5.16 Neuroplasticity and Skill Acquisition

Your most impressive training session this month may have taught you almost nothing. And that scrappy, awkward one where nothing quite clicked? That may be the one that changes you.

‌

This isn't a motivational poster. It's one of the most useful and least intuitive findings in sports science: the practice that looks best on the day is generally not the practice that sticks.

‌

Blocked, low-variability, heavily guided practice — one skill on repeat, in a predictable order, with correction after every attempt — looks excellent in the session and generalises poorly. Random, variable practice with reduced feedback looks worse on the day and produces better learning. Your brain, it turns out, marks your homework differently to your ego.

‌

Behind all of this sits plasticity — and it isn't a single process you can switch on. Your nervous system changes at five distinguishable levels: synaptic strength, structural synaptic change, myelination, cortical map reorganisation and network-level reconfiguration.

‌

Each level runs on its own timescale, from seconds to years, and skill acquisition exploits all five at once. Below: the mechanisms, the three stages of motor learning, the contextual interference effect, how to design feedback, and why sleep is part of the practice.

‌

Key takeaways

1. Plasticity operates at five levels, each on its own timescale. Early learning is largely synaptic; expertise involves whole networks reorganising over years.
2. Performance and learning are not the same thing. How you look within a session is a poor — and sometimes inverted — indicator of what you'll retain.
3. The contextual interference effect: random practice produces worse acquisition and better retention than blocked practice.
4. Fade your feedback. Constant, immediate feedback creates dependency; reduced-frequency and delayed feedback improve retention.
5. Sleep is part of the practice. Motor consolidation depends on it, and a skill practised without subsequent sleep is not fully learned.
6. Trial-to-trial variability falls later than average performance rises. A plateau with high variability means the skill is not yet consolidated.

‌

Beginner section: Looking good today versus being good next month

Here's the deal. If you want to look good in training, practise one thing repeatedly, in a predictable order, with a coach telling you what you did after every attempt.

‌

If you want to be good next month? Do close to the opposite.

‌

The reason is that performance and learning are different things. Performance is what you can do right now, with the current arrangement propping you up. Learning is a durable change in capability — the part that's still there once the props are removed.

‌

Think open-book exam versus closed-book. Predictable drills with constant guidance are the open book: flattering scores, no guarantee anything got stored. Practice conditions can improve one while actively harming the other.

‌

You've seen this in the gym. Someone nails a technique drill all evening with the coach's voice in their ear, comes back a week later, and it's gone. The session flattered them; it didn't change them.

‌

‌

The three stages every skill moves through

The classic description of skill learning has three stages, and you can feel each one from the inside (Fitts & Posner, 1967).

‌

Stage one is cognitive. Everything is deliberate, errors are large and inconsistent, and you have to think about each part separately — feet, hands, breath, all queued up in your head. First session learning to deadlift, basically.

‌

Stage two is associative. The basic pattern exists and you're refining it: errors get smaller and more consistent, and the movement stops feeling like a list of instructions.

‌

Stage three is autonomous. The skill runs without conscious supervision, and you can attend to other things while doing it — the tactics, the defender, the next rep — instead of the movement itself.

‌

The line on the graph that tells you the truth

Now look at the second line in that graph, because it's the more useful one.

‌

Variability — how different each attempt is from the last — keeps falling after average performance has largely flattened. Consistency arrives late to the party.

‌

That hands you a practical test you can run on yourself this week. If your average is decent but individual attempts still look quite different from one another, the skill is not consolidated — no matter how good the good ones are.

‌

One-line recap: stop asking how good your best attempt was. Start asking how much your attempts resemble each other.

‌

‌

Sleep is part of the session

The other thing worth knowing early: sleep is part of practice. Not a recovery bolt-on afterwards — part of the training itself.

‌

Motor skills continue to be consolidated during sleep after the session ends. In some studies, performance improves overnight with no additional practice at all — your brain keeps working on the skill while you're unconscious (Walker et al., 2002).

‌

So practising something hard and then sleeping four hours is not the same training stimulus as practising it and sleeping eight. Same session on paper; a different session in your nervous system.

‌

Advanced section: Mechanisms, contextual interference, and feedback

Five levels of plasticity

‌

• Synaptic strength. Long-term potentiation and depression alter the efficacy of existing synapses over seconds to hours, largely through AMPA receptor trafficking and phosphorylation, with NMDA receptors acting as coincidence detectors. This is the substrate of within-session and early between-session learning.
• Structural synaptic change. Dendritic spines are formed and eliminated within hours to days. This requires new protein synthesis, which is one reason the hours after practice matter and why sleep and nutrition are not merely supportive.
• Myelination. Oligodendrocyte precursor cells respond to activity, and activity-dependent myelination alters conduction velocity and, importantly, the timing of arrival between converging pathways (Sampaio-Baptista & Johansen-Berg, 2017). For coordination tasks, getting signals to arrive together matters as much as getting them to arrive fast.
• Map reorganisation. Cortical representations expand for trained effectors. String players show enlarged representation of the left-hand digits, and the effect is larger in those who began earlier (Elbert et al., 1995). This operates over weeks to months.
• Network reconfiguration. With expertise, the same task recruits different and often less extensive circuitry. Novices show widespread prefrontal activation for tasks that experts perform with more focal, efficient activity.

‌

Draganski and colleagues provided one of the clearest human demonstrations of the middle levels: Three months of juggling practice produced measurable grey matter increases in motion-processing areas, which partially regressed after three months without practice (Draganski et al., 2004). Both halves of that result matter. Structural change is real, and it is not permanent without maintenance (Zatorre et al., 2012).

‌

Contextual interference: The effect that changes how you programme

Shea and Morgan’s 1979 experiment is the canonical demonstration. Participants learned three movement patterns either in blocks (all of pattern one, then all of pattern two, then all of pattern three) or in random order. During acquisition, the blocked group performed clearly better. On retention and transfer tests days later, the random group performed clearly better (Shea & Morgan, 1979).

‌

‌

The dominant explanations are the elaboration hypothesis, that interleaving forces comparison between movements and richer encoding, and the reconstruction hypothesis, that having to retrieve and rebuild the motor plan on each trial strengthens the retrieval process itself. Both predict that the extra difficulty is the mechanism rather than a cost.

‌

Two qualifications keep this honest. The effect is more reliably demonstrated in laboratory tasks than in complex sport skills, and it interacts with skill level: Complete novices may need some blocked practice to establish a basic pattern before interleaving becomes productive. The practical rule that survives is to use blocked practice to establish the movement and switch to random and variable practice once it exists at all.

‌

‌

Feedback: Less is usually more

Feedback is one of the most over-applied coaching tools. The general finding, developed extensively by Salmoni, Schmidt and Walter and by Winstein and Schmidt, is that augmented feedback improves performance while it is present and can impair learning when it is present too often (Salmoni et al., 1984; Winstein & Schmidt, 1990).

‌

The proposed mechanism is guidance dependency. If external feedback is available after every attempt, the learner uses it instead of developing their own intrinsic error-detection capability. Remove the feedback and performance collapses, because the error-detection was never built.

‌

• Reduce frequency. Feedback on roughly a third to a half of trials generally produces better retention than feedback on every trial (Winstein & Schmidt, 1990).
• Delay it slightly. A short delay before feedback allows the learner to generate their own estimate first, which is what builds error detection.
• Summarise rather than itemise. Feedback averaged over several attempts often outperforms attempt-by-attempt feedback (Salmoni et al., 1984).
• Use external rather than internal focus. Wulf’s body of work indicates that directing attention to the movement effect — the bar path, the target — generally produces better outcomes than directing it to body parts (Wulf, 2013).
• Fade deliberately. Start with more feedback and plan its withdrawal, rather than providing constant feedback indefinitely.

‌

Consolidation and sleep

Learning does not end when practice ends. Motor memories are consolidated offline, and sleep plays a substantial role (Krakauer & Shadmehr, 2006). Walker and colleagues showed overnight improvements in motor sequence performance that were sleep-dependent rather than simply time-dependent, and slow-wave sleep and sleep spindle activity have both been associated with motor consolidation (Walker et al., 2002).

‌

There is also a well-documented interference window. Practising a second, competing motor task shortly after learning a first can disrupt consolidation of the first (Brashers-Krug et al., 1996). This argues for not stacking two novel technical skills back to back, and for separating the introduction of new techniques rather than overhauling several elements at once.

‌

Combined with the material in Article 5.9 on cholinergic gating of plasticity, this gives a coherent picture of what a good technical session looks like: Attended, moderately difficult, varied once the basic pattern exists, with feedback faded, not immediately followed by a competing novel skill, and followed by adequate sleep.

‌

‌

Practical section: Designing practice that actually sticks

1. Establish, then interleave. Use blocked practice only until the basic pattern exists, then mix it with other skills and conditions.
2. Vary the task, keep the context. Change loads, tempos, implements and angles. Do not change the time, place and order, because that is what habit formation depends on.
3. Fade the feedback. Plan a reduction schedule rather than coaching every rep indefinitely.
4. Cue externally. Point attention at the effect of the movement rather than the body part producing it.
5. Protect attention. Skill work with divided attention is physically less eligible to consolidate. Phone away.
6. Do not stack novel skills. Introduce one new technical element at a time and leave a gap before the next.
7. Sleep after technical work. Consolidation depends on it, and there is no substitute.
8. Test retention, not acquisition. Judge a technical change by how it looks after 48 hours or a week, not at the end of the session.

‌

That last point is the one most worth adopting. Assessing a technical change at the end of the session in which it was introduced will systematically favour the approaches that produce the least learning.

‌

‌

Sport applications

• Open-skill sports. Random and variable practice matters more here than anywhere, because the competitive environment is itself unpredictable.
• Closed-skill sports. Gymnastics, diving and Olympic lifting still benefit from variability in load and setup, but the case for random practice is weaker where the competition condition is fixed.
• Youth development. Early cortical map effects appear larger with earlier onset, which supports broad movement exposure rather than early narrow specialisation.
• Rehabilitation. Guidance dependency is a real risk when a therapist provides constant correction. Feedback fading should be planned into the programme.
• Late-career technical changes. Chunked patterns resist mid-sequence editing, so expect a worse-before-better phase and schedule it far from competition.

‌

Common mistakes

• Judging a technical change by end-of-session performance. Acquisition and retention frequently diverge. Test after a delay.
• Using blocked practice long after the pattern exists. It flatters within-session performance and produces worse retention and transfer.
• Coaching every repetition. Constant feedback builds dependency instead of intrinsic error detection.
• Cueing internally by default. External focus on the movement effect generally outperforms attention to body parts.
• Introducing several technical changes at once. Competing novel motor tasks interfere with each other during consolidation.
• Treating sleep as optional after skill work. Motor consolidation is sleep-dependent. The session is not finished when practice stops.

‌

Coaching cues

• Block it until it exists, then mix it up.
• Vary the task, not the time and place.
• Coach less often than feels natural.
• Cue the effect, not the body part.
• One new technical change at a time.
• Judge the change next week, not tonight.

‌

FAQs

Why do I perform worse when I mix up my practice?

That is the expected result and it is not a problem. The contextual interference effect describes exactly this: Interleaved practice produces poorer performance during acquisition and better retention and transfer afterwards. The extra difficulty of retrieving and rebuilding the movement plan on each attempt appears to be the mechanism rather than a cost, so the worse in-session performance is a sign the arrangement is working.

‌

How much feedback should a coach give?

Less than most give. Reviews of augmented feedback indicate that feedback on every trial improves performance while present and impairs retention, because learners come to rely on it instead of developing their own error detection. Feedback on roughly a third to a half of attempts, slightly delayed, and sometimes summarised across several attempts, generally produces better learning. It should also be planned to fade over time.

‌

Does the brain really change from training, or is that a metaphor?

It changes measurably. Three months of juggling practice has been shown to increase grey matter in motion-processing areas, with partial regression after three months without practice. String players show expanded cortical representation of the left-hand digits. Activity-dependent myelination alters conduction timing along trained pathways. These are structural findings, not metaphors, though the everyday use of the word neuroplasticity is often much looser than the evidence.

‌

How important is sleep for learning a skill?

Substantial. Motor sequence learning shows sleep-dependent overnight improvement rather than simply time-dependent improvement, and slow-wave sleep and spindle activity are associated with consolidation. Practically this means a technical session followed by poor sleep is a weaker stimulus than the same session followed by good sleep, and it argues against scheduling important technical work in periods of unavoidable sleep restriction.

‌

Should I focus on my body or on the target?

Generally the target, or more precisely the intended effect of the movement. A large body of work on attentional focus indicates that an external focus — the bar path, the ball, the landing spot — tends to produce better outcomes than an internal focus on body segments. Internal cues have their place in early acquisition or when correcting a specific positional fault, but they are over-used as a default.

‌

Is it too late to learn a new skill as an adult?

No, though the picture is nuanced. Adult plasticity is well documented and the mechanisms described in this article all operate throughout life. What does change is that some map-level effects appear larger with earlier onset, and adults often have well-chunked competing patterns that must be unpacked. The practical implication is that adult skill learning benefits from more deliberate practice design rather than from more volume.

‌

Recommended videos

Each video below was chosen because it covers a specific part of this article in more depth than text alone allows.

‌

‌

‌

‌

‌

‌

Related reading on FitXplor

• 5.5 The Athlete’s Brain
• 5.9 Serotonin, Noradrenaline, and Acetylcholine
• 5.14 The Neuroscience of Habit Formation
• 1.5 The Nervous System and Athletic Performance
• 1.3 Biomechanics of Human Movement

‌

References

Fitts, P. M., & Posner, M. I. (1967). Human Performance. Brooks/Cole.

Shea, J. B., & Morgan, R. L. (1979). Contextual interference effects on the acquisition, retention, and transfer of a motor skill. Journal of Experimental Psychology: Human Learning and Memory, 5(2), 179–187.

Salmoni, A. W., Schmidt, R. A., & Walter, C. B. (1984). Knowledge of results and motor learning: a review and critical reappraisal. Psychological Bulletin, 95(3), 355–386.

Winstein, C. J., & Schmidt, R. A. (1990). Reduced frequency of knowledge of results enhances motor skill learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16(4), 677–691.

Wulf, G. (2013). Attentional focus and motor learning: a review of 15 years. International Review of Sport and Exercise Psychology, 6(1), 77–104.

Draganski, B., Gaser, C., Busch, V., Schuierer, G., Bogdahn, U., & May, A. (2004). Neuroplasticity: changes in grey matter induced by training. Nature, 427(6972), 311–312.

Elbert, T., Pantev, C., Wienbruch, C., Rockstroh, B., & Taub, E. (1995). Increased cortical representation of the fingers of the left hand in string players. Science, 270(5234), 305–307.

Walker, M. P., Brakefield, T., Morgan, A., Hobson, J. A., & Stickgold, R. (2002). Practice with sleep makes perfect: sleep-dependent motor skill learning. Neuron, 35(1), 205–211.

Krakauer, J. W., & Shadmehr, R. (2006). Consolidation of motor memory. Trends in Neurosciences, 29(1), 58–64.

Zatorre, R. J., Fields, R. D., & Johansen-Berg, H. (2012). Plasticity in gray and white matter: neuroimaging changes in brain structure during learning. Nature Neuroscience, 15(4), 528–536.

Sampaio-Baptista, C., & Johansen-Berg, H. (2017). White matter plasticity in the adult brain. Neuron, 96(6), 1239–1251.

Brashers-Krug, T., Shadmehr, R., & Bizzi, E. (1996). Consolidation in human motor memory. Nature, 382(6588), 252–255.