury or less than ideal performance.  The assumption of the theory is best illustrated with this quote from its original description:

Injuries relate closely to proper joint function, or more appropriately, to joint dysfunction. Problems at one joint usually show up as pain in the joint above or below.

Purpose of this critique: On first glance (and still) of the JBJ I thought it was beautiful. So simple, so elegant.  It is a common idea to look both above and below a painful joint to hopefully find the criminal that is creating the painful victim.  So at first blush I loved it, but as I thought more about it, I kept seeing exceptions to the rule.  And if exceptions exist to this theory, then maybe the theory is less useful and does not accurately describe function.

This post is not specifically evaluating whether an assumed dysfunction at one joint leads to injuries at other joints (i.e. questioning regional interdependence), although I think that is important to do.  I am merely pointing out exceptions in the tendencies at each joint and how these exceptions suggest to me that joints may not be governed by these tendencies and are obviously more complicated.  If there are common exceptions at every joint then the JBJ does not accurately reflect reality.

A caveat about critiques:I could not create this observation (the JBJ) about the body – I am not that clever or astute. I have a lot of respect for Gray Cook and Mike Boyle in creating and publishing their ideas.  However, I think that theories are meant to be tested.  The book Movement is attempting to become a university textbook and the ideas within it should be subject to some rigour.   No one would consider it poor taste to critique a theory about the origins of the universe just because you had a lot of respect for the physicist.  In the same light, the JBJ theory should be questioned because it makes bold statements about the ideal functioning of the body.  Further, looking and questioning theory helps me understand it.  All this being said, I think that anything I write here Gray Cook already knows.  After writing this post in rough months ago, I essentially found a “rebuttal” by Gray Cook called expanding on the joint by joint approach here.  He explains the limits and utility of the JBJ well and also provides more food for discussion on how the body should function.  He even comments on exceptions

Of course you will find exceptions, but the more you work in exercise and rehabilitation, the more you will see these common tendencies, patterns and problems.

Bret’s Notes: I agree – I could not have come up with the JBJ approach as I’m not that clever either. I think it’s a brilliant model and is incredibly simple and elegant. I've learned a ton through Gray and Mike over the years but everything should be questioned and scrutinized in the best interest of scientific advancement.

I’m more of a biomechanics and CSCS, so I don’t have experience in physical therapy. But coaches and trainers typically train folks in various states of dysfunction, which gives me some confidence in critiquing the JBJ approach. Based on my limited experience in rehab, I agree with Gray in that the body tends to break down in the manner described below, and in general I’m a supporter of the JBJ, but like most models, it needs additional clarification.

Here is the joint by joint in a nutshell. I list the joint and what is assumed to be needed at that joint

1st MTP: mobility Midfoot: stability ankle: mobility knee: stability hip: mobility lumbar spine: stability thoracic spine: mobility scapulothoracic region: stability glenohumeral joint: mobility lower cervical spine: stability upper cervical spine: mobility

And in Gray Cook’s words this is how he describes it:

A quick summary goes like this—

1. The foot has a tendency toward sloppiness and therefore could benefit from greater amounts of stability and motor control. We can blame poor footwear, weak feet and exercises that neglect the foot, but the point is that the majority of our feet could be more stable.

2. The ankle has a tendency toward stiffness and therefore could benefit from greater amounts of mobility and flexibility. This is particularly evident in the common tendency toward dorsiflexion limitation.

3. The knee has a tendency toward sloppiness and therefore could benefit from greater amounts of stability and motor control. This tendency usually predates knee injuries and degeneration that actually make it become stiff.

4. The hip has a tendency toward stiffness and therefore could benefit from greater amounts of mobility and flexibility. This is particularly evident on range-of-motion testing for extension, medial and lateral rotation.

5. The lumbar and sacral region has a tendency toward sloppiness and therefore could benefit from greater amounts of stability and motor control. This region sits at the crossroads of mechanical stress, and lack of motor control is often replaced with generalized stiffness as a survival strategy.

6. The thoracic region has a tendency toward stiffness and therefore could benefit from greater amounts of mobility and flexibility. The architecture of this region is designed for support, but poor postural habits can promote stiffness.

7. The middle and lower cervical regions have a tendency toward sloppiness and therefore could benefit from greater amounts of stability and motor control.

8. The upper cervical region has a tendency toward stiffness and therefore could benefit from greater amounts of mobility and flexibility.

9. The shoulder scapular region has a tendency toward sloppiness and therefore could benefit from greater amounts of stability and motor control. Scapular substitution represents this problem and is a common theme in shoulder rehabilitation.

10. The shoulder joint has a tendency toward stiffness and therefore could benefit from greater amounts of mobility and flexibility.

It is theorized that if a joint moves away from this idealized function (and demonstrates the faulty “tendencies), one will experience dysfunction up or down the chain. This is termed regional interdependence.  What I question is whether this is a useful or accurate viewpoint on human function considering that every joint has an exception to these tendencies and many of these exceptions are not just minor outliers.  They are quite robust and prevalent.

Bret’s Notes: While I agree with Gray’s summary and am a big proponent of the “regional interdependence” theory, I’m also of the belief that all joints need specific levels of mobility and stability training. I’m with Greg here – there are certainly lots of exceptions to the JBJ, and some of the exceptions are incredibly important, which casts doubt on the simplicity of the JBJ model.

 The Exceptions

1st MTP (mobility assumed to be good)

Having “instability” or too much movement of the 1st metatarsal (e.g. it dorsiflexes) does not allow the the MTP to dorsiflex. Therefore, you need stability at this joint for mobility to occur.  Therefore the assumed deficit is a lack of mobility whereas the actual tendency towards dysfunction is a lack of stability. The perceived dysfunction at this joint is termed functional hallucis limitis and has been questioned by many biomechanists.   See a review here.

Midfoot (stability assumed to be good)

The pronation (or a sloppy midfoot) bogeyman shows poor correlation to injury. Yes, the foot must supinate (e.g. create stability) to lock out the midfoot for power production but we can’t assume that having a lot of pronation means that we lose force production and are at risk for injury up the chain. This has not been proven as a consistent risk factor for injuries yet it persists despite many reviews questioning its significance.

The video below shows a former world recorder holder in the marathon and 10k demonstrating huge amounts of pronation. It is hard to argue that he is leaking energy or that this pronation leads to some other damage up his kinetic chain.  This runner is 38, still running and still competing.

Ankle (needs mobility, tendency towards restriction):

An exception would be the tight calf muscles demonstrated in runners with a corresponding increased mechanical efficiency (a recent cherry picked paper here and here).  There is also a lack of research linking tightness in dorsiflexion with prospective injuries.

Further, I am not arguing that increasing flexibility will negatively influence the performance in runners.  I have heard this argued and think the jury is still out.  Here are a few abstracts showing no change in running economy following acute and chronic stretching regimes (herehere and here).  Note, I did cherry pick here.  There is some research suggesting acute stretching does influence running performance, point is, it is still up for debate.

Final point: athletes can get by without restricted dorsiflexion in many sports.  Do we always want to go changing this?  Can you with certainty conclude that a lack of dorsiflexion is a true dysfunction? I think a massive post on restricted dorsiflexion and injury, form and performance would be cool.  Any takers?

Bret’s Notes: Greg raises some excellent points. I’m of the belief that lack of dorsiflexion is certainly dangerous in weight training as it can lead to form decrements such as lumbar flexion during heavy squatting. However, it’s probably not as dangerous for sports performance where you don’t have heavy weight on the back. Furthermore, there’s probably a big difference between the biomechanics of “stiff joints” in weaker, sedentary individuals versus athletes. For example, joint stiffness adaptations in athletes may be caused by a shifting of the optimal length and/or alterations in sarcomeres, titin stiffness, or connective tissue stiffness. However, in a perfect world, most coaches would agree that they want their athletes to possess sufficient ankle dorsiflexion ROM, so I tend to focus on what’s “optimal,” rather than what’s “acceptable.” Therefore I agree, with Gray and Mike; most people would benefit more from mobility-related training than from stability-related training for the ankle joint.

Knee (stability is assumed to be needed)

Of course we need stability but do we actually see unstable knees except when a ligament is blown out? While laxity of ligaments (abstract here)may predispose to injury we can't see this with the naked eye.  Depending on how you define stability, are you sure that it is the knee that is “unstable” when you see the knee looking sloppy?  The knee just follows the path set by the hip.  There isn’t an instability that can be seen or measured until you damage some internal restraint.  We don’t train the knee for stability we train the hips in association with the knees.  The knees benefit from the control of the gmax, gmed, gmin and hip rotators and on an unseen level the hamstrings.

An exception to the JBJ rule is the type of mobility we need in our knees (i.e. the tendency for the knees to lose mobility). We need full extension.  A loss of this following injury or surgery is huge for dysfunction.  Check this link here. Further, we need the joint surfaces to translate and rotate – that is why some therapists do Mulligan techniques, manipulations or mobilizations.

Bottom Line: The knees are stable because of what we do at the hips or the passive restraint system of the ligaments (untrainable).  Mobility of the knee is also important. Because of these exceptions I don’t see there being a tendency to sloppiness at the knee because that sloppiness (if relevant) is from the hip and therefore I question the JBJ.

Bret’s Notes: For the past few years, I’ve been heavily influenced by researcher Chris Powers, who feels that knee issues are almost always due to problems at the hips. However, a conversation with a high-level biomechanist in Auckland in addition to a couple of recent journal articles led me to believe that optimally-functioning knees are not just about the hips – we need strong quads to influence pressure distribution, strong hamstrings as co-contractors for stability, and I’m still open for VMO potential, but more research is needed in this area. So knee health is highly dependent on hip mechanics, but I feel the knee can be trained for improved stability and improved biomechanics, meaning that the inherent forces and stress distributions (patellofemoral joint contact area is increased via quadriceps strengthening, the ACL joint is spared from hamstring co-contraction, etc.) can be reduced through strengthening muscles acting on the knee joint. So I agree with Greg about the importance of hip stability but with Gray and Mike about the importance of knee stability.

Greg's response to Bret: I am not arguing that knee muscles are involved in stability and health of the knee.  What I want to emphasize is that the sloppiness we see at the knee joint is more than likely due to alterations in how the hip or spine controls the position of the knee or even the soleus. Poor mechanics of the knee are primarily controlled by something else other than the muscles that just cross the knee.

Hip (assumed need – mobility). 

The rationale here is that if you don’t move in your hips you will move your lumbar spine and predispose yourself to injury. I love this idea and am reading a great PhD thesis by Janice Moreside (a student of Stu McGills) on this idea exactly.

The obvious exception to the lack of mobility tendency would be the many biomechanists who argue that knee injuries are related to alterations in control of the femur (increased hip internal rotation and adduction). You can certainly make a strong case for hip “stability” being required to prevent injuries and certainly being a greater risk factor for dysfunction. Because of research funding and the difficulty of research, hip extensibility has less research supporting its relationship with injury. Regardless, these competing biomechanical rationales suggest that the JBJ theory has problems.

However, it should be noted that the authors of the JBJ already recognized this exception and wrote:

The exception to the rule seems to be at the hip. The hip can be both immobile and unstable, resulting in knee pain from the instability–a weak hip will allow internal rotation and adduction of the femur–or back pain from the immobility.

However, we can also question whether losses in sagittal plane range of motion does lead to increased strain, changes in movement at the lumbar spine and subsequent injury in the spine.  I’ve questioned this before here. One example is the research that looks at individuals who have a loss of hip extension on a Thomas test don’t actually create more motion in the lumbar spines during running.  Yes, this is one study and more needs to be done.  I am just saying that this is again something worthy of discussing.  Last, in Dr. Moreside’s PhD thesis, increases in hip extension following training were not associated with decreases in the amount of spine extension that occur with an active hip extension movement (I am following this up with a post summarizing Dr. Moreside’s great hip and spine research if you are interested in reading more of this).

Hip Bottom Line: The Hips need both stability (motor control training, movement, strength, endurance) and mobility.

Bret’s Notes: I am in complete agreement with Greg here. I don’t know what’s more “important” for injury-prevention purposes – hip mobility or hip stability, but suffice to say they’re both incredibly important. Hip instability begets knee pain. For more info on this topic I highly recommend reading THIS paper (click on the link and download the pdf). There exists research indicating that hip instability contributes to low back pain and anterior hip pain as well.

I believe that the exception at the hip provides the second biggest blow to the JBJ approach. In theory, it would be lovely to see this beautiful, alternating pattern of joints needing mobility and joints needing stability. While it’s a nice overall model, the exceptions must be noted. This brings me to my next point – the biggest single blow to the JBJ approach:

 The Pelvis!

The pelvis joint is completely ignored in the JBJ model. The pelvis needs mobility, but even more important is pelvic stability. I am of the belief that pelvic instability is a huge criminal in terms of creating mechanical insults to the spine. I’m not sure if Gray and Mike purposely left out the pelvis in attempts to simplify the model and allow for the “alternating” approach, or if they simply overlooked it, but I feel it’s time we started giving the pelvis much more attention.

Many individuals are unable to adequately tilt their pelvis in various directions, and they lack the ability to dissociate the spine and the pelvis. If the joint doesn’t function properly dynamically then I doubt it functions properly statically. Individuals typically possess incredibly poor motor control in this region and could benefit greatly from static and dynamic strength training for the pelvis. Core stability doesn’t just involve the lumbar spine and hips; the pelvis is just as important. Furthermore, research has shown that muscle function is unique depending whether core muscles are acting on the spine/ribcage or the pelvis.

I believe that this is a major area for future research and improvements in human movement mechanics and I have personally achieved success with my clients in implementing pelvic neuromuscular training strategies.

Lumbar spine (assumed to need stability and have a tendency to being sloppy):

I won’t even get into this. Too big, too messy, too contentious.

I would just ask “do you think it is all that bad to bend your lumbar spine during simple unloaded activities”?  Can you name 5 sports that see huge ranges of spine motion as being necessary for success in that sport?  Do some athletes even use a flexed spine to lift heavy weights (although they can still be stable, a great study by Stu McGill is here.  Lots of lumbar flexion during lifting but that position was buttressed with stability derived via cocontraction).

Take a look at the range of motion in the spine in the graph below.

Do you think it is only occurring in the thoracic spine?  Please note, the pictures are not exactly linked with the ROM curves.

Bottom line: the lumbar spine probably needs some mobility and stability.

Bret’s Notes: The lumbar spine indeed loses considerable motility as we age, and while I agree that stability is more important than mobility in this region (research shows that increased spinal flexibility doesn’t reduce back pain, but again, I focus on what’s “ideal,” not what’s “acceptable”), I don’t feel that it’s ideal to accept mobility losses. Smart training can allow people to keep their 3D lumbar spine mobility (or even build lumbar spine mobility) while not posing too much of a threat for injury, as long as end-ranges of spinal motion are avoided. That said, even when we think we’re stabilizing the core, the spine is moving. This has been shown with squats, deadlifts, sprinting, and even kettlebell swings.

Over time, with proper training, the brain probably figures out the best compromise between performance and spinal health and determines just how much ROM to allow in each segment and how much muscle activation is required in the various core muscles to “tune the stiffness” and stabilize in non-neutral positions.

Greg's outrageous comment to Bret: This might come as a surprise but of the studies done to date there isn't consistent support that spine stability exercises are markedly more effective than other exercises (motor control exercise, graded exposure, general exercise) for the treatment of back pain.  Pain is a separate beast and we have to be cautious in thinking that just  hammering stability will decrease pain.  Prevention of injury may be something different.

Thoracic spine (needs mobility): I don’t have a strong exception that I absolutely stand behind. I just think that like any joint it needs to both move and be challenged with physical stress (stability).  There are researchers that suggest the thoracic spine is also subject to the same form and force closure demands that the SI joint exhibits.  I believe they might argue that “instability” or a tendency to sloppiness can also occur here if you look for it (see Diane Lee and LJ Lee in their Thorax book  Just another exception to consider.

 Bret’s Notes: Sure the ribcage adds considerably to the t-spine’s stability, so this portion of the spine has a built-in stabilizer. However, some research shows that the thoracic intervertebral discs suffer from an alarming number of herniations just as the lumbar discs do. Of course, this isn’t well-correlated with pain, but I’m sure that most individuals “break down” in terms of erector spinae weakening and can benefit biomechanically from increased static thoracic extension strength, especially in weight-training.

When under load, you can have all the mobility needed but without strength the upper back could round too far forward and cause problems over time. Furthermore, I feel that poor postural habits in addition to weak muscles contribute to the stiffening of the t-spine as muscle force is required to pull the t-spine into proper position in the first place.

Scapula (needs stability): this one is easy using biomechanical reasoning. We need the scapula to get out of the way of the arm bone. There is cross sectional evidence suggesting that those with impingement type pain have less posterior rotation and elevation of the scapula. This joint has to move…it must posteriorly tilt and upwardly rotate.  I suppose you could flip things on their heads and say that a lack of mobility in one direction is really a lack of stability in the opposite direction.  The motor control deficit here is that we don’t control the scap to get out of the way.  Again, I recognize that stability can be compromised when we don’t want the scap to move but the point of the article is to show that every joint has an exception and that all capacities are important.

Glenohumeral (needs mobility): Well of course it needs mobility, can’t argue that, but biomechanically we also argue that the head of the humerus needs to be positioned properly in the socket…this is stability. Deviations from this (anterior or superior glide) are biomechanically considered to be linked with pain.  Again, mobility and stability are both important.


The tendencies at some joints can certainly occur and they might even be linked with pain and dysfunction (although, I think this can be debated in another post and was only briefly touched on here, e.g. pronation), but the point of this blogpost was to highlight how often we have exceptions to the JBJ. So we have observed tendencies but also common exceptions at the same joint.  When we have this many exceptions to the JBJ rule we might want to consider questioning whether this is a theory that adequately describes human function.  Is it even helpful if there are so many exceptions?  What I would stress here is that Gray Cook probably already knows this stuff and would be flexible in his programming and assessments about function anyway. I bet he doesn’t even need the JBJ and nor do you. But, lots of other people may not view the JBJ with such flexibility and can become dogmatic and rigid.  This post was for those individuals.

Bret’s Notes: I’m sure that Gray and Mike are well-aware that each joint needs both mobility and stability, but the JBJ model is directed at dysfunction and portrays how the effects of gravity, poor posture, and sedentarism typically affect joint behavior. I still like the JBJ and teach it to my students, but I also point out the exceptions and stress to them that all joints need varying degrees of mobility and stability, and I especially stress the contentions at the hips and pelvis. The model created by Gray and Mike is both genius and useful, but it needs further explanation to be more complete.

Gray’s main point in this quote hits the nail on the head: “The whole purpose of the joint-by-joint concept is to realize generalities…The examples are there to make you think above and below the area you’re working on and in the things you’re asking for.”

So the way I see it, the JBJ model could be improved dramatically just by adding in the following two adjustments:

1st MTP: mobility

Midfoot: stability

ankle: mobility

knee: stability

hip: mobility and stability

pelvis: mobility and stability

lumbar spine: stability

thoracic spine: mobility

scapulothoracic region: stability

glenohumeral joint: mobility

lower cervical spine: stability

upper cervical spine: mobility

Greg's Last Points

The exceptions pointed out above illustrate to me that most joints need both mobility and stability and we can make an argument to train capacity in all realms of joint function to maximize a happy body.  I would even go so far to suggest that some tendencies to dysfunction suggested in the JBJ (and by Bret) are not even dysfunctions...just the normal variability that the body possesses.  I disagree somewhat with Bret with the tendencies he suggested above.  That is cool.  My opinions are provisional and will certainly change with more information.  Its interesting that you can't look at the same research and have different final opinions.

Last, I won’t even discuss pain.  The link between assumed “altered biomechanics”, “poor form” and future injury and pain is extremely weak and may even be non-existent.  This is something again worth discussing, just not here.

Postural correction and changing posture. Can we treat our patients like puppets?

Audience: Therapists and Strength CoachesPurpose: To justify the use of a variety of exercises (even general exercises) for training, rehabilitation and injury prevention and question the application of movement specificity principles.

The Gist of this Post: Specificity of training is an important component of rehabilitation and strength and conditioning but I think the application of specificity can be taken too far when we attempt to mold our posture.

A related post by Tony Ingram touches on many of these ideas in relation to pain and posture.


The godfather of specificity was a former professor of mine, Digby Sale.  For a brief review see here.  Very briefly the research suggests:

"Evidence supports exercise-type specificity; the greatest training effects occur when the same exercise type is used for both testing and training. Range-of-motion (ROM) specificity is supported; strength improvements are greatest at the exercised joint angles, with enough carryover to strengthen ROMs precluded from direct training due to injury. Velocity specificity is supported; strength gains are consistently greatest at the training velocity, with some carryover. Some studies have produced a training effect only for velocities at and below the training velocity while others have produced effects around the training velocity"

Another great review article is by Cronin et al (2002) link here:  A quick quote from the abstract:

"It has been suggested that training at a specific velocity improves strength mainly at that velocity and as velocity deviates from the trained velocity, the less effective training will be. However, the research describing velocity-specific adaptation and the transference of these adaptations to other movement velocities is by no means clear".

My thesis: The applications of specificity can be taken too far in three ways

1. The repeated performance of an exercise leads to plastic deformation of tissues or changes in motor control that cause significant changes in posture and movement capabilities (aka form).

I question if the body is really this malleable and ahem, stupid.  It assumes that consistently training certain movements makes you move in that specific way and you lose the ability to move in other ways (i.e. your posture and form become changed).   I believe it is an inappropriate extension of you become what you train.  It is a training belief related to the idea that hip flexors become shortened because we sit all day (see a previous post here).  I don't doubt that you can train habits of movement that might carryover into other tasks.  What I question is whether are tissues are so readily plastic and they can't control their destiny because of some passive changes in the make up of the tissue. 

An example...

I was listening to a podcast where the speakers objected to what they felt was the rampant, unjustified, often silly and apparently detrimental usage of the front and side plank (bridge) by physiotherapists and trainers the world over.   What the speakers argued was that performing these repetitive planks with no motion between the hips and the thorax would somehow create runners that will run like robots and lose the ability to dissociate the hips from the thorax.  As if twenty minutes of planking a week will somehow carryover to the automatic movements that occur during running.

I just don't buy this and consider it such a pessimistic and wholly unfounded structural view of the body.  It assumes that the body is stupid and a few minutes of planking will somehow override what ever neural control mechanisms, not to mention physical forces, that create subtle movement in the spine when we run. 

A brief review of 3D spine kinematics during running can be found here (Schache et al 2002 ) and here (Saunders et al 2005).

We don't change form through simple exercises...or do we?

The belief that planking makes you rigid and run like a robot has not been tested but assumes that planking will somehow stiffen up all the muscles of the trunk within the neutral zone and also cause our brain to change the automatic way in which it recruits muscles during locomotion.  That is some powerful planking to override our nervous system like that.  It is very difficult to change running kinematics even when we try to change running kinematics by volitionally changing our posture.  But somehow, a little bit of planking can do this despite us trying to run normally. 

Same holds true for the knock on the curl ups.  I think it is a fair to critique curl ups  for other reasons (e.g. there are better exercises, you may not think they are functional, you don't like the idea of compressing a disc with some flexion) but I don't think we can turn people into kyphotic zombies.  Unless you've been bit by a Kyphotic zombie and those are biochemical changes not biomechanical. Curl ups get critiqued because it is assumed that doing a lot of curl ups will end of shortening the rectus abdominis and will therefore be constantly flexed.  I just don't see any research suggesting that this happens.  While we might increase the stiffness of the rectus abdominis this is different than making that muscle shorter at its resting length.  The muscle has a stress strain curve where there is hardly any resistance to movement around its resting length (i.e. the neutral zone) while strength training might shift that curve if stiffness increases it still has a "toe region" of the neutral zone where hardly any passive force is created.  Certainly not enough to crank down the thoracic spine and all of the other opposing muscle groups

Some Research on changing posture and form through exercise

Here is a sampling of studies looking at both strengthening and stretching programs designed to change Scapular position or posture in general .  This is ridiculously difficult to do.  None of the following studies were able to do it:

- a review here by Con Hrysomallis looking at Shoulder position

- a review by Hrysomallis looking in general at the ability to change posture:

-Wang et al (1999) Stretching and strengthening exercises: their effect on three-dimensional scapular kinematics.:

- McClure et al (2004) Shoulder function and 3-dimensional kinematics in people with shoulder impingement syndrome before and after a 6-week exercise program:

-Hibbard et al (2012) Effect of a 6-Week Strengthening Program on Shoulder and Scapular Stabilizer Strength and Scapular Kinematics in Division I Collegiate Swimmers:

Serendipity of the internets

I have been writing and thinking about this post for months and along comes a post from Bret Contreras arguing that strength training alone does not change form. He argues that motor control training changes form.  This is a component of what I am trying to say.  Here is a link to Bret's piece

MY ARGUMENT DOES HAVE SOME RESEARCH AGAINST IT...Sort of weaken my arguement (and engender some healthy doubt or hope :) ) there are some papers that do show a change in posture albeit inconsistent. 

 1. Here is a great paper by Scannell and McGill(2003) - Stu does all the great stuff!

But, there was not a change in the stiffness of the spine nor did this lordotic static change carryover to a functional task

There were no changes in the size and location of the NZ of each group recorded during the mid-training and posttraining tests.

Relative to the pretraining test, all 3 groups sat in more lumbar flexion during the mid-training test (P=.005) (lumbar flexion increased by 4° in subjects with hypolordosis, by 5° in subjects with hyperlordosis, and by 5° in control subjects) and the posttraining test (P>.5) (flexion increased by only 1° more in all 3 groups relative to the mid-training test results). The changes in the sitting position between the pretraining and mid-training tests were seen in all 3 groups and therefore cannot be considered a treatment effect

Above Figure Description: The changes in the neutral zone (depicted by the black bars) and the lumbar position during sitting, standing, and walking (50% level of the amplitude probability distribution function) across the pretraining test (1), mid-training test (2), and posttraining test (3) are shown. The subjects with hyperlordosis stood in less lumbar extension during the mid-training and posttraining tests. Subjects in all 3 groups sat in more lumbar flexion during the mid-training test. No changes in the lumbar spine position during walking were found during the mid-training and posttraining tests. Positive values greater than the neutral zone represent extension.

An interpretation of the Scannell Paper

The Scanell paper certainly shows a change in resting posture (e.g. lordosis) but we see no change in lordosis during functional activities nor is their change in the stiffness of the spine.  This suggests that we aren't changing passive properties of the spinal tissues with these exercises but we are doing something else to change resting posture.  Did the participants change habits while doing the exercises and became more comfortable standing in a more neutral posture?  Was their standing posture a choice?

What I take from the conflicting research is that if there is a change in form with basic exercise it is not robust nor is it consistent.  And there is not sufficient information or even biological plausibility to assume that doing the plank daily for 3-5 minutes will somehow result in all of us moving like robots during an activity like running which appears governed by more central patterns.  I can't even imagine trying to run with a non-moving spine.  

A few more papers showing changes in posture with exercise:

2.  Misconception #2: Exercises must be specific in terms of every variable to be considered "specific"

This happened to me when I was trying to publish a paper on strength and conditioning for golf and I had the pickiest reviewer who kept saying that none of the exercises I was advocating as being specific to golf (e.g cable chops, one arm cable punches/pulls, weighted swings, med ball rotational tosses, Swing fan swinging) were specific to the golf swing as they all either had slightly different mechanics or different speeds.  I believe that I asked if a weighted sled exercise was specific to running and I was told that no it was not. I just think it is realistic to assume that a definition of specificity recognizes that some differences in terms of velocity or kinematics is acceptable and that we will still get benefits in the task we want to improve.  The early papers mentioned by Behm and Sale and by Cronin et al certainly support this.

3. Misconception #3: Non-specific or general strength exercises can't carry over to performance/injury reduction for specific tasks.

This is essentially the opposite of number two.  It is a pretty big debate and would rear its head with those arguing for functional exercises versus non specific exercises.  You could also ask "Are specific or functional tasks always better than general exercise?".  MOst of the time people assume (myself included) that specific tasks are better but I think we would ignoring a lot of research that suggests otherwise.

 A specific example can be seen in respect to doing a plank exercise for a runner.  Obviously, no runner gets into this position.  It is not specific in terms of body position with respect to gravity, the movements of running and certainly not velocity.   So the knock against this exercise is that it is not specific to running and therefore can't help the runner.  And I say, who cares? Why does it have to be specific to running? Can't an exercise give us benefits that carryover to other tasks?  Of course they can.  The plank obviously trains the trunk muscles and the lateral hip muscles.  We know that many runners with pain have weak abductors (whether this is a cause or correlation is hotly contested) and at the same time we have some evidence to suggest that training the hip musculature (with exercises that are nothing like the running movement) can be effective in returning people to running and decreasing pain in those with knee injuries. 

What is the mechanism for non-specific exercise benefits

Interestingly, the reason why these exercises are effective often has nothing to do with the purported biomechanical rationale for training the hip musculature in runners.  Specifically, we can advocate training the hip musculature under the assumption that this might change hip valgus and femoral internal rotation during the stride as there is some evidence to suggest that these mechanics are occasionally linked with knee dysfunction.  However, when we implement these hip exercise programs (again with exercises that are basic and totally not specific to running) we will see decreases in pain and return to running WITHOUT changes in the assumed dysfunctional mechanics (check out these papers here, here, here,  here  and here). Of course there are some studies showing that form does change (here and here - albeit no change in kinematics, just moment)  To me, this suggests that there is something generally beneficial about these exercises and that specificity does not have to occur.  But, if you want to change running form than the intervention should probably specifically try to change running form through feedback and training. (see here, here and here)

In terms of performance, I am biased, I do like creating exercises or stealing them that are somehow specific to the task at hand.  However, I recognize that there is a great deal of research that shows that exercises that are not specific to the sporting task can still improve performance (e.g. squats for sprinters).  Again, arguing that specificity is not always necessary.   My main point is the body craves variety and our programming and rehabilitation should reflect this desire.  We don't  have to be so rigid in our prescriptions or believe that there is only one way to get benefits.  The body adapts, lets exult in this.

Future Posts Related to this Topic

This post is a bit of a jumble with a lot different ideas.  I have few posts written half-assed in my head that are related. If anyone wants to write any of the following with me please let me know.  Or stay tuned for these upcoming posts.

These posts will ideally be a gathering of information that generates questions.

1. A catalogue of exceptions to the Joint-by-Joint Approach.

This post will use the framework of the joint by joint theory to consider the research that looks at injury risk factors.  It is not so much a critique of the approach but more a means to understand it limits and explore how it can still be useful.

2. A catalogue of examples that support the Joint-by-Joint Approach

3. Is there an optimal way to move? A catalogue of theories and evidence for ideal movement 

4. Can form be changed with via mechanical changes in tissue?

5. Changing running form through feedback and training.