Brain-training games reliably make people better at brain-training games. What large trials have struggled to find is anything beyond that: the practice rarely spreads to untrained tasks, let alone to work, study or everyday memory.

So, do brain training games work? For the games themselves, yes. For memory, focus and thinking in daily life, the evidence is weak, and the best-known long-term claim, a lower dementia risk after one kind of speed training, is still disputed. The World Health Organization’s 2026 guideline on dementia risk says cognitive training may be offered to older adults without dementia, a conditional recommendation resting on low-certainty evidence that points to a small gain in thinking-test scores and little to no effect on dementia over five years.¹

In practice: play a brain game if you enjoy it, as a game. If you want to remember names, give a better talk or pass an exam, practice that task itself, because across the studies below, practice improves what gets practiced.

Do brain training games work beyond the game itself?

A 2016 review found extensive evidence that brain-training programs raise scores on their own exercises, so the real question is whether the gain spreads to unpracticed tasks, which psychologists call transfer.²

Researchers separate two kinds. Near transfer is a gain on a task that closely resembles the one you practiced, such as a new version of a card-matching game after weeks of card matching. Far transfer is a gain on something different: a reasoning test, a school subject, or remembering what was agreed in a meeting.

In 2014 scientists published two rival open letters, one calling brain games unproven and one defending them. In 2016 Daniel Simons and colleagues reviewed the evidence behind the dispute: the published intervention studies that leading brain-training companies cite on their own websites. Their verdict was graded by distance: plenty of evidence that practice improves the practiced tasks, less for tasks that closely resemble them, and little for unrelated tasks or thinking in daily life. They also found major shortcomings in the design or analysis of many of the studies.²

1. The trained task: extensive evidence that practice improves the exercises people actually do
2. Closely related tasks: less evidence of improvement
3. Distant tasks and everyday thinking: little evidence of improvement

Picture someone who plays a matching game on the train each morning (an illustration, not a study). After a month her scores are far higher. That is real learning, but of the game itself. Whether she now forgets fewer names at work is the question the trials ask, and it is where the evidence thins out.

The practical lesson is to ask one question of any claim that a program is “proven”: proven to improve what? Its own exercises, a lab test, or something you care about in your life?

The BBC trial: big practice gains, no spillover

A large randomized test of everyday brain training, run with viewers of a BBC science program and published in Nature in 2010, found that practice improved every trained task but found no evidence that the gains transferred to untrained tests of thinking, even closely related ones.³

The study

Moderate evidence

Putting brain training to the test: Owen and colleagues, Nature, 2010

Volunteers were randomly assigned to practice reasoning and planning tasks, a broader mix of memory, attention and math tasks like those in commercial games, or a control task of looking up answers to obscure trivia questions online. Both training groups improved sharply on every task they practiced. On four benchmark tests they had not practiced, they improved barely more than the control group, and on two of them numerically less, and the number of sessions people completed made almost no difference. At the rate observed, the authors estimated, gaining one extra digit on a memory-span test would take almost four years of training.³

The caveats are real: the dose was modest, and the authors could not rule out effects from much longer training; nobody over 60 took part; and the volunteers chose to sign up, with most people who registered not finishing.³

If you enjoy a brain game, nothing here says stop; the trial just gives little reason to expect a payoff outside the game.

Working-memory drills: better at the drill, not smarter

Training working memoryworking memory: The small mental workspace that holds and handles the information you are using right now, such as a phone number you are about to dial. Its capacity is limited, which is why juggling several things at once quickly overloads it.Full entry in the glossary, the mind’s small scratch space for holding information while you work on it, is linked to better scores on similar memory tasks but not on intelligence, reading or arithmetic, according to a 2016 meta-analysismeta-analysis: A study that combines the results of earlier studies on the same question into one overall estimate. Pooling makes the estimate more precise, but it cannot repair the studies it pools: a meta-analysis of surveys is still survey evidence.Full entry in the glossary of 87 publications by Monica Melby-Lervåg, Thomas Redick and Charles Hulme. The idea behind it is appealing: working-memory capacity is closely tied to intelligence, so a bigger workspace might lift other skills too. Early studies seemed to agree, but many were tiny and, the authors argue, too small to give reliable answers.⁴ Where the workspace fits is explained in how memory works, from encoding to retrieval.

What decided the question was the comparison group. Against people who did nothing, some studies found gains on nonverbal reasoning tests. Against a control groupcontrol group: The group in a study that does not get the treatment being tested, so its results show what would have happened anyway. It may get nothing, usual care, a placebo, a place on a waiting list or a different activity, and which one it gets shapes what the comparison can show.Full entry in the glossary given some other engaging activity, known as an active control, gains on intelligence, reading, word decoding and arithmetic were not reliably different from zero, either straight after training or at follow-up an average of five months later. Across studies, bigger working-memory gains did not predict bigger gains elsewhere, which undercuts the theory the training rests on.⁴

The authors’ explanation is ordinary: people improve on a memory drill largely by developing strategies for that drill’s particular material, and people who know they are in the “real” training group may simply try harder on the tests afterward.⁴ Someone who spends weeks on a digit game may learn to group digits into pairs. That raises the digit score, but a report or a conversation does not arrive as a string of digits.

Older adults in ACTIVE kept some test gains for a decade

The strongest long-term evidence comes from ACTIVE, a US trial funded by the National Institutes of Health that, from 1998, randomly assigned about 2,800 independent adults aged 65 and over to ten small-group sessions of memory, reasoning or speed-of-processing training, or to no training.⁵

This was not a phone game: trainers taught memory and reasoning strategies in person, and speed training used computer exercises in which people located items flashed for ever shorter times.

1. 1998–1999About 2,800 adults aged 65+ recruited in six US cities
2. 5–6 weeksTen group sessions of memory, reasoning or speed training
3. 11 and 35 monthsFour-session boosters for some trained participants
4. 10 yearsFollow-up of thinking tests and daily tasks
5. 20 yearsUS Medicare records checked for dementia diagnoses

Ten years on, the reasoning and speed groups still scored better than controls on tests of the skills they had trained, while the memory group’s advantage had faded. All three groups reported less difficulty with everyday activities such as preparing meals, handling money and shopping: about 6 in 10 trained participants had held their starting level, against 5 in 10 controls. But on tests that measured everyday tasks directly, training made no clear difference. The daily-life results are self-reports, and fewer than half of participants remained at ten years. The trial was publicly funded, but one author, Karlene Ball, disclosed that she consults for and owns stock in Posit Science, the company that markets the speed-training software.⁵

The lesson is about scope. ACTIVE tested coached, in-person training for people over 65. A 2017 US National Academies report called its evidence encouraging but inconclusive, and found no evidence that its long-term benefits apply to commercial brain-training apps.⁶

Further reading

• 

  The Invisible Gorilla: How Our Intuitions Deceive Us

  by Christopher Chabris and Daniel Simons

  The lead author of the 2016 brain-training review and his co-author explain why quick fixes for the mind rarely deliver.

As an Amazon Associate WiserHours earns from qualifying purchases.

The dementia finding is still disputed

Whether any brain training lowers dementia risk is unsettled. Later analyses of ACTIVE linked speed training, or a subgroup of it, to fewer dementia cases, but the results rest on borderline statistics, and WHO’s 2026 review found that cognitive training may have little to no effect on how many people develop dementia over five years.¹

In a 2017 analysis of the ten-year data, Jerri Edwards and colleagues reported a lower risk of dementia in the speed-training group than in controls, and no clear difference for memory or reasoning training. By our count from the paper’s table, about 8 in 100 of the speed group and 11 in 100 controls met the study’s definition of dementia over ten years (raw counts, ignoring deaths and dropouts). The result only just reached statistical significancestatistical significance: A result that a statistical test says would be unlikely to turn up if there were no real effect. It does not carry the everyday sense of significant: it says nothing about how big the effect is or whether it matters in practice.Full entry in the glossary, dementia was not an outcome the trial was designed to test, and it was defined from test scores and reported diagnoses rather than a doctor’s assessment. Edwards disclosed past consulting for Posit Science and for the company that earlier owned the speed-training software.⁷

A 2026 study led by Norma Coe followed about 2,000 participants through US Medicare billing records to the end of 2019. Taken as whole groups, none of the three kinds of training clearly differed from control. Only a subgroup, speed-trained people assigned to booster sessions, had a clearly lower rate of diagnosed dementia (about 40 in 100 over 20 years, against 49 in 100 controls); boosters went only to people who had attended most of the first sessions, and Ball again disclosed her Posit Science ties.⁸

William Dickens of Northeastern University argued that, once corrected for the number of comparisons, no single result stands out; he read the overall pattern as a shared benefit from all three kinds of training, concentrated among the youngest participants.⁹ The study’s authors replied that each training arm was a separate question set out in advance, so no such correction was needed.¹⁰

Myth

Brain games are proven to protect against dementia.

Fact

That is not established. One US trial's link between speed training and fewer dementia diagnoses is disputed, and WHO's 2026 guideline found cognitive training may have little to no effect over five years.

The question is open. So when an ad tells an aging parent that ten minutes a day will keep dementia away, remember that no app, game or program has yet earned the word “prevents.”

What the Lumosity settlement shows about brain-training claims

In January 2016 the US Federal Trade Commission announced that Lumos Labs, the company behind the Lumosity program, had agreed to pay $2 million to settle charges that its advertising made unfounded claims.¹¹

The FTC alleged that the company claimed its games would improve performance at work, at school and in sport, delay age-related decline and protect against dementia and Alzheimer’s disease. The proposed court order required competent and reliable scientific evidence before any future claim of real-world or health benefits.¹¹

But Lumosity simply did not have the science to back up its ads.

The case was about advertising, not a finding that the games cause harm, but its lesson carries to any app: a claim of better scores on its own games is plausible; a claim of sharper work performance or protection from dementia needs evidence that the large trials above have not provided.

If you enjoy brain games, play them for what they are

To get better at something specific, practice that thing rather than a game. Because near-transfer gains were reliable, Melby-Lervåg and colleagues concluded that training closely resembling the target skill is likely to be more fruitful than generic memory games.⁴

Here is the swap in an ordinary week (an illustration, not a tested plan). Say you give an app ten minutes on the commute. If colleagues’ names slip, spend those minutes recalling who will be at tomorrow’s meeting, then check the list. If a presentation is coming, say its first two minutes out loud. For exams and courses, see study methods built on testing yourself; for concentration at work, ways to focus that have research behind them.

WHO’s 2026 guideline gives everyday mental activity a modest nod: as a conditional recommendation, it says cognitive stimulation, which covers pursuits such as reading, storytelling and playing games, may be encouraged, and it grades that evidence as very low in certainty.¹ For people over 65, a 2020 Cochrane review of computerized training lasting at least 12 weeks found only small benefits straight after training, of uncertain importance and on low-quality evidence, and no evidence that the gain in overall thinking lasted a year.¹²

WiserHours

Before you pay for a brain-training app

• It promises better scores on its own games, not protection from dementia
• Any study it cites compared users with an active comparison group
• I know whether the company funded or ran that research
• I can cancel an automatic renewal easily
• I am choosing it because I enjoy it, not because of a promise

Find this checklist online:
https://wiserhours.com/memory/brain-training-games/#checklist-brain-training-games-before-you-pay

Memory worries belong with a doctor, not an app

Memory problems that affect everyday life call for a doctor, not a training program: the NHS says they can have treatable causes, such as stress, anxiety, depression or sleep problems, and that treatment may work better if it starts early.¹³ The levels below sort UK and US advice by urgency; outside the UK, read GP as your family doctor.

• Now: when anyone becomes confused suddenly, the NHS tells people in the UK to dial 999 or get to A&E, and not to take the wheel themselves.¹⁴ In the US, the National Heart, Lung, and Blood Institute counts sudden confusion or trouble speaking among the warning signs of a stroke and says to call 911 right away.¹⁵ For thoughts of suicide or self-harm, the US 988 Suicide & Crisis Lifeline takes calls, texts and chats¹⁶; in the UK, the NHS points to 999 or A&E when a life is at risk and, in England, to 111’s mental health option for urgent help.¹⁷ Elsewhere, call the local emergency number.
• Soon: see a GP if memory problems are affecting your day-to-day life, the NHS advises, or if an older relative is becoming more forgetful.¹³ The National Institute on Aging in the US lists signs worth taking to a doctor: asking the same questions again and again, getting lost in familiar places, trouble following recipes or directions, growing confusion about time, people and places, and not taking care of oneself.¹⁸
• At your next appointment: occasional forgetfulness and taking longer to learn new things are often a normal part of aging, the National Institute on Aging says, but it still suggests talking with a doctor about noticeable changes in memory.¹⁸

The bottom line

Practice on a brain game buys skill at that game; large trials find little evidence that it reaches everyday memory or work. The long-term dementia benefit linked to one kind of speed training is disputed, and WHO rates the evidence for cognitive training as low in certainty. Play for enjoyment if you like the games, practice the real skill you want to improve, and take memory worries that affect daily life to a doctor.