Near vs far transfer: the honest limits of brain training
Practice makes you better at the practiced task (near transfer). Broad gains to everyday thinking (far transfer) are weak and contested.
Do not delay medical evaluation of real cognitive concerns in favor of brain-training apps.
- reliance on brain-training instead of seeking help for genuine memory concerns
What the evidence says
The sources are systematic reviews and meta-analyses of cognitive training in older adults. Butler 2018 (systematic review) reports moderate-strength evidence that cognitive training improves performance in the trained domain in cognitively normal adults, and evidence insufficient for preventing incident mild cognitive impairment or dementia; Sprague 2019 reports consistent evidence of transfer from cognitive training to the trained domains and inconsistent evidence of transfer to other cognitive outcomes; Mewborn 2017 (meta-analysis) reports a small significant improvement (g = 0.298), largest for working-memory interventions, including transfer effects; Gavelin 2019 reports small effects (g = 0.32 to 0.40) whose clinical value remains unclear; Gates 2019 (systematic review) reports low-quality evidence from one study and cannot determine whether computerised training maintains global cognition in midlife. Commercial brain-training products that promise general 'sharper thinking' claim more than these sources report.
- 'Brain-training games make you smarter overall' (they mainly make you better at the game (near transfer))
- 'If a game feels hard, it must be building general brainpower' (difficulty of a task does not equal broad transfer)
- Brain training does not prevent dementia or cognitive decline
- Brain training does not raise IQ or produce general 'smarter' thinking
- Gains do not reliably transfer to everyday tasks
- Agreement
- Direction recorded for 0 of 5 sources; not yet recorded for the rest.
- The claim's outcome
- Outcome not yet recorded
- Certainty
- A source states a certainty rating for 1 outcome.
- How the sources were read
- Some sources are based on abstracts, not full texts.
What each source measured and concluded (5)
- What it measured
- Cognitive decline and incident dementia (ADRD); intermediate cognitive test performance
- How it was read
- based on the abstract, not the full text
- What the review says it includes
- Systematic review of randomized controlled trials
- Studies and participants, as stated
- 34 unique RCTs (11 with medium/low risk of bias); samples 40 to 2832
- What it measured
- Efficacy of cognitive interventions on cognitive functioning (neuropsychological tests)
- How it was read
- based on the abstract, not the full text
- What the review says it includes
- Systematic review and meta-analysis of randomized-controlled trials
- Studies and participants, as stated
- 279 effects from 97 studies
- What it measured
- Objective cognitive function.
- How it measured it
- Objective cognitive tests (validated).
- How it was read
- read in full text
- What the review says it includes
- Umbrella review of 51 systematic reviews/meta-analyses
- Heterogeneity, as stated
- The extent to which such effects are of clinical value remains unclear, due to the scarcity of high-quality evidence and heterogeneity in reported findings."
- Heterogeneity, as stated
- Heterogeneity in reported findings
- What it measured
- Cognitive function.
- How it measured it
- Various objective cognitive tests and assessments (validated).
- How it was read
- read in full text
- What the review says it includes
- Systematic review of 75 RCTs
- Heterogeneity, as stated
- Conclusion: "There is inconsistent evidence of transfer from behavioral interventions to cognitive outcomes in healthy older adults.
- What it measured
- Global cognitive functioning.
- How it measured it
- Validated tests.
- How it was read
- read in full text
- What the review says it includes
- Systematic review of 1 RCT
- Studies and participants, as stated
- Enrolled: N=6742; Analyzed: N=not reported
Certainty ratings the sources state (1)
- Rating, as the source states it
- low
- Framework
- not named by the source
- Applies to
- Conclusion: "We found low-quality evidence from only one study.
- Stated by
- Gates et al. (2019)
Every source below is shown with its study design and links to the original study. This is the audit trail.
“Abstract only” means the review read that paper’s abstract, not its full text; “screened only” means it was included at re-screening and not extracted. Both stay in the trail with their DOI, because how a source was read is part of the audit.