Brain games make you better at brain games, and little else

The renewal notice lands a week before your subscription rolls over. It comes with a chart: your speed, memory and attention scores, all trending up since you started playing the app’s games over morning coffee. The improvement is real. You really are faster at those games.

What the chart can’t show is whether anything else changed. In 2016, the Federal Trade Commission charged the maker of a popular brain-training app with deceiving customers about exactly that. The settlement carried a $50 million judgment, suspended because of the company’s finances once it paid $2 million. Behind that case sits more than 15 years of research asking one question: does getting better at brain games make you better at anything else?

The short answer from the largest and most independent tests is no, or not much. Here is what those tests found, how regulators have treated the claims, and what to ask before you pay.

Your scores are rising, and that part is real

Researchers separate two kinds of gain. Near transfer means improving on the trained game or on tasks built much like it. Far transfer means improving on different skills, or in daily life: remembering names at a party, keeping track of a work project, staying sharp into your 80s. The sales pitch depends on the second kind.

The cleanest test of the difference was published in 2010, run with help from a BBC science program in Britain. A team at the Medical Research Council’s Cognition and Brain Sciences Unit in Cambridge, led by Adrian M. Owen, signed up viewers aged 18 to 60. Of 52,617 who registered, 11,430 completed the before-and-after testing and at least two training sessions over six weeks. They were asked to train for at least 10 minutes a day, three times a week.

One group practiced reasoning, planning and problem-solving tasks. A second practiced memory, attention, visual-spatial and math tasks similar to those in commercial products. A control group spent its sessions answering obscure trivia questions with help from the internet.

• On the trained tasks, gains were large in both training groups, effect sizes of roughly 0.7 to 1.6. By the usual convention, 0.8 or more counts as large.
• On four untrained benchmark tests, changes were tiny, and the trivia group improved about as much. On two of the tests, the trivia group’s gain was the largest.
• More sessions didn’t help. The number of sessions tracked closely with improvement on the trained games and hardly at all with benchmark gains.

The researchers offered one example of scale. In the group that practiced memory-style games, the number of digits people could hold in mind rose by three hundredths of a digit. At that rate, they calculated, gaining one extra digit would take almost four years of training. The trivia group gained two tenths of a digit without practicing memory at all. The Cambridge team wrote that “these results confirm that six weeks of regular computerised brain training confers no greater benefit than simply answering general knowledge questions using the internet.”

So a rising in-app score is evidence that you’ve learned the game, not that your brain works better elsewhere. Owen’s study tested adults no older than 60 for six weeks, so it can’t settle questions about older adults or longer training. Later research took those up.

The promise of sharper thinking: what pooled studies found

Since then, research teams have pooled hundreds of training experiments, and one variable keeps turning up: what the comparison group did.

Studies that compare trainees with people who did nothing at all are the ones most likely to report broad benefits. Studies that give the comparison group something engaging to do, such as trivia, puzzles or educational videos, mostly don’t. That matters because a do-nothing group can’t separate the training from simply practicing tests, expecting to improve and getting regular attention.

• Working memory training. A 2016 analysis by Monica Melby-Lervåg, Thomas S. Redick and Charles Hulme pooled 87 publications with 145 comparisons. Training reliably improved working-memory measures in the short term. Against controls who also did something, there was no convincing improvement in nonverbal or verbal ability, reading or arithmetic.
• Children. A 2020 analysis of 41 studies with 2,375 typically developing children, by Giovanni Sala and Fernand Gobet, found small to medium gains on memory tasks. Gains in intelligence and school achievement were essentially zero when studies used active controls. A 2019 analysis of children’s executive-function training put near transfer at 0.44 and far transfer at 0.11, which was not statistically significant.
• Video games. A 2018 set of three meta-analyses by Sala and colleagues, covering more than 300 comparisons each, found small or null effects and no evidence that playing video games causes better cognitive ability.

Sala and Gobet have since argued, in a 2023 position paper, that the overall far-transfer effect of cognitive training is null. That is the stated position of two of the field’s most prominent skeptics, though it fits the pooled numbers above.

In the library’s evidence vocabulary, near transfer is real. Far transfer to everyday life has been tested and not supported.

A 2026 test: the app against free puzzles

The most recent large test pitted a commercial adaptive training program against plain computer games. It ran as part of RECOVER, the National Institutes of Health research effort on long COVID, and was published in JAMA Neurology in 2026, led by David S. Knopman of the Mayo Clinic.

The trial enrolled 328 adults with thinking problems after COVID-19, median age 48, at 22 US sites. For 10 weeks, five times a week, they did 30-minute sessions. One arm used the adaptive program, which grows harder as you improve. The comparison arm played eight online puzzles and games, supplied by the same company, that lacked the adaptive engine. The designers chose that comparison on purpose: a program worth its added cost and effort, they reasoned, should beat something cheap or free.

It didn’t. On the main measure, participants’ own ratings of everyday thinking, the difference between the program and the puzzles was 0.0. Nor did any arm, including ones that added rehabilitation sessions or brain stimulation, beat its comparison on secondary questionnaires or neuropsychological tests. Every arm improved by about half a point on the five-point scale, and 74% of all participants said their assigned activity had helped their thinking, including about two thirds of the puzzle group. The author list includes two people affiliated with the program’s maker. The trial’s population was specific, and more than half had no measurable impairment on testing, so it is one data point. Thinking problems after COVID-19 themselves are taken up in the library’s brain fog article.

The promise of dementia protection: what reviews found

The most consequential claim is protection against memory loss. Cochrane, the international network that pools clinical trials under standard rules, has reviewed computerized training in three groups.

Group	What Cochrane found	Quality of evidence
Healthy adults 65 and older (2020)	8 trials, 1,183 people, most trained only 12 or 13 weeks. Against active comparisons such as educational videos, a slight gain in overall thinking at the end of training, from 2 trials. Little or no difference in memory. No data on daily activities or quality of life.	Low or very low
Healthy adults in midlife (2019)	One eligible study, 6,742 people over 50, six months, with heavy dropout. No data on overall thinking.	Low
People with mild cognitive impairment (2019)	8 trials, 660 people. Not one reported whether anyone developed dementia.	Very low

The reviewers called the late-life gains of uncertain clinical importance. In the impairment review, they judged direct harms unlikely but noted that the hours, and sometimes the fees, can weigh heavily.

So on the library’s scale, the short-term gain in healthy older adults is measured but below the threshold that matters, and dementia prevention by consumer apps is insufficient evidence, since the trials in people with mild cognitive impairment never measured it.

Two related bodies of work sit elsewhere. A 2021 meta-analysis of commercial programs in older adults, and its split between near and far transfer, is examined in the library’s article on dementia risk factors. The 20-year ACTIVE trial of in-person training, including its 2026 dementia-diagnosis results, is covered in the library’s article on structured brain exercise.

Two letters, two consensus claims

In October 2014, the Stanford Center on Longevity and Berlin’s Max Planck Institute for Human Development published a statement signed by more than 70 scientists. Brain-game marketing, it said, often exaggerates. Gains tend to be small, narrow and sometimes short-lived. It singled out, as perhaps the most damaging, the claim that games prevent or reverse Alzheimer’s disease, and said no study had shown that playing them cures or prevents Alzheimer’s or other dementias. It also warned that a single study by researchers with a financial stake does not show a product works, and that hours of solo drills come at the expense of exercise, reading or time with people.

Months later, 133 scientists and practitioners signed a reply. It said the statement was wrong to claim consensus and that dozens of randomized trials showed specific training benefits, some reaching everyday life.

The overlap is easy to miss. The reply agreed that marketing is often exaggerated, that programs should face independent, randomized trials against active controls, and that no study has shown brain training prevents or cures Alzheimer’s.

In 2016, seven researchers led by Daniel J. Simons of the University of Illinois tried to settle the dispute. They examined every peer-reviewed intervention study cited on the websites of leading brain-training companies, as listed on the proponents’ own site. The Illinois-led team found extensive evidence of improvement on trained tasks, less on closely related tasks, and “little evidence that training improves performance on distantly related tasks or that training improves everyday cognitive performance.” None of the cited studies met all the best practices the team set out.

Three federal cases and the line FDA draws

The enforcement cases against brain-training marketing have come from the FTC, which polices deceptive advertising. The evidence standard the agency applies is explained in the library’s coverage of memory-supplement advertising enforcement.

• A children’s attention game, January 2015. Its makers promised permanent gains in focus, memory, behavior and school performance, including for children with ADHD, and called the results scientifically proven. The game sold for $214.75 and brought in about $4.5 million from 2012 to mid-2013. The proposed settlement barred such claims without scientific support.
• A popular subscription app, January 2016. The FTC said the company claimed its games improved school, work and athletic performance, delayed age-related decline and protected against dementia and Alzheimer’s, and eased problems linked to stroke, brain injury, PTSD, ADHD and chemotherapy. Some website testimonials, the agency said, came from contests offering prizes such as a tablet computer or a trip, without saying so. Subscriptions ran from $14.95 a month to $299.95 for life. Besides the $2 million payment, the order required the company to notify auto-renewal subscribers from 2009 through 2014 and give them an easy way to cancel.
• A franchised one-on-one program, May 2016. With more than 80 centers in 25 states, it allegedly claimed clinically proven, permanent improvement in ADHD, autism, dementia, stroke and concussion, plus better grades and earnings. Its search ads were keyed to phrases about curing autism and ADD. A $4 million judgment was suspended on payment of $200,000, and future claims about cognitive benefits for people with health conditions must rest on human clinical testing.

What FDA does and doesn’t review

FDA’s general wellness guidance, updated at the start of 2026, gives sharper concentration, mental acuity and problem-solving, along with greater learning capacity, as examples of wellness claims. For low-risk products whose promises stop there, the agency doesn’t plan to step in. A game marketed to diagnose or treat autism, by the guidance’s own example, falls outside that policy.

Crossing that line means FDA review. In June 2020 the agency authorized a prescription video game for children aged 8 to 12 with inattentive or combined-type ADHD, to be used as part of a broader treatment program rather than alone. Its trial of 348 children succeeded on its main measure, a computerized attention test, with the game group improving 0.93 points against 0.03. On parent and clinician ADHD rating scales, the results leaned toward the game, but FDA’s review found no statistically meaningful differences. Reported side effects included lower frustration tolerance and headache. Prescription apps as a category get a fuller look in the library’s piece on apps that come by prescription.

The American Academy of Pediatrics’ 2019 ADHD guideline, meanwhile, groups cognitive training with mindfulness, diet changes and EEG biofeedback as approaches whose evidence is too thin, or whose measured benefit too small, to recommend.

When an app is the wrong tool

Direct harms are unlikely, as Cochrane’s reviewers noted. The indirect costs are money, time and false reassurance. The 2014 statement’s point about time applies here too; what trials found for exercise and for social connection is covered in the library’s articles on each.

• Changes in memory or thinking that you, or people close to you, have noticed. Whether they need a workup is a clinician’s call to make with you; an app can’t answer it. The library’s guide to memory assessment explains how evaluation works.
• A child with attention or learning problems. Whether ADHD or something else is behind them is a pediatric clinician’s question, and pediatric guidance doesn’t back cognitive training as an ADHD treatment.
• A diagnosis of mild cognitive impairment. Whether any training program fits into the care plan is for the clinician who manages the diagnosis to weigh, especially since no trial in that group has measured dementia.
• Rebuilding thinking skills after a stroke or brain injury. That is a question for the rehabilitation team; the library’s coverage of lingering concussion symptoms explains how guidelines treat self-directed computer programs.
• Sudden confusion, or sudden trouble speaking or understanding others. These can signal a stroke, and the CDC’s advice is to call 911 at once rather than drive to the hospital.

Questions to ask before you subscribe

We built this list from the trials and cases above.

1. Better at what? If the evidence shows gains only on the app’s own games, that’s near transfer.
2. Compared with what? A study where the comparison group did nothing can’t separate training from expectation and practice. Ask whether the program beat puzzles, trivia or videos.
3. Who ran the study? The 2014 scientists’ statement warned against relying on single studies by researchers with a financial stake.
4. Did anyone measure daily life? Ask whether a trial tracked work, driving, independence or dementia diagnoses rather than test scores.
5. How does canceling work? Check for automatic renewal, lifetime-plan terms and refund rules before entering a card number.
6. Does it claim to treat a condition? A treatment claim for ADHD, autism or dementia puts a product in FDA’s territory. Ask what it was authorized for and what its trial measured.
7. What happens to your data? That question gets fuller treatment in the library’s article on health-data privacy.

The bottom line

Practice makes you better at what you practice, and brain-training apps deliver that reliably. What large independent tests have not found is spillover into reasoning, school, work or protection against dementia, and the strongest claims of that kind drew federal enforcement. If you enjoy the games, the trials point to few direct harms; the costs are time and money. If you’re paying for a sharper mind, the most useful question is the one the 2026 trial asked: does this beat a free puzzle?

Sources

1. Federal Trade Commission. “Lumosity to Pay $2 Million to Settle FTC Deceptive Advertising Charges for Its 'Brain Training' Program (press release, January 5, 2016).” ftc.gov
2. Federal Trade Commission. “Marketers of One-on-One 'Brain Training' Programs Settle FTC Charges That Claims about Ability to Treat Severe Cognitive Impairments Are Unsupported (press release, May 18, 2016).” ftc.gov
3. Federal Trade Commission. “Makers of Jungle Rangers Computer Game for Kids Settle FTC Charges that They Deceived Consumers with Baseless 'Brain Training' Claims (press release, January 20, 2015).” ftc.gov
4. U.S. Food and Drug Administration. “General Wellness: Policy for Low Risk Devices (guidance, January 6, 2026).” fda.gov
5. U.S. Food and Drug Administration, Center for Devices and Radiological Health. “De Novo Classification Request for EndeavorRx (DEN200026 decision summary; 21 CFR 882.5803, digital therapy device for attention deficit hyperactivity disorder).” accessdata.fda.gov
6. U.S. Food and Drug Administration. “De Novo database record DEN200026 (received April 16, 2020; granted June 15, 2020).” accessdata.fda.gov
7. Stanford Center on Longevity and Max Planck Institute for Human Development. “A Consensus on the Brain Training Industry from the Scientific Community (Full Statement), October 20, 2014.” longevity.stanford.edu
8. Cognitive Training Data. “Response Letter (reply to the October 2014 Stanford statement; undated).” cognitivetrainingdata.org
9. Centers for Disease Control and Prevention. “Signs and Symptoms of Stroke.” cdc.gov
10. Nature. “Owen AM, Hampshire A, Grahn JA, Stenton R, Dajani S, Burns AS, Howard RJ, Ballard CG. "Putting brain training to the test." Nature, 2010. PubMed. DOI: 10.1038/nature09042.” pubmed.ncbi.nlm.nih.gov
11. Psychological Science in the Public Interest. “Simons DJ, Boot WR, Charness N, Gathercole SE, Chabris CF, Hambrick DZ, Stine-Morrow EAL. "Do 'Brain-Training' Programs Work?" Psychol Sci Public Interest, 2016. PubMed. DOI: 10.1177/1529100616661983.” pubmed.ncbi.nlm.nih.gov
12. Perspectives on Psychological Science. “Melby-Lervåg M, Redick TS, Hulme C. "Working Memory Training Does Not Improve Performance on Measures of Intelligence or Other Measures of 'Far Transfer': Evidence From a Meta-Analytic Review." Perspect Psychol Sci, 2016. PubMed. DOI: 10.1177/1745691616635612.” pubmed.ncbi.nlm.nih.gov
13. Psychonomic Bulletin & Review. “Sala G, Gobet F. "Working memory training in typically developing children: A multilevel meta-analysis." Psychon Bull Rev, 2020. PubMed. DOI: 10.3758/s13423-019-01681-y.” pubmed.ncbi.nlm.nih.gov
14. Psychological Bulletin. “Kassai R, Futo J, Demetrovics Z, Takacs ZK. "A meta-analysis of the experimental evidence on the near- and far-transfer effects among children's executive function skills." Psychol Bull, 2019. PubMed. DOI: 10.1037/bul0000180.” pubmed.ncbi.nlm.nih.gov
15. Psychological Bulletin. “Sala G, Tatlidil KS, Gobet F. "Video game training does not enhance cognitive ability: A comprehensive meta-analytic investigation." Psychol Bull, 2018 (online 2017). PubMed. DOI: 10.1037/bul0000139.” pubmed.ncbi.nlm.nih.gov
16. Perspectives on Psychological Science. “Gobet F, Sala G. "Cognitive Training: A Field in Search of a Phenomenon." Perspect Psychol Sci, 2023 (online 2022). PubMed. DOI: 10.1177/17456916221091830.” pubmed.ncbi.nlm.nih.gov
17. JAMA Neurology. “Knopman DS, Koltai D, Laskowitz DT, et al. "Evaluation of Interventions for Cognitive Symptoms in Long COVID: A Randomized Clinical Trial." JAMA Neurol, 2026;83(1):49-59 (online 2025). PubMed. DOI: 10.1001/jamaneurol.2025.4415.” pubmed.ncbi.nlm.nih.gov
18. Cochrane Database of Systematic Reviews. “Gates NJ, Rutjes AWS, Di Nisio M, et al. "Computerised cognitive training for 12 or more weeks for maintaining cognitive function in cognitively healthy people in late life." Cochrane Database Syst Rev, 2020. PubMed. DOI: 10.1002/14651858.CD012277.pub3.” pubmed.ncbi.nlm.nih.gov
19. Cochrane Database of Systematic Reviews. “Gates NJ, Rutjes AWS, Di Nisio M, et al. "Computerised cognitive training for maintaining cognitive function in cognitively healthy people in midlife." Cochrane Database Syst Rev, 2019. PubMed. DOI: 10.1002/14651858.CD012278.pub2.” pubmed.ncbi.nlm.nih.gov
20. Cochrane Database of Systematic Reviews. “Gates NJ, Vernooij RWM, Di Nisio M, et al. "Computerised cognitive training for preventing dementia in people with mild cognitive impairment." Cochrane Database Syst Rev, 2019. PubMed. DOI: 10.1002/14651858.CD012279.pub2.” pubmed.ncbi.nlm.nih.gov
21. Pediatrics. “Wolraich ML, Hagan JF, Allan C, et al. "Clinical Practice Guideline for the Diagnosis, Evaluation, and Treatment of Attention-Deficit/Hyperactivity Disorder in Children and Adolescents." Pediatrics, 2019. PubMed. DOI: 10.1542/peds.2019-2528.” pubmed.ncbi.nlm.nih.gov