Performance is not learning
In the earlier guides of this rung you met the motor system as machinery — the cortex, cord, and muscle that turn an intention into a movement, and the theories of motor control that try to explain how that intention is shaped. This guide asks a different question: how does a movement that was clumsy or impossible become smooth and automatic? That is motor learning, and it is the engine underneath almost everything a therapist does. A patient does not relearn to walk because the therapist's hands are magic; they relearn because practice is reshaping the nervous system.
The single most important idea in this whole field is also the most counter-intuitive: how well someone moves during a practice session — their performance — is not the same as how much they have actually learned. Learning is a relatively permanent change you can only measure later, when the coach is gone, the cues have faded, and the person tries the skill again, cold, in a new setting. A practice arrangement that makes today's session look beautiful can leave almost nothing behind; one that makes today look messy and effortful can build a skill that lasts. Hold on to this distinction — nearly every practical decision below flows from it.
Three stages of learning a movement
The classic map of how a skill develops describes three stages of motor learning, and although the boundaries are fuzzy, the picture is genuinely useful at the bedside. In the cognitive stage, the learner is figuring out what to do at all. Movements are slow, jerky, full of errors, and they cost enormous mental effort — the person has to think hard about every part, and they cannot hold a conversation while they do it. Watch someone use a walking frame for the first time after a hip fracture: their whole attention is consumed by where the frame goes and which foot moves next.
In the associative stage, the basic idea is in place and the work shifts to refinement. Errors get smaller and more consistent, the movement smooths out, and the learner starts to feel which adjustments help. This stage can be long — weeks to months for a complex skill. Finally, in the autonomous stage, the movement runs largely on its own. It is fast, accurate, resistant to distraction, and cheap in attention — the person can now talk, carry something, or scan traffic while doing it. Crossing into autonomy is exactly what frees up the mind for everything else life demands; a gait that still requires total concentration is not yet a useful gait.
Knowing the stage tells a therapist how to behave. In the cognitive stage you simplify, demonstrate, give frequent guidance, and keep the task achievable so the learner does not drown. As the person moves into the associative stage, you deliberately back off — fewer cues, more independent attempts, more room to make and feel their own errors. Pushing autonomous-stage challenges onto a cognitive-stage learner overwhelms them; smothering an associative learner with constant correction keeps them dependent. Reading the stage is one of the first clinical judgements in this whole rung.
How you schedule practice changes what sticks
Once a learner is practising, the next set of levers is the practice schedule — how the repetitions are arranged in time. Two contrasts matter most. The first is blocked versus random practice. In blocked practice you drill one task many times before moving on (AAAA, then BBBB). In random practice you interleave several tasks so the same one rarely comes up twice in a row (A B C B A C). Here is the surprise: blocked practice usually produces better performance during the session, but random practice usually produces better learning — better retention later and better ability to handle a new situation.
The leading explanation is that random practice forces the brain to keep rebuilding the movement plan from scratch each time, instead of just repeating the last attempt — that extra effort, sometimes called desirable difficulty, is what burns the skill in. The honest caveat is that a learner deep in the cognitive stage, or very frail, may simply need some blocked repetition first just to grasp the idea before randomness helps. So the practical move is a glide: more blocked early to build the basic pattern, then progressively more random as the person stabilizes. It is a dial, not a switch.
The second contrast is massed versus distributed practice — whether the repetitions are packed into one long block or spread out with rest between them. Distributed practice, spread across more shorter sessions, generally beats one exhausting marathon, partly because fatigue corrupts the very movement you are trying to engrave, and partly because the spacing itself seems to help memory consolidate. This is one quiet reason a little practice every day usually outperforms a single heroic weekly session, and why a sensible home program — short, frequent, repeated in the real environment — can do as much heavy lifting as the clinic visit itself.
BLOCKED A A A A | B B B B | C C C C -> looks great today, fades faster RANDOM A B C | B A C | C A B -> messy today, retained better MASSED [ one long, tiring session ] -> fatigue corrupts the pattern DISTRIBUTED [ short ] [ short ] [ short ] -> spacing helps it consolidate
Feedback: the fuel and the trap
A learner needs information about how the attempt went, and that is the role of feedback. Some feedback is intrinsic — the person sees, feels, and hears their own movement. The therapist's job is to add useful extrinsic feedback on top, and it comes in two flavours worth separating. Knowledge of results (KR) tells the learner the outcome: did you reach the cup, did you stay upright, how far did you walk. Knowledge of performance (KP) tells them about the movement itself: your knee buckled, you leaned to the left, your heel never struck the floor. Outcome versus form — both are useful, and they answer different questions.
Now the trap, and it is the same shape as the performance-versus-learning lesson. Feedback after every single attempt makes performance shoot up during the session — but it can quietly become a crutch. If the learner leans on the therapist's voice to tell them what happened, they never build their own internal error-detection, and when the voice is gone the skill collapses. So the better practice is, paradoxically, to give feedback less often as the learner improves: summary feedback after several trials, feedback only when the error crosses a threshold, or feedback the learner has to request. Fading the feedback forces the nervous system to start judging itself — which is the whole point.
Transfer: practising the thing you actually want
All the practice in the world is wasted if the skill stays trapped in the gym. Transfer of training is the degree to which something learned in one task or setting carries over to another — ideally to the patient's real life. The brutal general rule is that transfer is usually narrow: you tend to get good at what you actually practise, in the conditions you practise it, and not much beyond. Strengthening a quadriceps on a machine does not automatically teach a person to climb a kerb; practising sit-to-stand from a high plinth does not guarantee they can rise from their own low sofa.
This is the engine room behind one of the biggest shifts in modern rehabilitation. Because transfer is narrow, the field has moved hard toward task-oriented, specific practice: instead of abstract drills, you practise the actual goal — real walking on real floors, reaching for the real cup, climbing the real stairs the person has at home. You make the practice resemble the target as closely as is safe, and you vary it across the conditions the person will meet, so the skill is robust rather than brittle. When transfer is the bottleneck, specificity is the lever.
Two honest cautions before we leave transfer. First, motor learning is what makes practice stick, but it rests on neuroplasticity — and as the previous and next guides stress, plasticity has limits set by the size and site of the injury. Learning principles let you make the most of the capacity that remains; they do not manufacture capacity that is gone. Second, a relearned skill is not always the original movement returning. Sometimes the nervous system genuinely reclaims the old pattern, and sometimes it learns a clever workaround — the crucial difference between recovery and compensation. Both can be the right goal, but a therapist should always know which one they are actually training, and say so plainly.
The levers, pulled together
Step back and the whole guide collapses into a small set of dials a therapist adjusts to make a movement stick. None of them is exotic; the art is reading the person and turning the right one at the right time. Picture a man relearning to stand and pivot to his wheelchair after a stroke: early on the therapist demonstrates, breaks the transfer into pieces, and guides each one (cognitive stage, blocked practice, frequent feedback). Weeks later the same transfer is rehearsed in random order with other tasks, with feedback only when something goes wrong, practised onto the real chair beside the real bed — because that is where it has to work.
- Read the stage. Cognitive learner: simplify, demonstrate, guide. Associative or autonomous: step back and let them try, err, and self-correct.
- Schedule for learning, not for today. Glide from blocked toward random, and spread practice out (distributed) rather than cramming it (massed).
- Feed back wisely. Mix knowledge of results with knowledge of performance, then fade frequency as skill grows so the learner builds their own error detection.
- Practise the real target. Make the task specific to the person's goal and vary the conditions, because transfer is narrow.
- Be honest about the ceiling. Push the remaining plasticity hard, but name whether you are training recovery or compensation.
Notice how much of this asks the therapist to do less, not more — to back off cues, fade feedback, and tolerate visible, effortful struggle in the room. That restraint is hard, because helping a person succeed in the moment feels like good care. But the deepest message of motor learning is that the productive struggle is the treatment. The next guide turns to the biology that makes any of this possible — neuroplasticity itself — and to the harder question of where, honestly, its limits lie.