Learn physiotherapy · BPT-402 Neurological Physiotherapy
Motor Learning
You will be taught that external focus cueing is better than internal. Seven meta-analyses agreed. A 2024 reanalysis that modelled publication bias found the effect on performance shrank to g = 0.01 — essentially nothing.
In one line. Motor learning is a relatively permanent change in the capability to perform a skill, inferred from retention and transfer tests rather than from how well someone does during practice.
That distinction is the whole subject. Performance during a session can improve while learning does not, and several practice manipulations that look worse during practice produce better retention.
The external focus evidence has been challenged, hard
A 2021 meta-analysis of 73 studies and 1,824 participants found an external focus superior for performance (g = 0.264), retention (0.583) and transfer (0.584). [1] It is the basis of what most programmes teach about cueing.
A 2024 robust Bayesian reanalysis of those same data, fitting several plausible publication-bias models, found moderate to strong evidence of publication bias in every analysis. Corrected estimates were g = 0.01 for performance, 0.15 for retention, 0.09 for transfer and 0.06 for muscle activity. [2] Seven previous meta-studies had all reported superiority. This page does not pretend the question is closed, but the correction is the more recent and methodologically stronger analysis.
What the evidence shows
| Question | Finding | Source and quality |
|---|---|---|
| Is an external focus better? | On the original analysis, yes: Hedges' g 0.264 for performance (95% CI 0.217 to 0.310), 0.583 for retention (0.425 to 0.741), 0.584 for transfer (0.325 to 0.842) | 73 studies and 1,824 participants for performance; 40 studies and 1,274 for learning. Effects held across age, health status and skill level [1] |
| Does it survive correction for publication bias? | Largely no. g = 0.01 (performance), 0.15 (retention), 0.09 (transfer), 0.06 (muscle activity) after correction, with moderate to strong evidence of bias in all analyses | Robust Bayesian meta-analysis of the same dataset using several publication-bias models [2] |
| Does external focus help muscular endurance? | Reported yes: Cohen's d 0.58 (95% CI 0.34 to 0.82) versus internal focus and 0.42 (0.08 to 0.76) versus control. Internal focus did not differ from control | Small evidence base — seven comparisons across 14 study groups. Read alongside [2] rather than instead of it [4] |
| Does varied, randomised practice improve retention? | Yes overall, a medium beneficial effect — but almost negligible in applied settings, while effective in laboratory settings. Larger effect in older adults (large) than adults (medium) | 54 studies from 294 screened full texts, analysed with both a three-level mixed model and a random-effects model [3] |
| Does motor imagery add anything? | Possibly, combined with physical practice: standardised mean differences 0.83 to 1.87 across pooled studies | 22 studies, 934 children and adolescents aged 9 to 18. Risk of bias ranged from some concerns to high and the GRADE rating was low — treat as promising, not established [5] |
Where students get this wrong
1. Judging learning by how the session went
Learning is inferred from retention and transfer, not from performance during practice. This is why the contextual interference literature exists at all: random practice often looks worse in the session and better days later. [3] A patient who performs well while you cue them heavily may have learned less than one who struggled.
2. Teaching external focus cueing as established fact
It is the most confidently taught idea in this area and the one whose evidence has just been most seriously undermined. [2] You should still know it, still use it — a corrected retention estimate of g = 0.15 is not zero — but presenting it as a strong, settled effect is no longer defensible. If you cite the 2021 meta-analysis, [1] cite the 2024 reanalysis with it.
3. Assuming laboratory findings transfer to the clinic
The contextual interference effect was effective in laboratory settings and almost negligible in applied settings. [3] That is a pattern worth generalising: motor learning research is largely done on healthy volunteers learning artificial tasks, and the closer the setting gets to real life the smaller the effects tend to become.
4. Treating publication bias as somebody else's problem
Seven separate meta-analyses reported the same conclusion, and a reanalysis modelling publication bias overturned it. [2] Agreement between reviews is not independent confirmation when they draw on the same biased literature. This is the single most useful methodological lesson on the page, and it applies far beyond motor learning — see appraising a systematic review.
5. Reading a large effect size without reading its GRADE
The motor imagery pooled effects look impressive at 0.83 to 1.87, and the same paper reports risk of bias from some concerns to high and a GRADE rating of low certainty. [5] Effect size and certainty are separate questions, as on GRADE.
6. Applying findings from children or athletes to your patient
The motor imagery evidence is in 9 to 18 year olds, [5] the focus evidence largely in healthy adults and athletes, [1] and the endurance evidence in resistance exercise. [4] None of it was generated in the neurologically impaired adults where physiotherapists most often invoke motor learning principles.
What the evidence supports — and what it does not
Supported
- Measuring learning by retention and transfer, not in-session performance. [1][3]
- Varied and randomised practice, particularly in older adults. [3]
- External focus cueing as a reasonable default, with a small expected effect. [1][2]
- Motor imagery as an adjunct to physical practice, on low-certainty evidence. [5]
Not supported
- External focus as a large or settled effect. Corrected g = 0.01 for performance. [2]
- Expecting laboratory-sized contextual interference effects in clinic. [3]
- Treating agreement between meta-analyses as confirmation. [2]
- Motor imagery as established practice. GRADE low. [5]
- Transferring these findings to neurological patients without saying you are extrapolating. [1][5]
How certain is this?
Evidence grade: Low to moderate.
The grade is deliberately below what the raw effect sizes would suggest, because the most methodologically careful paper here is the one that found the field's headline result to be substantially an artefact of publication bias. [2] When a literature has demonstrable bias, the confidence intervals in its primary studies understate the real uncertainty.
The contextual interference review is large and well conducted, using two analysis models and reporting the laboratory-versus-applied split honestly rather than pooling it away. [3] The motor imagery review states its own GRADE rating as low. [5] The endurance meta-analysis rests on seven comparisons. [4]
What would change the grade: pre-registered trials of focus and practice-schedule manipulations in clinical populations, powered and reported regardless of result.
Common questions
Should I stop using external focus cues?
No. The bias-corrected estimate for retention was g = 0.15, [2] which is small but not zero, and external cueing costs nothing. What should change is your confidence and your teaching: it is a reasonable default, not a demonstrated principle, and you should say so when you cite it.
What is the difference between performance and learning?
Performance is how someone does now, during or immediately after practice. Learning is the durable change, tested after a delay (retention) or on a related task (transfer). Manipulations that degrade performance while improving retention are the reason the distinction matters — random practice is the classic example. [3]
Should I randomise practice order with my patients?
Cautiously. The effect was medium overall and almost negligible in applied settings, though larger in older adults. [3] Randomising task order is low cost and low risk, so it is reasonable to do — just do not expect the effect sizes reported from laboratory studies.
Is motor imagery worth adding?
As an adjunct to physical practice, possibly. Pooled effects of 0.83 to 1.87 look strong, but the same review reports risk of bias from some concerns to high and rates the certainty low, in 9 to 18 year olds. [5] It is cheap and safe, which is a reasonable basis for trying it; it is not an evidence-based substitute for practice.
Why does this page contradict my textbook?
Because the reanalysis is from 2024 and most textbooks predate it. [2] That is not a reason to distrust textbooks generally; it is a reason to check whether a strongly stated principle has been re-examined. A literature in which seven reviews agreed and a bias-corrected reanalysis disagreed is exactly where a student should look carefully.
References
- Chua LK, Jimenez-Diaz J, Lewthwaite R, et al. Superiority of external attentional focus for motor performance and learning: Systematic reviews and meta-analyses. Psychological Bulletin. 2021 Jun;147(6):618–645. doi:10.1037/bul0000335 PMID 34843301 Systematic review and meta-analysis
- McKay B, Corson AE, Seedu J, et al. Reporting bias, not external focus: A robust Bayesian meta-analysis and systematic review of the external focus of attention literature. Psychological Bulletin. 2024 Nov;150(11):1347–1362. doi:10.1037/bul0000451 PMID 39480294 Robust Bayesian meta-analysis
- Czyż SH, Wójcik AM, Solarská P, et al. High contextual interference improves retention in motor learning: systematic review and meta-analysis. Scientific Reports. 2024 Jul 10;14(1):15974. doi:10.1038/s41598-024-65753-3 PMID 38987617 Systematic review and meta-analysis
- Grgic J, Mikulic P. Effects of Attentional Focus on Muscular Endurance: A Meta-Analysis. International Journal of Environmental Research and Public Health. 2021 Dec 22;19(1):. doi:10.3390/ijerph19010089 PMID 35010348 Meta-analysis
- Behrendt F, Zumbrunnen V, Brem L, et al. Effect of Motor Imagery Training on Motor Learning in Children and Adolescents: A Systematic Review and Meta-Analysis. Children. 2021 Sep 8;18(18):. doi:10.3390/ijerph18189467 PMID 34574389 Systematic review and meta-analysis
About this resource
How to use this
Written to be learned from, not memorised.
This page reports both the established finding on attentional focus and the reanalysis that overturned it, because a student who knows only the first will teach it with more confidence than the evidence now supports. Faculty may use this page in teaching with attribution. It carries its review date and its next review date, so you can see at a glance whether it is current before you put it in front of a cohort.