Stretching Tendons: What can we do and why we should challenge our biases. Part 2

In a previous post we looked at how long term static stretching might influence the structural properties of the connective and muscular tissue that crosses joints.  The question was whether static stretching changes passive stiffness.  Interestingly, we saw good research showing different results.  One point of that post was to continually challenge our own biases by looking to other research to try to disprove what you think you know.  This post will continue in that vein and will also blow you away with the most interesting thing about long term static stretching that never gets any press Static stretching has been vilified for more than decade.  Like most things we've probably run too far in our practical applications of the research. Briefly, static stretching for more than 60 seconds will result in very tiny strength and power losses during simple activities between 1-5%. (Kay and Blazevich 2012, Simic 2013)  But these losses in performance don't really translate to more complicated sport performances and other researchers  have argued that this muscle inhibition found in a lab setting might not even occur during sport performance (McHugh 2010).

Recent systematic and narrative reviews suggest pre-participation static stretching does not reduce injuries but again there might be a catch.  Both McHugh (2010), Witvrouw (2007) and Jamtvedt (2010) have argued that pre-participation static stretching might be injury preventive if you segregate injuries related to muscular strains and the stretch shortening cycle.  This post can't go into the gory details of all of that research but the point is to be aware of these dissenting views.

OK.  Now here is the research that I find amazing.  And again, it strong challenges what we think static stretching does.

Terrible anterior pelvic tilt.  Those poor hip flexors
Terrible anterior pelvic tilt. Those poor hip flexors

How Long Term Static Stretching Could Increase Performance

Again, my previous bias was to not stretch with the exception of sports that needed that range of motion.  I still think there are more important things to do with your time but that is another topic.  We have some research that long term static stretching actually improves performance in things like strength, power and speed (Kokkonen 2007).  To me, those findings are just counter-intuitive and I had difficulty reconciling them with my views of stretching in general.

But there is one possible explanation for changes in performance after long term static stretching and that is related to Tendon Viscoelasticity and this is where shit gets amazing.

Biological materials (e.g tendon and connective tissue) exhibit viscoelastic behaviour.  Meaning they have a hysteresis loop when you load and then unload them.  Like they have a memory.  Their resistance to stretch or stiffness is influenced by previous loading.

How researchers measures tendon stiffness is to take an ultrasound transducer and measure achilles tendon length and then the subject contracts their calf muscle.  The tendon will then lengthen.  To measure stiffness you just keep contracting harder and harder and the tendon will lengthen more.  You then plot force (muscle contraction which is then converted to a joint torque/moment) versus tendon length.  The subject then slowly applies less contraction and you keep measuring tendon length.  This creates a hysteresis loop.  You have a stress-strain curve going up and then coming back down...Load and then Unload.  Below is a graph just showing the loading part of the curve and how that curve shifts to the left after strength training.  Meaning it gets stiffer. Acute stretching creates a shift in the curve to the right and the tendon is less stiff or more compliant (see Kubo) -

stretching and RT curves of stress strain.002
stretching and RT curves of stress strain.002

I think I'll do another post on how acute stretching or different contractions influence tendon stiffness but here is another good paper in the meantime (Kay et al 2015) but we here to talk about long term stretching.

What long term stretching does is amazing and counterintuitive.

I have shown graphically below in a slide which is a complete knock-off schematic of Kubo's research.

stretching and RT curves of stress strain.001
stretching and RT curves of stress strain.001

What do you see above?

- there is no change in the loading part of the curve.

- thus stretching did not change the stiffness part of that curve

- but there is a change in the slope of the curve when unloading (kind of like a concentric contraction after an eccentric contraction).

- notice how there is a loop between the loading and unloading curves? This means the unloading curve is less stiff meaning there was some loss in energy.  This is a quality of viscoelastic materials.  That energy was lost as heat.

- But what happens after long term stretching to that curve?  What happens to the heat loss?  Yup.  Less energy gets lost as heat after stretching!

- do you see the implications?

Crudely, the unloading part of the curve is STIFFER.  There was less energy lost as heat! Might this tendon be more efficient?  Might this small change be the mechanism for some of the research suggesting that long term stretching improves strength and performance!


What we can definitely say is that it doesn't appear harmful to performance.  We can't say that if you stretch your tendons are going to get loose and you will be sloppy and less energy efficient.  We don't have that evidence and that was something that I would have suggested in the past.  It seems pretty wrong to me.

You know what the other implications are? FASCIA. Seriously.  If you can't decrease tendon stiffness with long term stretching then I would wonder if it is really possible to change stiffer connective tissue like some fascia.  Tendons seem to be like springs that store and release energy.  And fascia (especially the ITB as argued by Carolyn Eng) seems to play a similar role for energy efficiency.  So the idea that our hands via manual therapy change this tissue  in the long term should be soundly questioned.  You certainly can't foam roll it into a greater length.  That's just silly.

OK.  Enough.  The point is to challenge our biases and that tendons are neat-o.  Things are often much simpler and more complicated then they seem.

What does stretching do to a joint? We really have no idea. Part I.


How's that title for a proclamation of ignorance? For someone who rarely stretches I sure love talking about it. I've been lecturing on stretching since the late 1990s and always loved to challenge my student's beliefs about what they thought stretching did.  The problem was, I didn't challenge my own biases for too long a time.

Ages ago I looked for research to confirm my bias that stretching wasn't necessary for injury prevention, performance or "postural corrections".  It was easy to find.  Ian Shrier was publishing work suggesting that stretching before competition did not reduce injury (Link here) and my favourite research discussed the mechanisms of stretching, which we can now elaborate on below.

Does stretching change stiffness

For more than a decade and a half I loved challenging students who said that stretching made a muscle less stiff.   While this is true immediately after a stretching bout we had some great research by Magnusson more than 18 years ago saying there was no change in the stiffness of the joint with long term stretching.  The implications were huge for postural correction folks.  It means you can't "loosen" up a muscle and effect some change in static posture because passive tension is unchanged.  Below is a simple graph that looks at how stiffness can change about a joint.  We create a stress-strain curve by moving a limb through its range of motion passively while measuring the force it takes to move the limb and the range of motion at the same time.  This creates a stress-strain curve.  The slope of the curve is its stiffness.  Some researchers also call this the Passive Resistance to Torque.  We can't say what structures contribute to the resistance to torque because we are just measuring joint ROM.  It could be capsular, muscle, tendon etc.  Researchers try to control for muscle activity by ensuring that EMG is silent.

In the curve below you can see that if a joint is less stiff the curve shifts to the right and if it is stiffer the curve shifts to the left (a left shift consistently happens with strength training).

Stress-strain curve
Stress-strain curve

The assumption for a long time was that long term static stretching shifted the curve to the right.  Because I was looking for ways to challenge the utility of stretching I found Magnusson's research showing that this just wasn't true.  What Magnusson showed as early as 1996/7 was that long term static stretching did not change muscle stiffness and therefore we could infer that there was no structural changes to the joint/muscle structures. I loved this contrarian shit.

Instead, what Magnusson described was an increase in muscle extensibility (joint ROM) that was proposed to be due to an increase in stretch tolerance.  Meaning if you stretch a lot, you get used to it and you can handle the discomfort and push yourself more.  Plenty of other researchers supported this finding and I was content for years to run with this idea.  I certainly did not go out of my way to challenge my own biases.  Below is a graph describing an increase in stretch tolerance.

stretch tolerance
stretch tolerance

I'd read whispers of papers challenging this idea but I didn't fully explore them...until this idea became so bloody POPULAR in 2010 with a review paper by Wepler and Magnusson.  Now everybody knew/thought that stretching created sensation changes rather than structural changes.  I'd lost my edge and had nothing special to add. :)  So it was time to really check my biases.  And sure enough there was plenty of research showing that long term static stretching did change the stress-strain curve.  What a blow to my scientific ego. Kubo (2002) showed this forever ago.  I remember reading his abstract and he wrote that stiffness did not change so I thought it was a done deal.  The problem was he described stiffness as related to a tendon not the entire joint complex.  What many called stiffness he referred to as Passive Resistance to Torque.  And guess what?  He showed that static stretching changed this - it changes joint stiffness in the long term.  And so did other research.  Take a look at the graph below.  This is what a shift in the passive resistance to torque looks like.  There are no less than 5 papers that document this shift.

a shift in the curve
a shift in the curve


Confused? You should be. Science sucks sometimes. We have well conducted research with very similar experimental protocols showing different results.  One group arguing that increases in ROM are purely neurally driven and the other suggesting there are actual tissue property changes.  I've switched sides of the fence personally.  I am not in the sensation only camp.  Both seem plausible and some work documents this change in the curve.  I don't know why there are differences in the outcomes across studies.

How is this relevant to anything?

I don't think this argues that stretching can change posture.  We aren't really seeing dramatic changes in the force that a muscle in its neutral position pulls on a bone to change a joint position.  It might change the kinematics we see when a joint gets away from the neutral zone though.  This would have relevance for the ease movement during end range positions.

In terms of performance many would say that we shouldn't stretch if we want to have a stiff and bouncy muscle.  This is often the argument for running.  We want a lot of passive energy return and stiffness is a good thing.  Well, I'm going to address that idea in Part II of this article and the potential conclusions will be SHOCKING!

Related blog posts:

1. What stretching does - this post is now outdated

2. Stretching is not a panacea and is neither good nor evil

3. Treating our patients like puppets: how postural correction is way overcooked.

4. Are you sure your hip flexors are tight?

5. Fascial science: stretching the power of manual therapy.

6. Foam rolling the crap out of your ITB.  Are you sure you are doing what you think you are doing?

Functional exercise is a poor term - how about some Comprehensive Capacity

originally posted at many years ago Here is an old post slightly reworked. Please read Eric Meira's post on functional exercise that is just fantastic.

A few years ago found myself defending the lowly clamshell exercise for runners. I was discussing with other physios whether the clamshell exercise was less "functional" than a band walk exercise (where you put elastic bands around your knees/ankles and walk forwards, sideways or backwards). I suggested that both were NOT (or equally) functional but agreed that both had their uses. I can tell you, I convinced no one :)

The consensus against me was that the clamshell sucked and that the band walk exercise was superior because it was more "functional" for runners. I tried to argue indeed the bandwalk might be a better exercise but not because the Bandwalk was more "functional".

The problem is the word "functional".

What the heck do you mean by 'functional".

When I heard someone say the bandwalk exercise was more "functional" than the clamshell I immediately thought malarky. Because my knee jerk reaction to hearing "functional" was my brain translating this to "movement specificity" or kinematic specificity. Meaning the exercise you are training matches the kinematics (specifically the displacements or joint angle motion) of the athletic task (in this case running).

Using kinematic/movement specificity to judge a Bandwalk or a Clamshell was a no brainer. In terms of function (using the kinematic specificity definition) they both suck. No runner lies on their side and lifts their leg up (e.g. the clamshell) but no runner runs sideways with a bloody elastic band around their knees.

So lets define "functional"

My knee jerk translation of "functional" to being "kinematically specific" is a little narrow.

Function to me means that the exercise has some sort of relevance to the task the athlete hopes to accomplish. Viewing it this way then exercises can be functional or relevant thru a number of different means:

1. Movement specific: this means the exercise task somehow looks like the athletic task it is trying to train. This means your exercise has similar form (due to motor control) to the goal task. Furthermore, it suggests that the neuromuscular recruitment pattern is similar to the exercises (e.g. muscle onsets, offsets, ratios etc) For example, squats are great functional exercise to get out of a chair.

2. Muscle or joint specific: this means the exercise is training similar muscles to the muscles that are being used in the goal task.

3. Velocity specific: this means if your goal task requires you to move fast than you should probably train fast. Exercises are therefore "functional" if they lead to some sort of carryover to the goal task's speed demands. We know that you don't actually have to move fast to get this carryover, sometimes just the intention to move fast will garner improvement

4. Movement direction: if your goal task requires a lot of deceleration than you should probably train the eccentric loading capability during your exercise. For a runner, you might think they have increased hip adduction during the impact phase of running which occurs for less than 100 milliseconds. Thus you should probably train this deceleration ability over that specific time frame for it to be functional.

5. Context: the exercise should be similar to the context of the goal task. The context might the exercises relationship to gravity or even to a societal or performance context (e.g you train to shoot free throws while people are screaming at you).

So what is more functional the lowly clamshell or Bandwalks?

Trick question! You can't answer this. There is no functional scoresheet. The better question is what is more useful to the runner or athlete. You have to be able to answer why you are prescribing an exercise in the first place. What is your intention with each exercise? What do you hope to accomplish? If you prescribe an exercise because it is "functional' than this is just begging the question. Functional isn't enough it has to lead to some specific gain. And if you think it is functional why is it beneficial?

A case example in the limits of functional justification: The BandWalk

I think most people would argue that the Bandwalk is more "functional" for the runner because it appears to train the muscles used in running in a more similar manner kinematically and more of the muscles that a runner uses when they are running.

But does it really do this? How much of the previous "functional" components does the band walk satisfy?

Movement Specificity: No runner runs with a band around their knee, they don't run sideways or backwards. On its surface it certainly isn't very similar in terms of joint angles and displacements. If you think movement specificity is important or practice makes perfect than training a movement that is so dissimilar to running can't be justified in this manner.

Movement Direction: Lateral band walks do not train the hip abductors in the manner that they are used during running. Eccentric control occurs over very short period under high impact loads when running. Band walks are extremely dissimilar to this. They are slow. They have an external force that is extremely different to the force vectors that create joint torques during running. We don't satisfy a movement specificity or even a velocity specificity argument. Considering this is the Bandwalk still functional?

Context? The band walk looks better than the clamshell that is for sure. The athlete is standing and shifting weight. But are they standing a shifting weight like a runner shifts weight? Nope. Is just standing and having weight shift enough to make it functional? Why not do dumbbell curls while standing and shift your weight back and forth.

Motor Control: No way. Doing bandwalks are nothing like the motor patterns used in running. Timing would be way off, no impact, no feedforward activation of muscles to damp vibration, hardly any storage of elastic energy. Do you think the "extensor paradox" occurs during a band walk. Not a chance. These two tasks are completely separate beasts. There is no way you can argue you are engraining some motor pattern during the band walk to enhance running.

Is the BandWalk Garbage?

Nope. It can certainly help runners. I rarely use it but I still think it is reasonable for a runner to train with it. Because it satisfies our simplest category of "function". It trains a group of muscles that runners need. These exercises build the capacity of those muscles and this probably transfers over to increased mechanical efficiency and maybe even some injury protection. However, if you think that it is functional enough to actually change a runners form than you might end up getting some poor results.

The problem with this last justification for strength coaches and rehab people is that it is too simple.

As therapists we want to think that we have some special knowledge about special exercises. We don't. There are no special exercises. Train hard, train smart, get strong, build power, build tolerance, build capacity, build endurance, build the ability to absorb load, dampen vibration, produce strength at all ranges etc. Have a generalized, good all around program and you will probably have good results. Maybe throw in some assessments to see where your athlete is lacking and then train the hell out of that and you will do better. But, there are no running exercises. And that is contrary to what both the research suggests and gurus suggest. Pick an expert in a field of strength and conditioning (powerlifting, olympic lifts, corrective exercise, pilates, yoga, core stabilizinationists etc) and they all swear by their success with runners. They are all probably right - thus no specific running exercises.

OK, so is the clamshell better?

No way! The clamshell sucks most of the time. I used to hate clamshells because I thought I was better than that. I thought runners should never do it. I thought it was remedial and "non-functional".

The clamshell works fewer muscles, has the person lying down, looks kinematically different than running and doesn't satisfy many of the functional principles I laid out earlier. BUT, it does train some muscles substantially different than the bandwalk. This is where it can be a useful exercise.

The clamshell sees the hip flex to 90 degrees and has the patient externally rotate the hip. At 90 degrees, because of changes in the line of pull of many hip muscles (GMax, GMed, Piriformis) the only muscles that externally rotate the hip are the deep external rotators. So training the capacity of these muscles might carryover to running.

I used to abhor the clamshell. Then I started testing more runners with the clamshell. A number who tested strong in many positions would tremble during the clamshell. Crazy, they had a lovely one leg squat, strong hip abduction but had trouble with 10 or 15 clamshells. What does that tell me? Such a massive deficit in function. Would you suggest clamshells here or something to address that specific movement? This seems like a case where I would suggest clamshells. If a runner can't do them I would want to address that deficit. I sure can't say if it is problem but if an athlete can't do something so simple, in a muscle group relevant to performance, then perhaps I should address it.

But, do I want to see every runner doing them as part of a "functional" program. Of course not. They suck for that. This is a case a where the exercise prescription is "functional" because it addresses a specific limitation in a specific runner.

OK. What do I suggest?

I have no exercise that addresses all of the components of "Functional" but that's why runners should get a comprehensive program.

If I have a bias I lean to training "Comprehensive Capacity". This means you train runners like they are athletes. Big multijoint-compound exercises that train strength, high load power (e.g cleans), low-load power (plyometrics), variable range (not just mid range but end range and weird stuff) exercises and unilateral exercises. The thing with the word "functional" is that is so broad to be meaningless to justify an exercise. Choose exercises based on some other specific capability of the neuromuscular system you hope to improve.

An aside: If you train "comprehensive capacity" and have a well rounded training program do you really need to do in depth assessments? Don't your exercises become your assessments? Or your exercise regime is so well rounded you address everything an assessment might tell you.

The Key Question: Why are you training what you are training?

For runners, I only think that we are training the muscle, joint, tendon and nervous system's capacity to tolerate stress that running imposes. This makes runners more powerful, mechanically efficient and may make them less injury prone.

Do exercises help change running form?

This is seems to be the underlying idea behind an exercise that kind of looks like running (eg. the clamshell). That training in that manner will improve your running form. But there is some evidence to suggest that this does not occur (Willy 2011). We probably aren't training form. We aren't correcting the biomechanics of running by choosing certain exercises. We aren't engraining some motor programs that carryover to running. If you want to do any of these things you have to do that while running and with some form of feedback. Our bodies are not puppets where muscles can be tightened or loosened to obtain some different posture or form. That is a motor control skill not something that changes with other exercises.

So why are you prescribing the exercises you prescribe for your runners?

Please let me know what other rationales are out there. I don't think this brief article really addresses everything




Related Posts:

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

2. Can we treat our patients like puppets? Changing posture through exercise?

3. Running Injury Prevention: What we know and more of what we don't

4. Runner's Strength: Some simple exercise videos for runners.

Treatment Fundamentals: a simple framework to reconceptualize pain and injury treatment

The following article was also published at

Medbridge Education.

During our roundtable discussion on Pain:

Where Does Biomechanics Fit on MedBridge

, a good question came up: “How do you keep up with the literature to guide your practice?” The daily volume of publications can make it difficult. My solution is to regularly refer to my fundamentals of treatment, which stay the same, and occasionally “redecorate” them with new ideas. This way, new research complements my fundamentals and rarely throws me for a loop. For your own practice, I’d recommend writing down what you consider to be your fundamentals of treatment. What do you hope to accomplish within your treatment session? What are your views of the body? What can you do to affect your patient?

My Fundamentals of Treatment (aka Axioms of Function)

1. Rule out red flags

When dealing with pain, I want to be sure that pain is the primary problem. Pain - secondary to cancer, autoimmune disorders, infections or anything sinister - must be dealt with appropriately.

2. Rule out serious tissue pathology

Although hurt doesn’t equal harm, sometimes harm equals hurt. I like the biopsychosocial approach to treating pain and injury, as it recognizes the importance of bio. Dealing with a runner who has experienced months of anterior hip and groin pain, I can’t automatically assume that they have chronic pain and a sensitized nervous system. They could have a stress fracture of the femoral neck, a tissue pathology to be addressed. Yet, recognizing the bio doesn’t mean freaking out about rotator cuffs tears, hip labral tears, degeneration, etc. – we know they can co-exist in healthy, pain-free individuals.

3. The body is strong and adaptable

After ruling out the pathologies, I can view the body positively. Regardless of pain levels, I can tell the patient that their tissue is strong and start treating the whole person. I want to change their beliefs about their body and convince them in the above axiom. The human body isn’t a stack of blocks that will fall apart if there’s something slightly off.

4. Pain is more about sensitivity than about injury

In my treatment, I use every opportunity to show that pain is modifiable, to prove that it's more about sensitivity than damage. The patient learns that their connective tissue is robust and their pain is due to a sensitive ecosystem meant to protect them. The pain often has little to do with their structure or strength; it’s more of an allergic response by the body. People don't die because of a bee sting, but because of their allergic reaction to it. Pain is the same way. You can remove the stinger (i.e. heal the tissue) but still be left with the protective responses driving the dysfunction. When you modulate a patient’s pain in 5 minutes and contrast it with the obvious fact that they didn’t heal in 5 minutes, these ideas start to click.

5. Treatment is about finding the appropriate stressor/load

I put stress on tissues because I know those tissues will adapt. But the body is an ecosystem, and there will be responses to that stress elsewhere. I have no idea which components of the biopsychosocial model I'm influencing with an exercise program, but I assume all might be involved and try to maximize it. Therapeutic neuroscience education is a stressor (see

Adriaan Louw’s great course

). We're not teaching patients to pass a test, but to actually change their opinion of their body and ecosystem. Do you think a spine stability program really needs to change spine stability to influence pain? Of course not. If proposed in the right manner, it can change a patient’s views on their strength, adaptability of their body, and malleability of their pain. I think exercise is a BPS intervention. I think the tissue is important, but I don’t fully know how it influences pain. Exercise is a great psychosocial intervention if we frame it so. We're challenging (a form of stress) the patient’s view of their body. If patients view themselves as weak, the pain-free exercises help them to change opinions about their body. As patients progress, I might even have them poke the bear – confront the pain. They learn that it can hurt - and they modify the movement to change the sensitivity – but then they persist with the activity with no flare up and no harm.

We find the right stress at the right time. This can work the other way, too. We challenge the patient’s beliefs about pain and injury, influencing how they experience pain and how they choose to move. Suddenly, you’ve changed their beliefs and they pass a spine stability test.

6. The patient is an active participant in their own care

My job is to do nothing that makes the patient rely on me. I don’t tell patients that they have scar tissue, adhesions that need breaking up, shifted ribs that need correcting, or anything that requires outside help – it would contradict the axiom that the body is strong, robust and adaptable. I can’t have patients thinking they’re an inherently unstable stack of blocks. So those are my fundamentals. Then, I add decorations.

Decorations: Useful Though Not Fundamental Axioms

1. Gauge your treatments by assessing sensitivity

Treatment at its core is about desensitizing and then building back up. We find aggravating variables and modify them (e.g. a squat hurts the knees so we change the biomechanics to unload the knees temporarily) but at the same time, we also build the tolerance to the offending activity. Some patients are so sensitive that we need to resort to imagined movements (

Graded Motor Imagery

), but then we still increase the load incrementally (Graded Exposure - for a great review see


2. Manual therapy is an adjunct to fundamentals

I use manual therapy to prove to patients that their pain is about sensitivity. Manual therapy isn’t about correcting anything, although it may help in the short run. It’s about changing what they feel and helping to believe in their adaptive potential. I often fake manual therapy. For example, I might “correct” the scapula with a scapular assistance test and then have the patient lift their arm – and it hurts less. After 1-2 repetitions, instead of pushing the scapula, I just twist the skin – still less pain. Then, I twist the skin in another direction - still less pain. Then, as they keep lifting, I stop “correcting” completely – still less pain! I explain that there is no way I’ve corrected their scapular motion. Rather, I’ve changed how they feel - their sensitivity. Perhaps they became more confident, less fearful. Perhaps something happened in the brain (likely, but I don’t always get into it). I use that change as a learning tool and then use exercises to reinforce the new perception (a nice discussion can be found



3. Your assessment reinforces their belief in strength

Patients think they are falling apart. They think they are tight and in need of correcting. I confront those beliefs in my assessment. Patients often believe they have tight hip flexors and weak glutes. I wonder where they got that pernicious and pervasive idea? Often, their hips extend equally on both sides, yet they only have pain on one side. I ask how can this be if their tight hips are the problem. I point out they don’t fall over when they walk or run, so how can their glutes be so weak as to cause problems? If they feel that they have pain because their core is weak, I ask, "How can that be?" A stable core takes less than 10% of maximum contraction of the anterior abs, even during a loaded barbells squat, and most of their pain is in sitting. These confrontations help change how patients view their body.

4. Comprehensive capacity trumps assessment-driven correctives

I wrestle with the relevance of regional interdependence. It seems logical and right, but only if I view the body as a machine or a structure, and not as this complicated ecosystem. I agree that hip movement will change knee movement. I agree that a thoracic kyphosis will change scapular movement and thus affect shoulder flexion. I question its relevance to pain. Hip strengthening exercises can be great to treat knee pain (review


), but so can the old boring knee extension or knee strengthening exercises (a cherry-picked paper


, but you can follow the related articles to get the point) and hip strength may not precede knee pain (review


). And guess what! The hip strengthening exercises don’t even have to change hip kinematics to improve pain (paper


, but ignore the title - there were no changes in kinematics that most people think are important for knee pain e.g. knee valgus). So, the hip can influence the knee, but it doesn’t have to influence it mechanically to alter pain and function.  That regional interdependent idea can certainly be questioned when it comes to biomechanical changes that are “driven up the chain” from the foot - this is especially evident in the transverse plane (see



My solution/alternative is

comprehensive capacity,

my term for a shotgun approach: I’m no sniper, I am a bumbling hunter.  This approach addresses the uncertainty we should all recognize.   We don’t have to just find “flaws” in how someone moves. We can look at the painful joint, at the joints above and below, and say, “I want you to be able to do everything!" I want strength, control in all ranges of motion, a variety of movement options, speed, agility, the ability to bounce and “pop” - all done fearlessly and confidently. After all, if I’m really honest, none of us knows how these mechanical and psychosocial variables interact. If you have an injured athlete, make them the best athlete and healthiest person they can be. That’s good rehab.

5. Postural and movement assessments reveal habits but not flaws

The dominant view is that when joints deviate from a neutral position during rest or physical activity, there’s a risk for pain. If a person has an anterior tilted pelvis, hip adduction and weak glutes, the default is to blame this abnormality for their back, knee, foot or shoulder pain. Those movement patterns can be relevant, but I question why they should be the default. If no previous research or ideas existed, would we still hold this view? Consider how common and varied these movements are in the pain-free population, how incredibly robust and adaptable the body is, how pain can contribute to these movements. Should we still default to these patterns as the cause for the pain? Will they become the new structural "damage/degeneration" bogeymen we dismiss on radiographs?

I might change the habits of movement, but only to give patients variety and uncouple those movements from pain. It’s a temporary fix, akin to a bandage for a cut or a 6-week heel lift for plantar fasciopathy. But I don’t want any patient to fear any movement. BUT, I recognize the possibility for exceptions...If the research makes a strong case that certain movement patterns in certain populations under certain conditions are related to injury (e.g. increased knee abduction moment during landing in females in sports with high ACL prevalence), then I take that research and redecorate my structure. But do I freak out when every girl who runs 2 km twice a week shows up on my treadmill with some knee valgus? No. (More on this



So those are my fundamentals. What are yours? Should I be redecorating?

Smudging pain neurotags and cortical body maps: explaining the weirdness of pain

Warning: this post is long Audience: health care providers

I use social media to help organize my thoughts and challenge what I think.  I usually ask a lot of questions because that has always been my style.  I tend to challenge those who are the most similar to me (even if I agree with them) because it helps me “know what I know”.

I recently posted a series of questions on Facebook about Pain Neurotags (the collection of brain cells that are activated throughout the brain to produce pain) and Cortical Body Maps (the regions of the brain that correspond to the physical body parts or to the movements that the body performs).


Below is a summary of the discussion and some relevant reading.

Relevant Reading

  1. Graded Motor Imagery Handbook (
  2. Explain Pain by Butler and Moseley
  3. How to Improve Proprioception (a great piece on Cortical Body Maps by Todd Hargrove at

I initially asked…

1. How do you see cortical body maps being similar to or different from neurotags? Specifically pain tags.  Would you view a sensory body representation as being part of a pain neurotag?

2. With pain neurotag smudging (disinhibition of other neurotags) would part of that be due to smudging of the sensory body map and some due to disinhibition of other neurotags?

Bold and italics are mine. I have cut out some information but by and large here is what everyone contributed with a few recaps and extra notes thrown in.

Sigurd Mikelson gave us...

Puh. Now there's a challenge... I'll have to do some thinking and come back to this, but for starters I think of cortical body maps as a more neuroanatomical construct than neurotags which I understand as functional synaptic networks. The term neurotag is also somewhat ambiguous as its "only" a construct of "some kind of activity" that we potentially want to see. The term (in lack of anything better) still serves as a subjective conceptualization of "something going on in there" that is thought to be associated with pain" and helpful in formulating dynamic, clinical frameworks. On the other hand one should be careful in taking them as "true", not that I'm implying you doing so. These considerations about terminology also goes for what is really a "sensory body representation", but yes, whatever that is, I guess it would also be part of what a pain neurotag is, whatever that is... In regards to your last question too, I think disinhibition of functional networks of "brain activity" is the least wrong explanation so far about what might be going on in ongoing pain states and behavioral "disorders".

Intelligence concept

Lars Avemarie provided a great definition from the GMI Handbook

"A neurosignature (or we call it a neurotag, like a graffiti tag1) is a pattern of activity in the neuromatrix. It is a physical linking of neurones at a particular time and it canlead to an output such as pain, movement or an emotion. We are not conscious of most of these brain constructions and their outputs, but rest assured that it will be a pain neurosignature, absolutely personalised for you that will be the basis behind you experience of pain (Figure 4.3)."


The Graded Motor Imagery Handbook. G. Lorimer MoseleyDavid S. Butler, Timothy B. Beames, Thomas J. Giles. Noigroup Publications, Adelaide, Australia, 2012

In specific reference to the question “Would you view a sensory body representation as being part of a pain neurotag? " Lars responded:

Yes, a sensory body representation can be part of a neurosignature. To my current knowledge.

Ben Cormack (at asked the clarifying question…

...could we say we may use selected neurones within our cortical maps when we activate a pain tag/signature?

and Sigurd responded:

from my understanding, these "selected neurons", if you talk about neurons located around/in S1, S1, thalamus, insula, ACC are multi-modal and will therefore respond to any stimuli regardless of whether the stimuli is nociceptive or non-nociceptive. Ianetti and Moreaux have done some important work on this.

Whatever gives meaning to the stimuli would partially determine the threat value and emergence of a potential painful experience (or continuous state when thinking of longstanding pain), which would also involve all of the above. The problem is that we don't have solid evidence to suggest that synapses/neurons give meaning to anything, or at least, not enough to formulate/prove a theory that will hold under any condition.

Nevertheless, the neuromatrix-theory binds together these diffuse bits of "knowledge" into a dynamic framework useful for clinical practice. Or, maybe it should be understood as a dynamic model, rather than a theory, like the bio-psycho-social model (which rather should be practiced as a socio-psycho-biological model). Maybe. There are epistemological challenges with both.

Quick Recap

Our first three contestants have put forward the idea that neurotags are activations of different brain cells, from different parts of the brain that can produce pain, movement or emotions.  Within a neurotag can be our “cortical body representations” or our body maps.  The sentiment is that our body maps are more neuroanatomical constructs and the neurosignature is a functional creation that would include activation of body maps as well as the activation of other brain cells.

We have also introduced the idea that a Pain Neurotag can be activated or modulated by other factors than mere information from the body’s booboos. Siggurd has noted the importance of meaning in the creation of the pain experience.  This is important because we should view the brain as evaluating all the information it has in relation to the “danger” signals from the body and from other information it possesses.  Ultimately, if the brain feels sufficiently threatened pain will be felt.

A new point: There are multiple areas of the brain involved in producing pain.  Butler and Moseley call these Ignition Nodes. These nodes are used for:

  • Sensation
  • Movement
  • Emotions
  • Memory

Butler and Moseley write “Pain just uses this parts to express itself”

With persistent pain these nodes can become sensitized and more easily activated. This means that multiple factors besides the booboo in your back can cause pain.

Phillip Snell then added…

I very much like Sigurd's notion of the neuromatrix-theory as a dynamic clinical model. I find that as I consider your post Gregory I'm tempted to frame neurotags as memories and cortical body maps as an overall sense of self. In that construct, I see that those neurotags are subsets that help to define the overall sense of self that the cortical body matrix represents. Memory and sense of self are plastic and each influence the other. Good topic to sit with this AM!

Jason Silvernail stepped in and essentially said he would take this topic to the highest authority...Lorimer Moseley :)

“Good question Greg and great comments Sigurd. I think at this point we don't know enough for sure to be able to say one way or another. Greg your suggestion makes sense as part of a framework.

All frameworks are wrong, but some are useful. I think the bigger question is how can we make these concepts useful - in helping practitioners and patients understand, in formulating hypotheses to test, in developing treatment approaches, etc.

In general I have the same overall sense of the interrelation between cortical maps and neurotags as Phillip does - but that division probably reflects my own conceptual model rather than any underlying truth. Would love to get some neuroscientists in a room and ask them. Hey maybe we can ask Lorimer Moseley in San Diego in February…”

I then produced a screen shot from Pg 76 of Explain Pain that describes the phenomenon of ‘smudging of the neurotag”.  The relevant quotation from that page being:

another change which is known to occur in the outer brain, the cortex, is “smudging” - brain areas normally devoted to different body parts or different functions, start to overlap. In fact, the longer pain persists, the more advanced the changes int he brain. We think both types of changes might be strategies by which the brain ‘looks out’ for you - by making the body part difficult to use (smudging of the motor areas in the brain, thus limiting movement) or by making neraby body parts sensitive too (smudging of sensory areas in the brain).

Todd Hargrove jumped in with an answer that I tended to agree with.  I agree a lot with Todd thats why I argue so much with him.  Todd wrote:

Thanks for including me. Why are you asking?

I only browsed the answers above and saw many great points. Here are my answers, which duplicate a lot of what has already been said.

“How do you see cortical body maps being similar to or different from neurotags?”

The cortical maps are anatomically discrete areas of the brain whose activity represents the movement, position and health of the body parts. Neurotags are the patterns of brain activity that create a particular experience of pain. These patterns are widespread throughout the brain, and include activity in the cortical maps. So the neurotag is far broader than the maps.

“Would you view a sensory body representation as being part of a pain neurotag?”

Yes, for the reasons stated above.

“With pain neurotag smudging (disinhibition of other neurotags) would part of that be due to smudging of the sensory body map and some due to disinhibition of other neurotags.”

Personally I never think of pain neurotags as being smudged, as their purpose is not to represent or map the body, but to create an action signal. I think of the cortical maps as being smudged, because when they are smudged they inaccurately map the body, just as smudged map would fail to clarify the outlines of a geographic area. But I do think of the pain neurotags as being disinhibited, or as being activated by lots of stimuli that we don’t want to create pain, such as thinking about herniated discs, arriving at a workplace, or doing a movement that at one time caused nociception but no longer does. I think the easiest way to understand and describe this problem is not by talking about neurotags or disinhibition, but by reference to Pavlovian learning through association.

For example, if I want to explain to someone why movement can help with their pain, I might say that they learned through experience to associate movement with pain, because of a previous injury. The pain continues even though the injury is gone, just as a dog will salivate when a bell is rung, after it learns to associate the bell with dinner. The dog will eventually stop salivating if you ring the bell without bringing dinner, and you will eventually stop feeling pain if you do that movement you have been avoiding without causing reinjury. I wouldn't say anything about disinhibition or neurotags, not because those ideas are wrong, but because these concepts are unnecessarily complex in my opinion, and as Jason said, they are all in service of good communication and understanding, and should be judged on that basis.

Another quick recap…

Todd highlighted the previous views that the cortical body maps are anatomical representations of the body and that Neurotags are Functional/dynamic “patterns of brain activity” that create some experience.  A Neurotag is again viewed as something that can encompass a cortical body map.

Todd and Jason discuss the utility of using Neurotags in explaining pain which I think is a very fair point.  Todd gives a great example of how we can discuss pain becoming associated with movement (an example of the pain neurotag become more easily activated with other stimuli) without actually ever discussing pain neurotags with a patient.

Todd also provides us with another great reference from Lorimer Moseley on Neurotags and memory.

and Jo Nijs work on altering pain memories through exercise

Ben further added...I like the concept of memories that we then create predictions from. This seems to make sense.

Somewhere down the line if we alter the memories we alter the prediction. I think this has elements of associated learning in it and why I identified with the paper Todd Hargrove reposted on the thread.

I then added:

Thanks Todd Hargrove that's a great response and how I generally view it (but less eloquently put). I got the neurotag smudging from Moseley. See my above pic. I tend to view neurotags as something that would contain cortical body maps as you described.

I suppose smudging is just another way to describe disinhibition of other neurotags. I'm not sure it's the pain neurotags themselves that become disinhibited though. I think it's other neurotags that become disinhibited. Thus they fire with the initial pain tag. Hence we have spreading and diffuse and moving pain

Lars adds:

Lars Avemarie Todd Hargrove Great comments, specially this "Personally I never think of pain neurotags as being smudged, as their purpose is not to represent or map the body, but to create an action signal. I think of the cortical maps as being smudged, because when they are smudged they inaccurately map the body, just as smudged map would fail to clarify the outlines of a geographic area. "

This is the same way I view it currently.

Moseley talkes about this here:

“One aspect of the changes that occur when pain persists is that the proprioceptive representation of the painful body part in primary sensory cortex changes. This may have implications for motor control because these representations are the maps that the brain uses to plan and execute movement. If the map of a body part becomes inaccurate, then motor control may be compromised – it is known that experimental disruption of cortical proprioceptive maps disrupts motor planning”


Moseley, G. Lorimer. Reconceptualising pain according to modern pain science. Physical Therapy Reviews 2007; 12: 169–178.

And Hodges here:

“First, chronic pain is commonly associated with reduced tactile sensitivity on the painful body part [70,71,133] and disorganisation of somatotopic maps in SI [71,134–138]. Further, the magnitude of somatosensory disorganisation is a strong predictor of pain severity [134,135,139]. Conversely, tactile discrimination training, which should promote organisation of somatosensory maps, is an effective treatment for chronic pain, presumably by sharpening somatotopic maps of the body in primary somatosensory cortex [140,141]. Intriguingly, the effectiveness of tactile discrimination training increases when CRPS patients view the body part being trained [141].”


Spine (Phila Pa 1976). 2011 Oct 1;36(21):1721-7. doi: 10.1097/BRS.0b013e31821c4267.ISSLS prize winner: Smudging the motor brain in young adults with recurrent low back pain. Tsao H1, Danneels LA, Hodges PW

With Todd adding:

I think that pain neurotags can be activated by other neurotags, as you say. For example, the neurotag for thinking about work activates the pain neurotag. But does this happen through disinhibition (a mistake) or through learning (neurons that fire together wire together)?

Hopefully the latter because the remedy for unlearning is more clear to me than the remedy for improving inhibition, which might be systemic or genetic. I understand that many intractable problems like migraine, epilepsy, etc. are characterized by failures in inhibition.

By the way, you might be interested in this post, which has four or five paragraphs summarizing what Moseley had to say at a conference about neurotags, disinhibition and maps, including lots of very concrete examples. Still very confusing, but I think it supports my idea that correctable disinhibition is more in the cortical maps than in unwanted connections between pain and non-pain neurotags.)

Its OK to not know what to think of these things

Lars then quotes a great article:

"Chronic pain is impossible to ignore, and commonly dominates the patient’s mental life. Ironically, however, there is also evidence that chronic pain is associated with ‘neglect-like’ symptoms [151–153] or ‘body perception disturbance’ [154,155], in which the affected limb is misperceived. These symptoms reflect a constellation of symptoms with intriguing similarities to disorders of body representation that follow right parietal lobe damage [156]. Some patients report that their affected limb feels ‘‘like dead weight’’ and that focused attention is required to move the limb, while others report feeling that the painful limb feels ‘‘foreign’’ or ‘‘strange’’, as if it were not part of the patient’s body [152–154].

In one large survey of CRPS patients, 84% of patients re-ported at least one such neglect-like symptom [152]. Such reports are intriguingly similar to previous reports of ‘asomatagnosia’ following parietal lobe damage, in which patients report feeling like the contralateral side of their body is absent [156]. In some cases, these feelings of foreignness result in hostility towards the limb (‘misoplegia’) [154,157] and even desire to have the limb amputated [154,158]. In other cases, there is clear evidence for sensory abnormalities related to the affected body part, including unawareness of limb position [155,157], referral of sensations to adjacent body parts [139,159], and displacement of the perceived body midline towards the affected side [160].

In addition to the neglect-like symptoms caused by chronic pain, visual hemi-neglect itself also alters pain perception. In particular, Liu and colleagues [161] observed a mislocalization of painful stimuli to the ipsilesional size of the body, and a misidentification of stimulus modality in neglect patients."

"Our review of the neural mechanisms that integrate pain and body representation has several implications for the search for multisensory therapies. First, this search should not be abandoned as hopeless, since there is clear evidence for neurophysiological mechanisms for multisensory interactions involving the nociceptive system, on which such therapies should rely. Second, there is a need for future research to use multiple sensory modalities, and not only vision, as potential interactors with pain. Third, multisensory interactions involving nociception reveal a strong principle of spatial organization. Since chronic pain is often very specifically localized to a single body part, and since reorganization of spatially-mapped cortical areas appears relevant to chronic pain, future research might usefully investigate how to harness the spatial organization of body-pain interactions in order to modulate chronic pain."


Curr Biol. 2013 Feb 18;23(4):R164-76. Spatial sensory organization and body representation in pain perception.Haggard P1, Iannetti GD, Longo MR.

Wrapping Up

And that was basically it with the discussion.   Back to the original questions:

How do you see cortical body maps being similar to or different from neurotags? Specifically pain tags.  Would you view a sensory body representation as being part of a pain neurotag?

The consensus was that cortical body maps represent discrete anatomical or movements in the cortex and that Pain Neurotags are probably something larger that can contain those body maps.  From reviewing Moseley’s writings in the GMI handbook he views those bodymaps as neurotags as well.  This makes sense since those are still parts of the brain that would become activated when moving or sensing a body part.

Moseley uses this definition:

“A neurotag is a network of interconnected neurones, we will call brain cells, that are distributed throughout the brain. When a neurotag is activated it produces an output. The output defines the neurotag”

And goes on to write:

What is important conceptually is that to activate any particular neurotag, two criteria have to be met:

  1. The member brain cells have to fire. That is, the neurones that make up the neurotag have to fire.
  2. Nearby brain cells have to NOT fire. If non-member brain cells fire, the neurotag is changed, is imprecise, is wrong if you like.

An important point is that each member brain cell of the neurotag is open to being influenced by other brain cells which is communicates.  This is how other variables like thoughts, memories, movements and emotions can influence a pain neurotag

Moseley goes on to write:

When pain persists, the pain neurotag becomes sensitised and disinhibited. Sensitisation of the pain neurotag refers to an increase in the excitability of the member brain cells such that they are more easily activated. The principle of this increase in excitability is the same as that of individual neurones that become potentiated, or ‘wound-up’. Sensitisation of the pain neurotag is like the whole network of brain cells being ‘wound-up’. Sensitisation of the pain neurotag offers the most sensible explanation as to why, as pain persists, pain is evoked more easily and by a wider array of internal and external stimuli than it is initially.

Moseley describes the smudging of a neurotag:

Remember that for a neurotag to be activated, it requires sufficient activation of the member brain cells and sufficient inhibition of the non-member brain cells? Well, disinhibition refers to a decrease in this inhibition of non-member brain cells. This sounds complex but it just means that they’re not dampened down. Perhaps an easier way to think about it is to say that the neurotags lose precision. This effect probably only involves neurotags that are relevant to the pain neurotag. Disinhibition might manifest as spreading pain, pain that moves, pain that is less precisely defined anatomically or qualitatively. The pattern of spread will not adhere to the distribution of a peripheral nerve, or to that of a nerve root. Instead, pain will spread to a whole limb, a body region or to a whole side of the body. Disinhibition of movement neurotags will manifest as imprecise movements or perhaps in the extreme, dystonia.

...and to the last part of the questions

With pain neurotag smudging (disinhibition of other neurotags) would part of that be due to smudging of the sensory body map and some due to disinhibition of other neurotags?

I’m not sure this was solidly answered.  Everyone can agree that cortical body maps become disinhibited/smudged and this can contributes to influencing the pain neurotag.  I’ve found less discussion in the literature or in our discussion on whether there is “disinhibition of other neurotags” besides the body maps.

Moseley writes:

Disinhibition is the most likely explanation for the rangeof body-related neurotags that can become disrupted as pain persists. The most investigated of these involves an area of the brain called the primary sensory cortex, or S1.

Todd suggested that it might be more useful to think less about a pain neurotag becoming disinhibited and more that that neurotag becomes coupled with some other neurotag via associated learning.  Todd rightly notes that talking about this association with pain and movement is a great way to unlink pain from movement in patients.  And perhaps we don’t even have to talk about smudging at all with patients if we just talk about learning.  The key here is that other neurotags can more easily activate the pain neurotag.  I don’t know, and I don’t think we resolved, if this coupling is due to disinhibition or some other mechanism like learning.  I’m not sure it matters in order to explain the symptoms that patients feel or to help guide treatment.

Thanks to everyone for helping out with this.