Brain-Training Games Make You Better at Brain-Training Games.
Mind · Evidence Review · 8 min read · What the largest review of the field found, the one trial that is a genuine exception, and what the $2 million fine was actually for
Based on the published research of Daniel J. Simons, PhD, Professor of Psychology, University of Illinois Urbana-Champaign, and six co-authors. Faculty profile · Google Scholar · Wikipedia · ResearchGate
The gains are real. The question is how far they travel, and the answer is: about as far as the task you practised.
The download usually follows a specific moment. A name that would not come. A word that went missing in front of eleven people. The second time in a week of standing in a doorway with no idea what the errand was.
So the question gets typed at some point after eleven at night, and it is a reasonable question. Do brain training games work, and is fifteen minutes a day going to hold the line?
This is a reading of published research. Blue Leaf Journal is not a clinical body; the findings below belong to the researchers named, and the plain-English rendering is ours.
The short answer: practising a cognitive task reliably improves performance on that task. The evidence that the improvement carries over to closely related tasks is weaker, and the evidence that it carries into everyday thinking is very thin. That is the conclusion of the largest review of the field, written by seven researchers who read the literature the industry itself cited.¹ One large clinical trial is a real and specific exception, and it is worth knowing exactly what it did and did not show.²
The short version, with numbers
- The 2016 review in Psychological Science in the Public Interest runs to 84 journal pages and was written by seven researchers across six institutions.¹
- Its finding, in the authors’ own words: brain-training tasks improve performance on the trained tasks themselves, with “less evidence” of improvement on closely related tasks and “very little evidence” of improvement on distantly related tasks or on everyday cognitive performance.¹
- The authors also record that many of the underlying studies “suffered from methodological problems and did not conform to best practices for research.”¹
- The exception: the ACTIVE trial, 2,832 adults averaging 74 years, ten sessions of 60 to 70 minutes. At ten years, 73.6 per cent of reasoning-trained participants were still above their own baseline against 61.7 per cent of controls, and 70.7 per cent of speed-trained against 48.8 per cent.²
- In the same trial, memory training produced no significant result at ten years.²
- In January 2016 the Federal Trade Commission announced that Lumos Labs, maker of Lumosity, would pay $2 million to settle charges of deceptive advertising for its brain-training programme.³
- What none of this measured: whether any of it changes the specific experience of losing a word mid-sentence at 52. That symptom was not the outcome in these studies.
In this article: Scope and method · Why transfer is the whole question · The evidence, study by study · The trial that is a real exception · How strong is this evidence · The one move · FAQ
Scope and method
This review covers commercial brain-training programmes and the laboratory cognitive-training studies they draw on, in adults, with an emphasis on whether trained gains generalise. It does not cover cognitive rehabilitation after stroke or brain injury, which is a separate literature with different aims. It does not cover exercise, sleep or hearing correction, each of which has its own evidence base for cognition and none of which is a game.
Sources were selected as follows: the field’s single most comprehensive published review, the largest randomised trial of cognitive training in older adults, and the regulatory record where advertising claims were formally tested. Where the evidence is contested, the disagreement is named rather than resolved.
Why does “does brain training work” depend on one word?
Because everything turns on transfer, which is the technical term for whether a gain travels.
Key terms
Near transfer (improvement on a task closely resembling the one practised, such as a different version of the same memory game).
Far transfer (improvement on a task that shares little surface resemblance with the training, such as reasoning after speed practice).
Everyday cognition (performance on the things a person actually does: following a conversation, managing medication, finding the word).
Active control group (a comparison group given a different but equally engaging activity, rather than nothing at all, so that expectation and attention are matched between groups).
The distinction matters because a headline that says training improved cognition is usually reporting near transfer, and the buyer is hoping for everyday cognition. Those are separated by two full steps of evidence, and the literature does not carry a reader across them.
The Simons review is unusually direct about this. Brain-training tasks improve performance on the trained tasks themselves. There is less evidence for closely related but non-identical tasks, and very little evidence that training improves performance on distantly related tasks or improves everyday cognitive performance.¹ The wording is careful, and the care is the point: the researchers are describing a gradient, not issuing a verdict.
If the reason for the download was a specific moment rather than a general worry, the moment is usually the more useful thing to look at. Start here.
The source for this piece
Daniel J. Simons, PhD
Professor in the Department of Psychology at the University of Illinois Urbana-Champaign. His research covers attention, perception, memory and research methodology, and he is a co-author of the selective-attention work widely known as the invisible gorilla experiment.
What we read: Simons and colleagues, “Do ‘Brain-Training’ Programs Work?”, Psychological Science in the Public Interest, 2016, volume 17, issue 3, pages 103 to 186.¹
Where to follow his work: Illinois faculty profile · curriculum vitae · Google Scholar · ResearchGate · Wikipedia · Psychology Today
The evidence, study by study
| Source | Size | Population | Design | Finding | Limitation |
|---|---|---|---|---|---|
| Simons et al., 2016¹ | 84-page review of the field’s literature | Adults, across the training literature | Systematic critical review | Gains on trained tasks; less on near tasks; very little on far tasks or everyday cognition | A review inherits the weaknesses of what it reviews; the authors state that many primary studies did not follow best practice |
| ACTIVE, Rebok et al., 2014² | 2,832 at enrolment; 44 per cent followed to 10 years | US adults, mean age 74, 76 per cent women | Randomised controlled trial, four arms, 10 sessions | Reasoning and speed training still above baseline at 10 years versus controls; self-reported daily function better in all trained arms | Memory arm not significant; daily-function outcome was self-reported; 56 per cent of the sample was lost by year 10 |
| FTC v. Lumos Labs, 2016³ | Regulatory action | Consumers of a commercial programme | Deceptive-advertising charge, settled | $2 million settlement over the programme’s advertising claims | A settlement is a legal outcome, not a scientific finding about whether training works |
The trial that is a real exception
ACTIVE, the Advanced Cognitive Training for Independent and Vital Elderly study, deserves to be described accurately rather than used as a slogan by either side.
It enrolled 2,832 volunteers across six US cities, randomised into three training arms and a control group. Training was ten sessions of 60 to 70 minutes over five to six weeks, with some participants randomly selected for later booster sessions. Effects were measured immediately and again at one, two, three, five and ten years.²
At ten years, 73.6 per cent of reasoning-trained participants were still performing reasoning tasks above their own pre-trial baseline, against 61.7 per cent of controls. In the speed arm the comparison was 70.7 per cent against 48.8 per cent. The memory arm produced no significant result. On self-reported ability to manage complex daily tasks, at least 60 per cent of participants in every trained arm reported less difficulty than the 49 per cent of untrained participants.²
“Showing that training gains are maintained for up to ten years is a remarkable result because it suggests that a fairly modest intervention in practicing mental skills can have long-term effects on older adults’ cognition and daily function beyond what we might reasonably expect,” said George Rebok, one of the study’s investigators, in the Johns Hopkins announcement.²
Three qualifications belong beside that. The participants averaged 74 at enrolment, which is two decades older than most women asking this question. Only 44 per cent of the original sample was still being followed at ten years. And the daily-function outcome was what participants said about themselves, not what anyone observed.²
What these studies did not measure
None of this literature measured the symptom that sends women to the app store. The outcomes are laboratory tasks and, in ACTIVE, a self-report questionnaire about daily activities. The experience of a word disappearing mid-sentence in a meeting, in a woman in her early fifties, is not an outcome in any of these studies, and the population in the largest trial is roughly twenty years older than she is.
The limitation the authors named themselves: the Simons review states that many of the studies it examined suffered from methodological problems and did not conform to best practices, and it recommends preregistration, adequate randomisation, sufficient sample sizes and correction for multiple comparisons.¹ A field that needs those recommendations is a field whose positive findings should be held loosely in both directions.
How strong is this evidence?
Moderate
Split by claim, which is the honest way to grade it. That practice improves performance on the practised task is strong: it replicates and nobody disputes it. That the improvement reaches everyday thinking is preliminary, resting on one large trial with a self-reported functional outcome and substantial attrition, against a wider literature the reviewers describe as methodologically weak.
The grade is not lowered because the products are commercial. It is lowered because the primary studies mostly lacked active control groups and preregistration, which are the two design features that separate a real effect from an expectation effect.¹
Who else has measured this
The 2016 review is not one psychologist’s opinion. It carries seven authors across six institutions, and the trial that complicates it carries eight investigators of its own.
Walter R. Boot, PhD and Neil Charness, PhD, both of Florida State University, are co-authors of the review, Charness through the university’s Institute for Successful Longevity. Their contribution to this literature is largely the case for active control groups, without which a training study cannot separate a genuine cognitive gain from the effect of expecting one.¹
the report, Association for Psychological Science · PubMed
Susan E. Gathercole, PhD, of the MRC Cognition and Brain Sciences Unit and the University of Cambridge, and David Z. Hambrick, PhD, of Michigan State University, are co-authors whose separate work on working memory and on the limits of practice runs directly into the transfer question.¹
free full text of the review · ResearchGate
Christopher F. Chabris, PhD, of Union College and Geisinger Health System, and Elizabeth A. L. Stine-Morrow, PhD, of the University of Illinois and the Beckman Institute, complete the authorship. Chabris is Simons’s long-standing collaborator on attention research.¹
Wikipedia · the review on PubMed
George W. Rebok, PhD, Professor in the Department of Mental Health and the Center on Aging and Health at the Johns Hopkins Bloomberg School of Public Health, is one of the ACTIVE investigators and the lead author of its ten-year report.²
Johns Hopkins faculty profile · Alzheimer’s Disease Research Center · Loop, Frontiers · LinkedIn
The Federal Trade Commission tested a commercial claim rather than a scientific one, and the record it produced is public. In January 2016 it announced that Lumos Labs would pay $2 million to settle deceptive-advertising charges over its brain-training programme.³
FTC press release · FTC, “Mind the gap”
The gain is real. It just stays where you put it.
What this actually changes
Not much about the app, and quite a lot about the worry underneath it.
A woman in her early fifties who cannot retrieve a familiar word is, on the evidence, unlikely to be looking at the problem ACTIVE was designed for. She is more often looking at something with a nearer explanation: interrupted sleep, the menopause transition, or the ordinary cost of holding nine open threads at once. Those have their own literatures, and unlike the training literature, they have measured the actual symptom. The imaging work on menopause and brain metabolism is one place that has been done properly, and the question of what counts as a normal midlife memory slip is another.
The reframe worth taking from all this is narrow and useful. The app is not a fraud, and it is also not a treatment. It is practice at a task, priced monthly, and it will make a person measurably better at that task. Whether that is worth the subscription is a spending question rather than a medical one, and it can be answered without any anxiety attached to it.
When this belongs with a doctor
Cognitive change is worth a clinical opinion rather than an app when: it is noticed by other people and not only by you; when it interferes with tasks you have done competently for years, such as paying bills or following a familiar route; when it has worsened noticeably over months rather than fluctuating; or when it arrives alongside changes in mood, personality, language or walking.
A sentence that gets it taken seriously in a ten-minute appointment: “This is a change from my own baseline, not a lifelong pattern, and it has been getting worse over about six months.” Ask specifically about thyroid function, B12, medication review and sleep, since all four are checkable and all four affect cognition.
The one move
Before you subscribe to anything, spend three minutes writing down what you actually want the fifteen minutes a day to fix.
Be specific. Not “my memory.” The thing itself: the word that goes in meetings, the name at the school gate, the reason for walking into the room. Then read your own sentence back and ask whether a laboratory task resembles it in any way.
If it does not, the money and the fifteen minutes are almost certainly better spent on the nearest measurable thing, which for most women in this position is sleep. That is not a consolation prize. It is where the evidence actually is.
Frequently asked questions
Do brain training games work?
They reliably improve performance on the trained task. The largest review of the field found less evidence for closely related tasks and very little evidence for distantly related tasks or everyday cognitive performance.¹ So the honest answer depends entirely on what you mean by work.
Does Lumosity work?
In January 2016 the Federal Trade Commission announced a $2 million settlement with Lumos Labs over deceptive advertising for its brain-training programme.³ That is a finding about advertising claims, not a scientific verdict on the software, and the general transfer evidence above applies to it as it does to any similar product.
Is there any cognitive training that has been shown to last?
Yes, with conditions. In the ACTIVE trial, reasoning and speed training still separated from controls at ten years, while memory training did not.² The participants averaged 74 years old, which is the detail most summaries leave out.
Why do the studies disagree with each other?
Largely because of design. The Simons review reports that many primary studies lacked features such as active control groups and preregistration, and it recommends both.¹ Without an active control, an improvement can come from expecting to improve rather than from the training.
Should I be worried about forgetting words at fifty-two?
Usually not on its own, and there are nearer explanations worth ruling out first, including sleep, the menopause transition and simple cognitive load. Worth a clinical opinion if others notice it, if it interferes with long-familiar tasks, or if it has worsened steadily over months.
The one move
Write down, in one specific sentence, what you want the training to fix. Then ask whether a laboratory task resembles it. Three minutes, before any subscription.
If the honest answer is that the load is the problem rather than the memory, that is worth ten quiet minutes on one page. Start here.
Related reading: why the page gets read and not retained, what happens when the sentence disappears mid-meeting, and what seven sleep trackers got right and wrong.
On the researchers. Daniel Simons is an experimental psychologist and methodologist, not a clinician; his review evaluates the design and findings of the training literature and makes no claim about any individual person’s cognition. George Rebok is a gerontologist and one of eight ACTIVE investigators. The Federal Trade Commission is a regulator, and its settlement concerns advertising rather than science.
Sourcing, disclosure and disclaimers
This is not medical advice. Blue Leaf Journal publishes general information for a general readership. Nothing here is a diagnosis, a treatment recommendation, or a substitute for care from a clinician who knows your history.
We are not clinicians. Blue Leaf Journal is an independent publication and its bylines carry no medical or academic credential. We read published research and translate it. The findings belong to the researchers and institutions named and linked above; the plain-English rendering is ours, and so is any error in it.
No affiliation and no endorsement. Blue Leaf Journal is not affiliated with Daniel J. Simons, Walter R. Boot, Neil Charness, Susan E. Gathercole, Christopher F. Chabris, David Z. Hambrick, Elizabeth A. L. Stine-Morrow, George W. Rebok, the University of Illinois Urbana-Champaign, Florida State University, the University of Cambridge, the MRC Cognition and Brain Sciences Unit, Union College, Geisinger, Michigan State University, Johns Hopkins University, the Association for Psychological Science or the Federal Trade Commission. None of them has reviewed, approved or endorsed this article.
No commercial relationship. No product named here paid for or was paid for this coverage. There are no affiliate links and no sponsored placements in this article.
How this was checked. Every figure and quotation above was checked against the cited source on 10 September 2026, and each source is linked below. If you find something we have got wrong, write to miriamalderton@blueleafjournal.com and we will correct it and say that we did.
References
1. Simons, D. J., Boot, W. R., Charness, N., Gathercole, S. E., Chabris, C. F., Hambrick, D. Z., and Stine-Morrow, E. A. L. “Do ‘Brain-Training’ Programs Work?” Psychological Science in the Public Interest, 2016, 17(3):103–186. Association for Psychological Science summary · PubMed · free full text
2. Rebok, G. W., et al. “Ten-Year Effects of the Advanced Cognitive Training for Independent and Vital Elderly Cognitive Training Trial on Cognition and Everyday Functioning in Older Adults.” Journal of the American Geriatrics Society, January 2014. Wiley · PubMed · Johns Hopkins Bloomberg School announcement, 13 January 2014, source of the ten-year percentages and the Rebok quotation.
3. Federal Trade Commission, “Lumosity to Pay $2 Million to Settle FTC Deceptive Advertising Charges for Its ‘Brain Training’ Program”, January 2016. Also “Mind the gap: What Lumosity promised vs. what it could prove”.
Keep reading
Miriam Alderton is Research Editor at Blue Leaf Journal. She reads the methods section first and the abstract last, and every figure in this piece is linked to its source above. The limitation of this review is that it rests on three sources rather than a systematic search of its own, and it inherits whatever the largest of them got wrong.
Written by Miriam Alderton for Blue Leaf Journal. Updated: 10 September 2026.
