Motor learning: the coordination myth and 5 reflexes that slow progress

1/4h LabO #252, Watch the episode on YouTube

Five phrases come back every day in the field, they sound rigorous and caring, and yet contemporary motor learning science has been taking them apart one by one for 60 years. A single number is enough to measure the gap: a baby learning to walk falls seventeen times an hour, and it is precisely this rate that produces their astonishing speed of learning. And what if what we do to help, to secure, correct, repeat, explain, were exactly what blocks our athletes' progress?

What is motor learning? The brain learns from the gap

One mechanism comes back with every myth, so let's set it down first. The brain learns through prediction error. With every movement, it anticipates an outcome. If what it predicted matches exactly what happens, it has no internal reason to recalibrate. It is the gap between forecast and reality that feeds cerebellar learning. No gap, no learning. It's that simple.

This principle has been a solid consensus since the work of Karl Friston and Daniel Wolpert. And the consequence for your practice is direct: everything that removes the gap, identical repetition, the imposed model movement, the sanitized context, the avoided error, removes at the same time the engine of progress.

Keep a second pillar close at hand as well: motor learning is very largely implicit. It is built without verbalization, or nearly. The baby learns walking, running, fine grasping without the slightest verbal instruction or tutorial. This grid, the gap on one side, the implicit on the other, is enough to understand why the five reflexes that follow turn against progress.

Myth 1: "To learn well, you have to repeat"

You know the argument by heart. You repeat, you engrave the movement, it becomes automatic, and the more you repeat, the more solid it gets. Honestly, it isn't wrong. Repetition is necessary for motor learning. The baby repeats tens of thousands of times before walking, the expert accumulates thousands of hours. Nobody disputes that. Quantity, for that matter, is rarely what is missing. Where it gets stuck is the form of the repetition.

There is a massive literature on this since the early 1990s, starting from Schmidt and Bjork in the United States. The verdict never changes: identical repetition gives you flawless immediate performance in the session, and lasting retention lower than what varied repetition offers. You polish something that collapses as soon as conditions shift a little.

To grasp why, imagine a forest to cross. First option: you take the same path a thousand times. You know every stone, every root, you race along that path at full speed. Second option: at each pass, you vary, sometimes more to the right, sometimes through a clearing, sometimes around a rock. A thousand times as well, never the same route. The day it rains, a tree falls across the way, the light changes, who gets through? The first one is lost, their single path has just become unusable. The second improvises, because they know the forest and not just one path.

The motor system works the same way. When the athlete redoes the same movement under the same conditions, their brain has no gap to correct between what it predicted and what happened. Yet it is precisely this gap, the prediction error, that feeds cerebellar learning. No gap, no learning.

The field rule fits in one line. If your athlete does the same movement identically twice, change a parameter: the surface, the angle, the speed, a cognitive constraint, a signal that appears at an unexpected moment. It doesn't matter which one, but change something. A thousand varied repetitions will always beat five thousand identical ones, because you are building knowledge of the forest and not of a path.

Myth 2: "There is a reference movement for every skill"

This one dies hard. The reasoning seems unbeatable: look at the champions, they all do roughly the same thing, so there must be a reference movement, so the coach's job comes down to passing on that model and correcting deviations. Except that the idea is wrong on two levels.

The first level is theoretical. Bernstein in the 1940s, then Newell in the 1980s, set out the constraints model. Every movement emerges from the interaction between three poles: the task, the environment and the individual. And individual constraints, anthropometry, motor history, specific capacities, vary enormously from one athlete to another. So there is no optimal movement valid for everyone, only an optimal solution for this body, in this context.

The second level is empirical. When you really look, with fine kinematic analysis, at what champions do, you discover that two of the best servers in tennis don't have exactly the same serve, that two of the best goal scorers in football don't have exactly the same shot, that two gymnasts at the top don't have exactly the same rotation. What looks alike to the eye hides fine differences, and those differences are precisely the adaptation to individual constraints.

To feel what this changes in practice, take the tailor. You show them a photo of Federer in a suit and tell them: I want exactly this suit. The good tailor answers: fine, but his shoulders measure 50 cm and yours 44, his inseam is 88 and yours 80, so I'll make you the same silhouette, cut for you. If they insist on literally reproducing Federer's suit at Federer's dimensions, you'll be cramped, you'll have creases in the wrong places, and you'll end up tearing it as you move.

This is exactly what happens in biomechanics when a single technical model is imposed. The athlete plateaus, because the solution isn't adapted to their constraints. They tire needlessly, because their body compensates constantly. And their injury risk climbs, because you load structures that weren't made for it. The coach's role is not to impose Federer's suit. They pose the problem: here is the target, here is the situation, here is the constraint. And it is the athlete who invents their motor solution, made to their measure.

Myth 3: "Isolate the movement to work on it better"

This reflex puts forward an argument that sounds downright scientific. Analytical isolation is rigorous method: you isolate a variable to work on it better, like physicists, chemists, even biologists. Applied to sport, it gives: isolate the movement, perfect it out of context, then reintegrate it into the game. True for physics, true for chemistry. False for human motor control.

The image that sums it all up is learning to swim. Imagine telling someone: first we'll work on pure crawl technique, lie on this bench, I'll place your arms in the right position, you do the movements perfectly, and when you've mastered the technique on the bench, only then will we go into the water. The person can refine their arm movements for hours. The day they jump into the water, they discover that what they learned has almost nothing to do with swimming. Because in water there is buoyancy, resistance, breathing rhythm, body orientation, coordination between the legs and regulation of the breath. These couplings are precisely what make the movement.

Motor control works as an integrated system, where perception, prediction and action are structurally coupled. As soon as you isolate the movement from its context to perfect it, you destroy the informational ecosystem that made it functional. Pinder and his team set out in 2011 the criterion that remains the reference today: what matters is fidelity to the informational constraints of the performance context, not the simplicity of the drill.

Watch the nuance, it is decisive. Doing everything in match situations is not the goal. You can simplify: reduce the number of players, enlarge the space, slow the tempo, ease a time constraint. Simplify, yes. Decontextualize, no. The boundary runs through the informational invariants that guide the decision. In simplification, you keep them: a real defender, a variable target, a signal that changes. In decontextualization, you remove them.

That is why the rhythm ladder laid on the floor, all by itself, without the slightest external signal, is the archetype of the decontextualized drill. The system learns to execute a predetermined motor pattern to perfection, but it doesn't learn to read the situation and decide, which still represents 90% of what is at stake in a match. You can simplify the pool, you cannot drain it.

Myth 4: "The goal is for them to succeed, so avoid errors"

This reflex is almost automatic. The athlete falls, so you secure. They miss, so you simplify. They fail several times, so you correct upstream so that they no longer miss. It's protective, it's caring, it's what is taught in every classic coach and educator training. And pedagogically, it's devastating.

Go back to the number from the start. Karen Adolph, 2012: a baby learning to walk falls 17 times an hour. This rate is not a problem to fix. It is what produces the infant's absolutely astonishing speed of learning. The child learns to walk thanks to falling. The reason is mechanical, and you have already come across it: the brain learns through prediction error, a solid consensus since Friston and Wolpert. When the prediction matches reality exactly, there is no internal reason to recalibrate. No gap, no learning.

Schmidt and Bjork, them again, put a name to this in the 1990s: "desirable difficulties". These are the conditions that produce more errors in the session and improve delayed retention. In other words, a session that is too clean is often a session that teaches nothing.

Secure too systematically, and you create two drifts. First, athletes who no longer tolerate error emotionally, because they have learned, culturally, that error is bad, that it is a failure, a judgment on them. Second, athletes who collapse at the first real difficulty, because they have only ever progressed in conditions where you had prepared the ground for them to succeed.

The alternative has a name: organized tolerance of error. You preserve physical safety, obviously, that is non-negotiable. But you calibrate conditions that produce enough errors to stimulate learning without saturating. And above all, this is probably the most important, you educate your athlete to read error as information and not as a verdict on who they are. At bottom, 17 falls an hour is the signature of learning that works. Error plays the role of fuel.

Myth 5: "Explain precisely what they have to do"

This is probably the deepest reflex, the one we all trigger without thinking in 90% of our sessions. "Extend the arm, bend the knee, put the flat of the foot down, hip forward, keep the back straight." We have all learned to correct through speech.

An old parable sums up the trap: the centipede. It has always walked along quietly, its thousand legs organize themselves, it's fluid, it's efficient. One day, someone asks it: how do you do it, exactly? Which leg do you move first? The second? And the back ones, in what order? The centipede stops, analyzes, breaks it down, tries to move consciously, and finds itself unable to walk. That is precisely what happens in your athlete's brain when you over-verbalize with them.

Motor learning is very largely implicit. It happens without verbalization, or with very little. The most striking proof is the baby: walking, running, fine grasping, all the most complex human motor skills, acquired without any verbal instruction or tutorial. Nobody tells them to orient their pelvis or recruit this or that muscle. They learn by doing. At the other end of the spectrum, the expert, a top-level footballer, a virtuoso musician, a seasoned surgeon, knows how to do it but often can't say how. And when you force them to explain it, their performance degrades.

Beilock documented this in a famous study published in 2002. She takes expert golfers and asks them to verbalize precisely what they are doing as they hit the ball. Result: their performance drops significantly. Verbalization activates explicit consciousness, which interferes with perfectly mastered implicit motor processes. This is exactly what Gabriele Wulf has been documenting for 25 years with her work on external focus. Ask your athlete to think about their body, their segments, their internal parameters, and you turn them into a centipede, you make them worse.

So what do we do instead? We manipulate constraints instead of giving internal instructions. You move a target, change a rule of the game, modify a support, add a time constraint, introduce a signal that appears at an unexpected moment. The system finds the solution on its own, without being dictated what to do with its body. A rare, well-placed demonstration is almost always worth more than ten minutes of verbal explanation. And silence during execution is almost always worth more than speech.

Five reflexes to question again in your sessions

Let's recap the five myths: identical repetition, the universal ideal movement, isolating the movement out of context, avoiding error, precise verbal instruction. Five phrases we hear every day in our professions, all with an apparent logic, and all invalidated one by one by contemporary motor learning science for 60 years.

One thread runs through them. Learning needs a gap to correct, so vary. It needs its informational context, so don't decontextualize. It feeds on error, so organize it instead of fleeing it. It is largely implicit, so manipulate constraints rather than instruct. Once you have this grid in mind, you can audit your own philosophy of coordination training and spot, session after session, where you unintentionally remove the engine of progress.

These five myths are only a way in, there are many more. To go further on coordination training, a complete guide is waiting for you in the episode description.

FAQ

What is motor learning?

It is the process by which the brain learns to produce and refine its movements. Its central engine is prediction error: the gap between what the brain anticipates and what actually happens. This gap feeds cerebellar learning, and without a gap there is no learning. This process is also very largely implicit: it is built by doing, with little or no verbalization, as in the baby learning to walk.

What criticisms are made of the classic approach to training?

Five traditional reflexes are challenged by motor learning science: repeating the same movement identically, aiming for a single reference movement, isolating the movement from its context, trying to avoid errors, and verbally explaining every detail of the movement. Each starts from a reasonable intention, but slows progress because it removes the gap, the context, the error or the implicit dimension that learning needs.

Do you have to repeat to learn a movement?

Yes, repetition is necessary, the baby and the expert both repeat enormously. But form matters more than quantity. Identical repetition gives a fine performance in the session and weak retention as soon as conditions change. Varied repetition, where you modify a parameter as soon as the movement is reproduced the same way twice, builds far more robust knowledge of the situation.

Is there an ideal technique valid for all athletes?

No. Movement emerges from the interaction between the task, the environment and the individual (constraints model of Bernstein and Newell). Individual constraints, anthropometry, motor history, capacities, impose a solution specific to each body. In fine analysis, two champions of the same sport never have exactly the same movement. Imposing a single model makes the athlete plateau, tires them and raises their injury risk.

Are errors useful for motor learning?

Yes, error is the fuel of learning. A baby falls about 17 times an hour when learning to walk (Adolph, 2012), and it is this rate of falling that produces their speed of learning. The brain recalibrates thanks to the gap between prediction and reality. The right stance is to organize tolerance of error: physical safety preserved, and error read as information, not as a judgment.

Should you verbally explain the movement to the athlete?

As little as possible. Motor learning is largely implicit, and over-verbalization activates explicit consciousness, which interferes with mastered motor processes (centipede effect; performance drop in expert golfers forced to verbalize, Beilock, 2002). It is better to manipulate constraints (move a target, change a rule, add an unexpected signal), place a demonstration at the right moment, and stay silent during execution.

Related articles

• 3 mistakes that destroy motor learning (and how to fix them)

• Variability and motor learning: why repeating the same movement caps your progress

• Motor learning: the link between technique and physical qualities

• Archaic reflexes in children: why so many motor delays and learning difficulties, and what to do

Discover the Neuro-Postural Reprogramming training