Does Brain Training Actually Work? What the Science Says

Near transfer refers to improvement on tasks closely related to what was trained — train digit span, improve on digit span. Far transfer refers to improvement on unrelated abilities, such as reasoning, academic achievement, or real-world executive function. Near transfer is reliable across virtually every cognitive training study. Far transfer is where the scientific consensus is far more sceptical.

The distinction became publicly prominent in 2014 when an open letter signed by 75 cognitive scientists — published through Stanford — argued that commercial brain training companies were making far transfer claims that the peer-reviewed literature did not support. The letter was a direct response to the marketing of products like Lumosity and CogniFit, which implied that training specific cognitive games would improve general intelligence, prevent dementia, or enhance academic performance. The scientists were not arguing that cognitive training was useless — only that the far transfer claims were overstated relative to the evidence at the time.

The most comprehensive independent review of the brain training literature is Simons et al. (2016), published in Psychological Science in the Public Interest. The team reviewed hundreds of studies and concluded that near transfer is reliable and consistent across programmes, but far transfer evidence — the claim that brain games make you measurably smarter or more capable in everyday life — is weak to nonexistent for most commercial programmes. Methodological problems were common: lack of active control groups (comparing brain games to watching TV rather than to other mentally stimulating activities), short follow-up periods, and publication bias toward positive results.

Crucially, the same review identified physical exercise as one of the few interventions with genuine far transfer to cognitive ability — specifically to processing speed and executive function — with effect sizes meaningfully larger than those seen for cognitive training tasks. The implication is uncomfortable for the brain training industry: a 20-minute jog may do more for general cognitive performance than 20 minutes of targeted memory games.

Setting aside far transfer, three clinical contexts show strong evidence for meaningful brain training benefits. First, age-related cognitive decline: the ACTIVE trial (Advanced Cognitive Training for Independent and Vital Elderly) enrolled 2,832 participants aged 65 and older and followed them for 10 years. Participants who received cognitive training maintained specific trained abilities and showed reduced dementia risk signals compared to controls. The effects were task-specific but real and durable. Second, ADHD: working memory training, particularly Cogmed-style programmes, reduces working memory symptoms and, in children, reduces inattentive behaviour — though pharmacological treatment remains substantially more effective for most individuals.

Third, stroke rehabilitation: neuroplasticity research is clearest in the context of post-stroke recovery, where targeted retraining of damaged functions exploits the brain's capacity to recruit adjacent circuits for lost functions. Here, the specificity of training is a feature: you train the lost ability directly, and the brain rebuilds the circuit that was damaged. The mechanism of plasticity driving this is well-understood and the clinical outcomes are robust.

A persistent methodological problem in brain training research is the difficulty of controlling for general mental engagement. Studies rarely pit brain training against an equally engaging but cognitively demanding alternative — such as learning a musical instrument, studying a new language, or solving crosswords. When such comparisons are made, the advantages of commercial brain training programmes tend to shrink or disappear. This suggests that what matters may be sustained, effortful mental engagement itself, rather than the specific cognitive demands of the trained task.

This creates an important reframe: the question may not be 'does brain training work?' but 'what type of mental engagement is most efficient for which goal?' The answer likely depends on the specific ability you want to improve. If you want a faster visual reaction time, targeted reaction practice will beat general mental stimulation. If you want broadly preserved cognition into old age, the evidence favours the same lifestyle factors that benefit the rest of the body: movement, sleep, social connection, and continuous learning.

Intellectual honesty requires acknowledging what these games reliably do and don't do. They reliably improve the specific skills they target: playing Reaction Grid improves visual reaction time; playing Number Memory improves digit span; playing Pattern Recall improves visuospatial sequence memory. Near transfer is real. They also build the habit of daily cognitive engagement — which matters, because sporadic intense sessions are less effective than consistent practice. And they provide objective measurement: your scores are numbers, not impressions, which means you can track real change over time rather than wondering if you're improving.

What they won't do is make you measurably more intelligent in unrelated domains or prevent dementia through some general brain-strengthening mechanism. The goal is cognitive fitness — maintaining and improving specific cognitive capacities that degrade with disuse and age. That is a worthwhile goal on its own terms, without requiring inflated claims about general intelligence enhancement.