Do Brain Games for Seniors Work? What the ACTIVE Trial Found
No brain game is proven to prevent dementia. In the 10-year ACTIVE trial of 2,802 healthy older adults, speed training had a hazard ratio of 0.71 (0.50 to 0.998); memory and reasoning training did not show a clear effect.
What you need to know
- Brain games for seniors are not proven to prevent dementia. In the ACTIVE randomised trial (Edwards and colleagues, 2017; 2,802 healthy adults aged 65 and over; 260 dementia cases in 10 years), speed-of-processing training had a hazard ratio of 0.71 (95% interval 0.50 to 0.998).
- Memory training (0.79, 0.57 to 1.11) and reasoning training (0.79, 0.56 to 1.10) showed no clear difference from no training. The trial did not test crosswords, sudoku or commercial apps.
- The authors list limits: dementia was not the trial's primary outcome, it was not a clinical diagnosis, 938 people dropped out and 627 died, and they call for replication. No brain game is proven to prevent dementia.
- Our eight Brain Games each work one skill, are not the ACTIVE software, and make no dementia or IQ claim. A game score is not a diagnosis.
Brain games for seniors are not proven to prevent dementia, but one large trial linked one kind of training to a lower dementia risk after 10 years. The ACTIVE trial (Edwards and colleagues, 2017) followed 2,802 healthy adults aged 65 and over. The speed-of-processing group had a hazard ratio for dementia of 0.71 (95% interval 0.50 to 0.998). The memory and reasoning groups each had 0.79, with intervals that included 1, which means no clear difference from no training.
Below: which kinds of training were tested, how big the effect was and how far to trust it, what crosswords and apps have to do with it, what our own Brain Games train, and when a memory problem is a matter for a doctor. We read the full paper on 12 November 2026.
Do brain games for seniors work?
It depends on what you mean by work, and the trial answers only one version of the question. People search for brain games for older adults, elderly brain games, brain training for elderly relatives and brain training exercises for seniors, and they usually want one of three things. They want to get better at the game, to cope better with daily tasks, or to lower the chance of dementia. The trial speaks only to the last two, and weakly to the third.
The paper's introduction says cognitive training improves cognitive performance and delays functional impairment, but that its effects on dementia were not known. The authors add that multiple randomised trials indicate speed training improves everyday functioning. The dementia result is newer and weaker, because dementia was not the primary outcome of ACTIVE, so these results come from secondary analyses. We cover the wider evidence on brain games in our guide to whether brain training apps work, and do not retell it here.
Which kind of training had trial evidence, and which did not?
Three programs were tested, and only the speed program showed a clear signal. Participants were community-dwelling adults aged 65 and older, with an average age of 73.6 years (SD 5.9) and an average Mini-Mental State Examination score of 27.3. They were randomised to one of four groups.
- Memory training: instruction and practice in strategies for verbal episodic memory
- Reasoning training: instruction and practice in strategies for problem-solving and serial patterns
- Speed-of-processing training: computerised visual exercises designed to raise how much, and how complex, information a person can process quickly
- No-contact control: no training at all
Each training group had up to 10 sessions over 6 weeks. A subset also received up to four booster sessions at 11 months and up to four more at 35 months. People were assessed at 1, 2, 3, 5 and 10 years. The speed software was a specific computerised program aimed at improving the useful field of view. No result transfers automatically to other apps.
How big was the effect in the ACTIVE trial?
Over 10 years the trial recorded 260 cases of dementia, and speed training was linked to a 29% lower risk. The speed group's hazard ratio was 0.71 (95% interval 0.50 to 0.998, P = .049) against the control group. A hazard ratio compares how fast new cases appear in two groups, and 1.0 means no difference. The top of the interval, 0.998, is almost 1.0, so the data are also compatible with almost no effect.
- Speed of processing: hazard ratio 0.71, interval 0.50 to 0.998, P = .049
- Memory: hazard ratio 0.79, interval 0.57 to 1.11, P = .177
- Reasoning: hazard ratio 0.79, interval 0.56 to 1.10, P = .163
- No-contact control: the comparison group
The authors also found that each extra speed session attended went with a 10% lower hazard (unadjusted hazard ratio 0.90, interval 0.85 to 0.95). They moderated that finding as an association, because people who attended more sessions may differ from those who attended fewer. A later follow-up linked 2,021 trial members to Medicare claims for 20 years. Coe and colleagues (2026) report a hazard ratio of 0.75 (interval 0.59 to 0.95) for speed training with one or more boosters, and 1.01 (0.81 to 1.27) without boosters, with no main effect of memory or reasoning training.
The paper lists its limits. Dementia was defined from interviews and performance measures, not a clinical diagnosis. Of the 2,785 people in the analysis, 1,220 completed the 10-year follow-up, 627 died and 938 dropped out. The method of choosing who received boosters opens a range of interpretations. The authors call for replication with a clinical diagnosis. The paper also discloses that one author consulted from 1996 to 2005 for the company that owned the speed software. The authors state they have no conflicts of interest relevant to these analyses.
Do crosswords, sudoku and apps count?
The trial says nothing about them, because it tested three specific programs and not crosswords, sudoku or commercial apps. A claim that your favourite puzzle protects the brain needs its own trial. Our article on whether Sudoku increases IQ covers what studies show for one popular puzzle.
If you are choosing brain games for seniors, pick games you will keep playing, and use the ones without a clock if speed stresses you. That is a practical preference, not a medical claim. The trial gives no dose for any commercial game, and it gives no reason to expect the same result from a different program or a different group of people.
Other lifestyle factors have their own research. See our pieces on multilingualism and brain aging and hearing loss and cognitive test scores. None of them is a training program, and none of them is a promise.
In ACTIVE, 2,802 healthy older adults, 260 dementia cases and 10 years: only speed training showed a clear signal (hazard ratio 0.71), and its interval ran up to 0.998.
What do our Brain Games train, and what do they not claim?
Each of our Brain Games works one skill, and none is the ACTIVE software. The Memory Game, Digit Span Test and Sequence Memory Test work memory. The Stroop Test and Reaction Time Test work attention and speed. Logic Puzzles, the Tower of Hanoi and Code Breaker work planning and logic. Logic Puzzles and the Tower of Hanoi have no time limit and give hints.
Our games need no signup, and scores stay in your browser, so you can compare today with your own past rounds. Reaction time slows with age for everyone, as our page on average reaction time by age shows. We make no claim that any of our games lowers dementia risk or raises IQ. A game is practice for one skill, and it does not measure a cognitive index.
When is a memory problem a reason to ask a doctor?
When memory problems get in the way of daily life, ask a doctor. A game score is not a diagnosis, and neither is an online IQ score. Clinicians use short screening tools, and our guide to the MoCA score explains what the best-known one reports. We also answer whether an IQ test can detect dementia.
To put a cognitive score in context, our IQ percentile calculator converts an IQ to a percentile. It is not a screen for memory loss.
Where would your own score land?
Curious how your reasoning compares with other adults? The IQ Metrics test reports a result with its percentile.
Find your IQ score now! →For a result on the standard scale, take the IQ Metrics test.
Common questions
Do brain games for seniors work?
They are not proven to prevent dementia. In the ACTIVE trial of 2,802 healthy adults aged 65 and over, only speed-of-processing training was linked to lower dementia risk over 10 years (hazard ratio 0.71, interval 0.50 to 0.998), and dementia was not the trial's primary outcome.
Which brain training had evidence in older adults?
In ACTIVE, speed-of-processing training, a computerised visual program. Memory training (hazard ratio 0.79) and reasoning training (0.79) showed no clear difference from no training.
Do crosswords and sudoku prevent dementia?
The ACTIVE trial did not test them, so it cannot say. A claim about a puzzle needs its own trial.
When should an older adult ask a doctor about memory?
When memory problems affect daily life. A game or online score is not a diagnosis, and doctors use clinical screening tools such as the MoCA.
Sources for this story
- Edwards JD, Xu H, Clark DO, Guey LT, Ross LA, Unverzagt FW. Speed of processing training results in lower risk of dementia. Alzheimer's & Dementia: Translational Research & Clinical Interventions 3(4):603-611 (2017), doi 10.1016/j.trci.2017.09.002; full text read — Wiley / Alzheimer's Association
- Coe NB, Miller KEM, Sun C, Taggert E, Gross AL, Jones RN, Felix C, Albert MS. Impact of cognitive training on claims-based diagnosed dementia over 20 years: evidence from the ACTIVE study. Alzheimer's & Dementia: Translational Research & Clinical Interventions 12(1) (2026), doi 10.1002/trc2.70197; abstract read — Wiley
- Brain Games hub and the Memory Game and Sequence Memory Test pages, read 12 November 2026 — IQ Metrics
- Brain training apps: do they work? (our guide to the wider evidence, linked rather than retold) — IQ Metrics
Corrections: spotted an error? Email corrections@iqmetrics.org and we will update this story and note the change here.
Related stories
All news →Do Brain Training Apps Work? One Trial Survived 20 Years
A follow-up published in February 2026 tracked 2,802 adults for two decades after five weeks of cognitive training. One of the three training arms showed a lower rate of dementia diagnosis. The other two showed nothing.
Four Languages, a Brain 13 Years Younger: Read the Caveats
Research presented at a European neuroscience meeting in July 2026 used magnetic brain recordings to estimate biological brain age, and found multilingual speakers scoring years below their calendar age. The gradient is striking. The design cannot show that languages caused it.
When Is the Human Brain Fully Developed? What 124,000 Brain Scans Show
No single age: grey matter volume peaks at 5.9 years, white matter at 28.7 and network wiring turns near 32. Two large studies, 123,984 scans and 4,216 people, and no study we read shows a switch at 25.
Read the research.
Then find your own number.
Our IIF-certified assessment reports your score with its scale, percentile and confidence range — and a breakdown of the cognitive domains behind it.
Start IQ Test →