Brain-training games reliably improve one thing: your score on that game

If you practise a cognitive game, you get better at that game. That much is not in dispute, and it is most of what the science of commercial "brain training" actually establishes. The harder claims — that playing improves memory, work or driving in general, or delays the cognitive decline of ageing — are far weaker, and in one prominent case the U.S. Federal Trade Commission ruled that a leading company had no science to support them [s3].

The evidence divided the field in two

The state of the evidence was captured, unusually clearly, by two duelling open letters in 2014. A consensus statement from more than 70 scientists said brain games do not provide a scientifically grounded way to improve cognition or stave off cognitive decline; months later, a group of 133 scientists and practitioners countered that the literature was full of demonstrated benefits [s1]. Two groups, the same studies, opposite conclusions.

A comprehensive review in Psychological Science in the Public Interest set out to referee the dispute by grading the studies the brain-training companies themselves cited as their best evidence, against a defined set of methodological best practices [s1]. Its finding was a graded one. There was extensive evidence that training improves performance on the trained tasks; less evidence that it improves closely related tasks; and little evidence that it enhances distantly related tasks or improves everyday cognitive performance [s1]. The review also found that none of the cited studies met all the best practices needed to draw clear conclusions about real-world benefit [s1].

This is the "transfer" problem, and it is the whole game. Getting faster at a matching puzzle is only useful if that speed carries over — "transfers" — to remembering names, managing medications or reacting at the wheel. The evidence for near transfer is real but modest; the evidence for the far transfer the marketing implies is where it thins out to almost nothing [s1].

The strongest trial, honestly read

The most cited long-term evidence is the ACTIVE trial, which is genuinely good and often misquoted in both directions. ACTIVE randomised 2,832 older adults, mean baseline age 73.6, across six U.S. cities to training in memory, reasoning, or speed of processing — just ten sessions, with a few booster sessions — or to a no-contact control group, and followed them for a decade [s2].

At ten years, two of the three trained skills had held up. The reasoning intervention maintained an effect on reasoning ability of 0.23 standardised units, and the speed-of-processing intervention maintained an effect of 0.66 units on its target ability [s2]. Memory training, by contrast, no longer showed a maintained effect on memory performance at ten years [s2]. That skills trained in a handful of sessions were still measurable a decade later is a real result.

The more consequential finding is about everyday function. Participants in each training group reported less difficulty with instrumental activities of daily living — the errands-and-finances tasks that let people live independently — and at a mean age of 82, roughly 60% of trained participants were still at or above their baseline level of self-reported daily function, versus about 50% of controls [s2]. That is a transfer signal, from a rigorous trial. But it rests substantially on self-report, it is a difference of about ten percentage points, and ACTIVE tested a structured research protocol delivered by trainers — not a consumer app played casually at home. Crucially, ACTIVE was not designed to test, and did not establish, that training prevents dementia.

Where the regulator drew the line

The gap between "trained skills can persist" and "this product protects your brain" is exactly where the FTC intervened. In January 2016, the maker of Lumosity agreed to pay $2 million to settle charges that it had deceptively advertised its games as able to delay age-related cognitive decline and protect against mild cognitive impairment, dementia and Alzheimer's disease — claims the FTC said the company simply did not have the science to support [s3]. The complaint's language is worth keeping in view: the agency said the company had "preyed on consumers' fears" about cognitive decline [s3].

The honest summary

The defensible claims about brain training are narrow. You will improve at what you practise, and structured cognitive training in specific abilities such as processing speed can produce effects that last years and may modestly support everyday function, as ACTIVE showed [s2]. The claims that do not hold up are the ones that sell the products: that playing branded games transfers broadly to real-world cognition, or that it prevents the diseases of ageing [s1] [s3]. A crossword, a language app or a new instrument will make you better at crosswords, that language or that instrument — and each is a reasonable way to spend time. Whether any of them changes the trajectory of your brain is a claim the best evidence does not currently support.

This article is informational and is not medical advice.