The question the research is actually asking
Brain training research distinguishes two types of improvement. Near transfer means getting better at the trained task or very similar tasks. Far transfer means improving on untrained tasks that draw on the same underlying cognitive processes. Near transfer is easy to produce - practice anything and you improve at it. Far transfer is what the marketing promises, and it is where the evidence gets much thinner.
The key question is not "does brain training improve scores on a task?" but "does brain training improve cognitive processing speed in real-world situations?" To benchmark yourself without bias, take the Processing Speed test before and after any training period - it is less susceptible to pure practice effects than the games themselves.
Why this matters
If "brain training" only produces near transfer, you are essentially spending time and money to get better at one game - not improving your cognitive speed in the classroom, at work, or in daily life. That is the core critique that led 75 neuroscientists to publish an open letter in 2014 calling brain training claims "frequently exaggerated."
What the evidence actually shows
Near transfer: consistently demonstrated
Reliable findingEvery serious brain training study shows improvements on the trained tasks and closely related tasks. If you do 30 hours of digit-span training, your digit span improves. If you train on a specific visual search game, your performance on that game improves substantially. This is not controversial - it is just practice.
Far transfer to processing speed: small and inconsistent
Mixed evidenceA comprehensive 2014 meta-analysis by Melby-Lervåg and Hulme of working memory training (n = 2,000+) found significant near transfer but "no convincing evidence" of far transfer to other cognitive abilities including processing speed. A 2016 meta-analysis by Soveri et al. found similar results. Effect sizes for far transfer where they do appear average around d = 0.1 - small and often within measurement error.
Active placebo problem: a critical methodological issue
Important caveatMany brain training studies use passive control groups (people who do nothing) rather than active placebo groups (people who do something equally engaging but cognitively different). When active placebos are used - for example, online crossword puzzles - the apparent benefits of brain training shrink substantially or disappear. A 2010 BBC study (n = 11,430) found that six weeks of brain training produced no greater transfer than general internet browsing.
Exception: speed of processing training in older adults
Promising signalThe ACTIVE (Advanced Cognitive Training for Independent and Vital Elderly) trial - one of the largest cognitive training RCTs ever conducted (n = 2,832) - found that speed-of-processing training (specifically the "useful field of view" task) produced durable benefits in the trained ability and reduced dementia risk at 10-year follow-up. This is a notable positive result, but it applies to a specific type of training (perceptual speed under time pressure) in adults over 65 - not general brain game apps.
What works better than brain games?
The contrast between brain training effect sizes (d ≈ 0.1 for far transfer) and aerobic exercise effect sizes (d ≈ 0.3, or up to 0.5 in older adults) is striking. Physical exercise appears to work through mechanisms that genuinely change the neural substrate - increased BDNF, white matter preservation, improved cerebrovascular health - rather than task-specific skill acquisition.
| Intervention | PS effect size (far transfer) | Evidence quality |
|---|---|---|
| Aerobic exercise (12 wks) | d = 0.28–0.35 | High (meta-analyses) |
| Sleep optimization | d = 0.3–0.5 | High (large RCTs) |
| Speed-of-processing training (ACTIVE) | d = 0.2–0.3 | Moderate (1 large RCT) |
| Dual n-back training | d = 0.1–0.2 | Moderate (inconsistent) |
| Commercial brain game apps | d ≈ 0.05–0.1 | Weak (active placebo controlled) |
For a practical guide to what actually improves processing speed, see our article on how to improve mental processing speed and the best lifestyle habits for cognitive processing. The brain training myth article explores this topic further across cognitive domains.
The verdict: a calibrated assessment
What we can fairly say
- • Brain games reliably produce near transfer - you get better at the game
- • Far transfer to real-world processing speed is small and inconsistent
- • Speed-of-processing training (not general apps) has the best evidence for older adults
- • Aerobic exercise, sleep, and reduced stress produce larger real-world cognitive gains
- • Playing cognitively engaging games is not harmful - opportunity cost is the real concern
Our recommendation: use the Processing Speed test and Reaction Time test as objective benchmarks. If you enjoy brain games, play them - but spend equal or greater time on aerobic exercise, which has far stronger evidence for improving the underlying neural hardware, not just one specific game's score.
Measure real processing speed - not just game skill
Use Human Benchmark as an objective cognitive benchmark separate from any training app.
Take the Processing Speed test
