Brain training games make you better at brain training games. The largest test of them put 11,430 adults through six weeks of online practice. They improved on every task they trained. On four tests they had never practiced, they gained no more than a group that did no training at all (Study). The most thorough review of the field found little sign that the gains reach everyday thinking (Review).
What Brain Training Games Are
They are puzzle sets you subscribe to. Lumosity, BrainHQ, Peak and CogniFit are the names most people meet. You get a few short exercises a day: spot the shape, remember the sequence, tap the odd one out. Lumosity sold 40 games and asked for 10 to 15 minutes, three or four times a week. It cost $14.95 a month, or $299.95 for life (FTC).
The ads rarely sell the puzzles. They sell what the puzzles are meant to buy you: a sharper day at work, fewer forgotten names, a brain that ages more slowly.
One idea decides whether any of that holds up. Psychologists call it transfer, and it comes in two sizes.
Near transfer is the ordinary result of practice. You get better at the task you trained, and at tasks that look almost the same. Nobody disputes this.
Far transfer is what the ads promise. You train a puzzle, and something you never practiced gets better too: your memory for names, your driving, your odds of avoiding dementia. That is the part the evidence keeps failing to find.
Scientists argued about this in public. In 2014 more than 70 signed a letter saying the products do not deliver. Another 133 signed a reply saying they do (Review).
You Get Better at the Game
In 2010 a research team ran the experiment on television. Viewers of a BBC science program trained online for six weeks. 52,617 signed up, and 11,430 took part (Study).
A coin flip decided what each person did. One group trained reasoning, planning and problem solving. A second trained a wider mix, including memory and attention. The third group did no training. They looked up obscure questions with a search engine instead. Everyone was asked for at least 10 minutes a day, three days a week.
Everyone also sat four tests, before and after, that nobody trained on. They covered reasoning, memory for words, spatial memory and paired learning.
Every trained task improved. The four untrained tests came out differently. The group that never trained also improved on all four, by about as much. On two of them it improved the most (Study). The authors found no evidence of transfer, even to closely related tasks.
In 2016 seven researchers went through the field in detail. They graded every published study that the leading brain training companies cited on their own websites. That evidence was the industry’s own. The review ran to 84 pages (Review).
Their verdict came in three layers. Plenty of evidence says training improves the trained tasks. There is less for closely related tasks, and little for distant ones or for everyday thinking (Review).
Many of those studies also had flaws in design or analysis that ruled out firm conclusions (Review).
None of that says the games do nothing. Train on tracking moving dots and you will get better at tracking moving dots. The gain is real, but it stays inside the game.
Why the FTC Stepped In
In January 2016 the Federal Trade Commission charged Lumos Labs, the company behind Lumosity, along with its two co-founders (FTC).
The complaint listed five counts. Four of them covered ad claims the FTC called false or unsupported (Complaint). The ads said the training improved everyday tasks, school, work and sports. They said it delayed age-related decline and protected against mild cognitive impairment, dementia and Alzheimer’s. They said it reduced the impairment caused by stroke, brain injury, PTSD, ADHD and chemotherapy. And they said scientific studies proved all of it.
The fifth count was about testimonials. The site carried 160 of them, and 46 had been gathered through prize contests. The site never mentioned that (Complaint).
The company paid $2 million to settle. The order also set a $50 million judgment, then suspended it, because Lumos Labs could not pay (Order).
The rest of the order shapes what the company can say next time. Lumos Labs and its founders may not repeat those claims about thinking unless human clinical testing backs them. That testing must decide by coin flip who trains and who does not. Qualified researchers must run it (Order). Lumosity also had to email subscribers on auto-renewal and give them a one-step way to cancel.
No court ruled that the games are useless. The case was about the gap between what the studies showed and what the ads claimed.
The Speed Training Exception
One program does have a long-term record. ACTIVE was a trial run in six US cities. It enrolled 2,832 older adults who lived independently, average age about 74. A coin flip sorted them into four groups (Trial). Three groups got ten training sessions each: memory, reasoning, or speed of processing. The fourth group got nothing at all. Later, a second coin flip picked some trained people for refresher sessions.
Speed of processing training is not a bundle of games. It is a single timed visual task. Things flash briefly on a screen, you locate them, and the display keeps speeding up.
Ten years later, the reasoning and speed groups were still ahead on the ability they had trained. The memory group was no longer ahead on memory (Trial). All three training groups also reported less difficulty with everyday jobs such as cooking, managing money and taking medication. At an average age of 82, about 60 in every 100 trained people were still at or above their starting level. Among those who never trained it was about 49 in every 100 (Trial).
There is a catch. Those numbers came from people rating their own difficulty. When researchers watched people do the real tasks, training made no difference at all (Trial).
An independent US review graded the ten-year results as weak evidence. By then 57 in every 100 people had dropped out, most without a recorded reason (Review).
A 2017 analysis then reported fewer dementia cases in the speed group: 59, against 75 among those who never trained (Analysis). That number comes with three caveats. Dementia was never the trial’s main question. The result sat right at the edge where chance could explain it. And no doctor diagnosed those cases: a formula assigned them from test scores, a screening cut-off and diagnoses people reported themselves.
A 2026 follow-up looked again with better records. Researchers linked 2,021 ACTIVE participants to 20 years of Medicare claims (Trial). Speed training on its own leaned the right way, but by too little to rule out chance. The lower diagnosis rate turned up only in people who had also done refreshers.
That refresher group was not the whole trial. Refreshers went only to people who finished at least 8 of the first 10 sessions (Trial). A second coin flip decided who got them, so the comparison is fair. But it covers only the more diligent half of the trial.
A letter in the same journal pushed back harder. The analysis ran six comparisons. Adjust for that, the letter argued, and every result could be down to chance, the refresher one included (Reanalysis).
This is not a clean debunk, and the trial’s own team rejects that correction. The letter’s own models point to a small protective effect shared across all three training types (Reanalysis). So the disagreement is about whether speed training is special, rather than whether training helps at all.
ACTIVE was ten sessions of one timed visual task in a lab, plus refreshers years later. That is a long way from a phone subscription with 40 games on it.
What Newer Reviews Found
The 2016 verdict has held up, and newer work is sharper about why.
In 2023 two researchers pooled 14 separate meta-analyses covering 21,968 people (Meta-analysis). The training covered working memory drills, video games, music, chess and exercise games. At face value, far transfer looked small but positive. Then the authors compared it against groups doing a real alternative activity. They also allowed for the disappointing studies that never got published. Far transfer shrank to nothing, and the training types looked much the same as each other.
Children get the same answer. Across 41 studies in 2,375 children, near transfer shrank as the test moved further from the trained task. Intelligence scores and school results did not move (Meta-analysis).
One review tested the products you can actually buy. It pooled 43 studies in 2,636 older adults across seven programs, including Lumosity, BrainHQ and CogniFit (Meta-analysis). Near transfer showed up. Far transfer appeared only in how people rated themselves. When someone else scored their everyday functioning, it disappeared. In the healthy group, once the unpublished studies were counted in, only processing speed survived.
The friendliest reading of the evidence comes from a 2020 pooling of 16 trials in 1,543 people over 60 (Meta-analysis). Commercial games improved processing speed most. Working memory, planning and word memory improved less. Attention and spatial ability stayed where they were. But every one of those results was a score on a thinking test. Nobody measured everyday life, and nobody checked back once training stopped (Meta-analysis).
One group does better than the rest. In people who already have mild cognitive impairment, computer training improved memory for words and pictures, and working memory too. That came from 28 studies in 1,489 people (Meta-analysis). In dementia, 9 studies in 371 people showed a gain in word memory only. That result was thin enough that it may not hold up. Sessions with a supervisor worked better than sessions alone.
Two limits come with that. Reviewers also looked at everyday tasks in the same people. The gain right after training was small, and by three months it had left no trace (Meta-analysis). A Cochrane review in mild impairment rated the evidence very weak. No trial in it reported who went on to develop dementia (Review).
What Actually Helps Your Brain
In 2024 the Lancet Commission on dementia named 14 risk factors you can do something about. Together they account for about 45% of dementia cases worldwide (Report). That 45% assumes every one of them is wiped out, so treat it as a ceiling.
Hearing loss is joint top of the list at 7% of cases, tied with high LDL cholesterol. Social isolation is joint third at 5%. Physical inactivity accounts for 2% (Report). Brain training does not appear on the list at all.
The same Commission looked at computer training directly. It judged the evidence weak: small gains straight after training, and nothing that lasts. It also noted that the programs can be expensive (Report). What the Commission did link to lower risk was education and mentally demanding work over decades.
So the first two moves are not apps. If you keep turning the television up, book a hearing test, because hearing loss and dementia sit joint top of that list. If your week contains almost no people, that counts too, and the health cost of loneliness is not small.
Exercise has the trial evidence behind it. A 2026 pooling of 50 trials in 4,826 healthy older adults found a small but steady improvement in planning and mental flexibility (Meta-analysis). The dose was ordinary: three sessions a week, under an hour each, at moderate effort, for three to six months. Walking-type exercise worked and so did weights. The authors do warn that smaller studies with nothing to report may be missing.
A comparison of 58 trials in 4,349 adults over 60 ranked resistance training first for overall thinking (Meta-analysis). The schedule that came out best was two 45-minute sessions a week for 12 weeks. Aerobic exercise ranked first for memory. A separate pooling found that with the right type of exercise, the useful dose starts below the standard activity guidelines (Meta-analysis).
In 2025 the US POINTER trial reported its results. A coin flip put 2,111 at-risk adults aged 60 to 79 into a two-year lifestyle program (Trial). One version was structured, the other self-guided. Both groups improved. The structured group improved more on overall thinking.
The structured version asked for a lot. People did aerobic exercise four days a week, plus resistance work and stretching twice a week. They followed the MIND diet. They reviewed their health numbers regularly with a study clinician (Trial). They also went to 38 team meetings over two years, each one with a facilitator (Trial). Brain training on a computer was in there too, at least three sessions a week.
So the training was one ingredient in a package built mostly on movement, food, health checks and showing up for other people. The gain landed in planning and mental flexibility. Memory came out much the same in both groups (Trial). What helped was the whole package.
Frequently Asked Questions
Do brain training games prevent dementia?
No trial has shown that they do. A Cochrane review looked at computer training in people with mild cognitive impairment. Not one trial in it even reported who later developed dementia (Review). The 2024 Lancet Commission judged the evidence weak, with no sign of a lasting effect (Report). One older trial found a signal for a specific speed task, and it remains disputed (Reanalysis).
Is Lumosity a scam?
The FTC charged the advertising, not the puzzles. Lumos Labs paid $2 million in 2016 to settle the case. Its ads had promised better performance at work and school, plus protection against mental decline (FTC). The order bars those claims unless human trials back them (Order). The puzzles still make you better at the puzzles.
Are crosswords and Sudoku better than the apps?
Probably no different. Researchers pooled results across working memory drills, video games, music and chess. In every category, the carryover to untrained abilities came out at zero (Meta-analysis). What the Lancet Commission ties to lower dementia risk is years of education and mentally demanding work (Report). Pick whichever one you will keep doing.
How long do the gains last after you stop?
Mostly nobody has checked. The main pooled analysis of commercial games in older adults measured only the change right after training (Meta-analysis). In people with mild impairment, a small everyday-task gain left no trace at three to five months (Meta-analysis). In ACTIVE, the ten-year gains came with refresher sessions along the way (Trial).
Do the games help if memory is already slipping?
This is where results look best. Across 28 studies in 1,489 people with mild cognitive impairment, computer training improved memory test scores (Meta-analysis). Whether that changes daily life, or lasts once training stops, is unclear (Meta-analysis).
Key Takeaways
- The gains are real but narrow. 11,430 adults trained for six weeks, improved on everything they practiced, and gained nothing on four untrained tests (Study).
- The biggest review found little carryover. Grading the industry’s own cited studies gave plenty of evidence for trained tasks, little for everyday thinking (Review).
- The FTC acted on that gap. Lumos Labs paid $2 million in 2016 and can no longer make those claims without human trials (FTC).
- The one long-term signal is shaky. Over 20 years of Medicare records, fewer dementia diagnoses showed up only in people who also did refresher sessions (Trial).
- Hearing and company beat puzzles. Hearing loss is joint top of the risks you can change, at 7% of cases, and social isolation is joint third at 5%. Brain training is absent (Report).
- The package beats the app. A structured two-year program of exercise, diet, health checks and team meetings improved thinking more than self-guided advice (Trial).
Play Them Because They’re Fun
If you like these games, keep playing them. Enjoying something is reason enough, and nothing here suggests they do harm.
Just do not buy them as insurance. The ads once promised protection from decline, and the evidence never backed that up.
One swap is worth trying this week. Spend the fifteen minutes you would give the app on something from the list above. Two 45-minute strength sessions a week is the schedule that came out best for overall thinking (Meta-analysis). If you have been turning the television up for a year, book the hearing test. If you have not seen a friend in a month, make the call.
Brain training buys you a better score on the game. The evidence for a better life points to the things that get you moving, hearing and talking to people.
This article is for educational purposes and is not medical advice. Talk to a qualified clinician before changing your health regimen.
Bio-Hacking.Blog 
Leave a comment