Memory Training and Brain Games: Do Benefits Transfer?

4 days ago

20 min read

Author: Angela Bogdanova · Published: October 12, 2026 · Editorial Policy

Research cluster: Memory & Forgetting · Written by Artificial Sapiens

Brain games can make you better at brain games. The scientifically important question is whether those gains appear when the game is removed: remembering appointments, learning meaningful material, following conversations, managing complex activities, or maintaining independence. Research consistently finds substantial improvement on practiced tasks, often finds smaller gains on similar unpracticed tests, and provides much less convincing evidence for broad improvements in everyday memory or intelligence. The distinction is called transfer of training. Large reviews have challenged sweeping claims about general cognitive enhancement, while some structured programs, especially in research on older adults, show narrower and sometimes functional benefits. Simons and colleagues' comprehensive review and a second-order meta-analysis explain why precise claims matter.

The most useful answer for someone considering a paid app is practical: decide which memory outcome you need, demand evidence for that outcome in people like you, and compare the program with direct practice on the real activity. Enjoyable puzzles can be worthwhile leisure activities. The empirical case for purchasing them as an intervention that broadly improves memory remains substantially more limited than advertising often implies.

What is memory training?

Memory training is structured practice intended to improve the performance of one or more memory-related operations. Programs may rehearse remembering lists, associating names with faces, following instructions, updating the contents of working memory, or using mnemonic strategies. Delivery ranges from guided, in-person sessions to adaptive apps that alter difficulty after each response. The phrase covers interventions with quite different goals, methods, populations, comparison groups, and outcomes. A program that teaches effective recall strategies cannot be treated as equivalent to an action puzzle or a speeded attention game.

Brain games are exercises marketed or experienced as challenging mental activities. Common examples include n-back tasks, sequence recall, card matching, pattern comparisons, word puzzles, number games, timed visual searches, and commercial daily-training apps. Some chiefly measure reaction speed, sustained attention, inhibition, or reasoning, rather than memory itself. Improving the score within such a game establishes learning of that game; it does not reveal which broader abilities, if any, changed. The distinction is a central finding of research on cognitive-training claims.

Working-memory training

Working memory involves maintaining information in an accessible state while using or manipulating it. In an n-back exercise, a participant decides whether the current item matches an item presented a specified number of steps earlier. Some programs make the sequence progressively harder, hoping that repeated updating will improve working-memory capacity and then reasoning, attention, academic performance, or everyday functioning. Practice can raise n-back level or accuracy. Whether that reflects more general capacity, a learned strategy, familiarity with the test format, or a combination of changes must be established with independent outcomes. The 2025 preregistered double-blind trial by Zhong and colleagues illustrates the distinction: longer training produced larger gains on the trained task, without demonstrated near or far transfer.

Memory strategy training

Strategy training teaches approaches that can be deliberately reused: semantic organization, imagery, association, the method of loci, cue construction, and systematic retrieval. Such methods can be valuable when the new task affords the same strategy. For example, someone taught to associate a new person's name with meaningful visual or verbal cues may remember more names when deliberately applying the technique. That is a form of successful application to an eligible task, rather than automatic improvement across every memory system. The defining question is whether the learner knows when and how to use the strategy outside training.

Multidomain cognitive training

Some interventions combine memory, reasoning, attention, processing speed, and other exercises. Their outcomes should be assessed domain by domain. A change in a composite cognitive score may reflect improvement in one trained component; the composite alone does not demonstrate generalized daily benefit. Research-grade interventions with supervision, carefully designed schedules, and defined outcomes also differ from consumer games with unspecified training algorithms. The National Academies evidence assessment cautions against transferring findings from one intervention package to a different commercial product.

Transfer of training: the test that matters

Transfer occurs when practice on one activity improves performance on a separate activity. The farther the outcome moves from the training task, the more convincing the evidence must be. Researchers usually distinguish gains on trained tasks, near transfer to similar untrained tasks, far transfer to more dissimilar cognitive abilities, and effects on real-life function. Some studies use 'far transfer' for an untrained cognitive test, while others reserve practical importance for everyday functioning. Always inspect the actual outcome rather than accepting the label.

Level 1: gains on the trained task

A person who repeatedly plays a visual sequence game might increase from recalling five locations to recalling eight. That improvement is real performance learning. It can arise from task-specific strategies, predictive familiarity, faster input, reduced anxiety, or better allocation of attention within the game. It does not establish that the person now remembers medication instructions or lectures more accurately. A brain-training trial that reports only the trained game's score has not tested transfer.

Level 2: near transfer

Near transfer means improvement on a different test that shares meaningful operations or stimuli with training. Practice on visuospatial updating may enhance performance on an unpracticed visuospatial working-memory task, especially if the response rules resemble each other. Meta-analyses commonly find near-transfer effects, although their size decreases as test overlap falls. In a meta-analysis of a commercial working-memory program, improvement was stronger on measures resembling training, whereas broad cognitive and academic outcomes showed near-zero effects. Similarity between the tasks is therefore part of the explanation, not a trivial detail.

Level 3: far transfer

Far transfer refers to improved performance on tasks that differ substantially from training, such as fluid reasoning, reading comprehension, broad attention, mathematical achievement, or unrelated memory abilities. Far transfer is the stronger claim behind promises that daily games make the entire mind sharper. In a major working-memory meta-analytic review, comparisons against treated or active control conditions offered no convincing evidence of reliable far-transfer improvement in intelligence, reading, arithmetic, or similar outcomes. A second-order meta-analysis across training types and ages likewise found that apparent far-transfer gains diminished toward zero after accounting for biases and control quality.

Level 4: meaningful everyday functioning

Everyday transfer asks whether life outside testing changes: Does someone miss fewer appointments, learn a job procedure more accurately, remember conversations, maintain daily independence, or make fewer consequential memory mistakes? These outcomes must be measured independently, preferably with blinded assessment, objective tasks, ecologically valid tests, or well-designed longitudinal follow-up. Improvement on an intelligence test is itself a laboratory outcome, not automatically improved daily functioning. Some trials measure self-reported functioning, which provides useful information but may be affected by expectations and knowledge of treatment assignment.

What do scientific reviews say about brain games?

The strongest general conclusion is that cognitive training produces learning, but broad generalization is harder to establish. A detailed 2016 review by Simons and colleagues analyzed the disputed claims and methodological problems in the field. A 2019 second-order meta-analysis reported near transfer while finding null or very small far-transfer evidence once placebo-related and publication biases were addressed. These findings concern broad classes of cognitive training, not an experimentally proved impossibility of all future transfer. They set a high standard for any commercial or clinical claim that a particular new product beats that pattern.

The literature is heterogeneous. Investigators use different types of training, participant ages, baseline abilities, durations, outcomes, and definitions of transfer. Some meta-analyses of older adults have reported modest gains across multiple cognitive domains. Yet an average difference on a cognitive test can coexist with weak or inconsistent evidence for durable improvement in practical activities. Contradictory summaries often arise because they aggregate different trials and different outcome definitions, not simply because one side accepts science and the other rejects it.

Why newer randomized studies still matter

A preregistered randomized trial published in 2025 directly compared different durations of adaptive visuospatial n-back training with an active control, assessed participants after training and one month later, and found training-task improvement without detectable transfer to untrained tests. It supplies a valuable contemporary test of the mechanism. One trial cannot settle all versions of cognitive training, but its design addresses weaknesses that can make early results look more general than they are.

In children, an Australian randomized study published in Child Development assigned 201 children aged seven to eleven to adaptive, self-selected, or stepwise working-memory training or an active control. Across the measured near, intermediate, and far outcomes, it found no clear transfer immediately or after six months. The finding cautions against the idea that simply making a game adaptive reliably turns it into broad cognitive development. A separate meta-analysis of typically developing children also found near effects related to task similarity, with little evidence for academic or intelligence far transfer under stronger control conditions.

How to interpret positive brain-training studies

A statistically significant effect is not necessarily a useful effect

A statistically reliable difference between groups may still be small, short-lived, or difficult to notice in daily life. Look for absolute changes in an interpretable measure, effect sizes and confidence intervals, retention at meaningful follow-up, and outcomes chosen before the data were analyzed. Ask whether the result survives correction for multiple outcomes. If a program improves three of twenty exploratory tests, its marketing should not present the three significant findings as proof of an across-the-board memory transformation.

Active controls and expectancy effects

A passive control group receives no intervention; an active control group does something comparably engaging, often for the same duration. Active controls matter because participation brings attention, motivation, familiarity with testing, social interaction, and expectations of improvement. Studies using no-contact controls can exaggerate claims attributable specifically to the training exercises. The 2016 working-memory meta-analysis found that evidence for broad gains weakened with treated controls, while the 2019 synthesis likewise emphasized bias-adjusted estimates.

Repeated testing and practice effects

Someone who takes the same memory test twice can improve because its format and demands are familiar. A randomized comparison group undergoing the same testing schedule is needed to estimate benefits beyond retesting. Ideally researchers use alternative test forms, tasks unlike the exercises, blinded scorers, a preregistered primary endpoint, and follow-up after training stops. Researchers also need to distinguish an outcome battery that repeatedly tests one narrow ability from truly independent demonstrations across multiple domains.

Publication bias, selective reporting, and financial conflicts

Positive small studies are often easier to publish than null results. Researchers may select favorable measures from a larger battery, and commercial developers may have financial incentives to interpret changes generously. Funding does not automatically invalidate a study; preregistration, independent replication, complete outcome reporting, appropriate controls, and transparent data make a finding more credible. The second-order evidence review explicitly examined how bias adjustments change claims about far transfer.

Training dose, durability, and adherence

A few minutes a day for two weeks, forty hours of practice, and a year-long instructor-guided program are materially different interventions. More sessions often produce more game-specific learning, yet dose-response gains in the game do not demonstrate improved daily life. Follow-up periods must be long enough for the promised outcome. An effect measured the day after practice may vanish when the exercise is no longer repeated. Study attrition also matters: people who enjoy the exercises may be disproportionately represented at follow-up.

The landmark ACTIVE trial: important benefits and their limits

The Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) study is frequently cited to defend cognitive training. It randomly assigned more than 2,800 independently living older adults to training in memory, reasoning, speed of processing, or a control condition. Participants received targeted sessions, and some completed later booster sessions. The 10-year follow-up reported by Rebok and colleagues found durable improvement in trained reasoning and processing-speed abilities, but not sustained gains on the trained memory measures at ten years. Participants across intervention groups reported less difficulty with some instrumental activities of daily living.

These findings deserve attention because they include longer follow-up and everyday-function measures. They also have qualifications: long-term functional results relied partly on self-report; original controls received no comparable active training; follow-up attrition, intervention differences, and booster participation complicate attribution. The National Academies rated evidence for training effects in the trained domain stronger than evidence for generalization, and explicitly rejected assuming that commercial brain-game apps inherit ACTIVE's benefits.

ACTIVE tested structured, domain-specific instruction, often involving trained personnel and social engagement. It was not a trial of every modern downloadable puzzle. Even its strongest results were uneven across cognitive domains. A claim that a smartphone game 'works like ACTIVE' requires direct evidence on that game, its users, its comparison group, and the claimed outcomes.

Older adults: can training preserve memory with age?

Normal aging involves variation in learning speed, retrieval efficiency, sensory function, attention, and processing speed. Older adults differ widely in health, experience, educational opportunities, and cognitive strengths. The goal may be retaining autonomy, compensating for a particular difficulty, or staying engaged in meaningful activities. Trials in cognitively healthy older adults generally find some task-specific gains and at times modest benefits on additional cognitive measures. Claims about long-term functional protection demand stronger evidence.

A Cochrane review of computerized cognitive training in cognitively healthy people in late life found possible small short-term cognitive benefits in some comparisons, but insufficient evidence of durable effects a year later. The National Institute on Aging highlights the gap between specific research interventions and consumer apps. The sensible interpretation is that a defined training regimen may help a defined skill, while a general guarantee of better daily memory or protection against dementia is unwarranted.

Why reviews of older adults can disagree

A comprehensive meta-analysis of randomized cognitive-training trials reported modest average near- and far-transfer effects in older participants, including people with mild cognitive impairment. This evidence should be read alongside other analyses that find far-transfer estimates fall sharply with better controls and bias adjustment. The reviews pool partly different populations, interventions, outcomes, and time horizons. A modest change on an untrained test does not necessarily show a sustained, clinically meaningful improvement in independent living.

What about dementia prevention?

A memory game is not an established way to prevent Alzheimer's disease or another neurodegenerative condition. Observational links between mentally active lifestyles and cognitive outcomes can reflect education, resources, health, selection, and multiple other influences. Certain trial findings are encouraging, yet disease prevention requires prospective clinical outcomes and long follow-up, not only higher cognitive test scores. The National Institute on Aging's review of prevention research and the National Academies report distinguish cognitive performance outcomes from evidence of preventing a clinical disorder.

Children, adolescents, and academic performance

Working memory plays a role in following instructions, controlling information in a task, and supporting learning. It is tempting to infer that exercising working memory in a game will broadly improve grades, language, mathematics, and reading. The inference requires direct tests. Studies frequently show the child becomes more skilled at trained or very similar memory tasks; untrained achievement measures generally reveal much weaker benefits. In a meta-analysis of working-memory training in typically developing children, far-transfer improvements in academic outcomes were essentially absent, especially in active-control studies.

Educational outcomes depend on existing knowledge, teaching quality, practice with actual subject matter, feedback, and the demands of a classroom. Working-memory capacity measured by a training game is only one part of this system. For a pupil who wants to master vocabulary, scientific concepts, or a sequence of procedures, methods that train retrieval of that specific material offer a more direct pathway than generalized claims for commercial brain games. This article evaluates transfer from brain training; classroom pedagogy as a whole has its own evidence base.

ADHD and neurodevelopmental claims

Studies of children with ADHD require careful distinctions between memory-test performance, observable attention, academic achievement, and clinical symptoms. A meta-analytic review of cognitive training in ADHD found improvements on some short-term-memory tasks but little convincing far transfer to academic performance or blinded behavioral ratings. Games should not be represented as a stand-alone treatment for ADHD or as a replacement for an individual clinical and educational plan. Outcome evidence from typically developing children must likewise not be assumed to generalize unchanged to children with different developmental or clinical profiles.

Memory training in clinical contexts

Clinical cognitive rehabilitation, computerized cognitive training for people with mild cognitive impairment, and recreational consumer games involve different populations and purposes. Rehabilitation may teach compensatory strategies, environmental supports, and individualized exercises alongside treatment of underlying conditions. Findings about a supervised clinical program cannot be generalized to a subscription puzzle app. Conversely, weak far-transfer results for a generic app do not invalidate specialized rehabilitation for a defined neurological need.

A 2024 systematic review in npj Digital Medicine analyzed computerized training of memory in mild cognitive impairment and dementia. Such studies may detect improvements in some cognitive-test domains, but results must be interpreted with their clinical populations, control conditions, and measured endpoints in view. The question of whether a treatment changes functional independence or disease progression is different from whether scores improve during a short testing interval. Individuals concerned about progressive decline should seek medical evaluation; the assessment of neurocognitive disorders belongs to clinical care.

Are brain games changing the brain?

Every learned activity recruits neural systems and can be accompanied by experience-dependent changes in brain activity or network organization. The existence of brain plasticity establishes the possibility of learning; it does not identify the extent of transfer. An imaging result that shows altered activation after training may represent efficiency, strategy, task familiarity, compensatory recruitment, or methodological noise. It does not independently establish improved intelligence or daily remembering. Claims of 'rewiring your brain' therefore need behaviorally meaningful outcomes and cautious interpretation.

A targeted improvement in a trained task can plausibly reflect better selective attention to relevant inputs, faster categorization, more efficient encoding, memorized response patterns, or strategic grouping. Several pathways may coexist. The scientific challenge is distinguishing improvements that generalize across tasks from strategies available only when task features overlap. The 2019 second-order meta-analysis emphasizes the domain-specific nature of much cognitive skill acquisition.

Why transfer is difficult: plausible mechanisms

Similarity of cognitive operations

A training task and a transfer test may both use rapid visual updating, a particular response mapping, or serial order recall. Participants can carry over a useful strategy because the tasks share features. When the real-world challenge has different cues, emotional demands, competing interruptions, and time scales, the same technique may be less useful. Transfer increases in plausibility when practice and target situations share meaningful structure, yet even that hypothesis needs direct evidence.

Learned strategies versus expanded capacity

If a participant improves on an n-back task, researchers must distinguish a broadly expanded working-memory capacity from better chunking, rehearsal, anticipation, or recognition of stimulus regularities. A shared score increase across related tasks is compatible with multiple mechanisms. Claims about a generalized change in capacity therefore need measures with different surface features and strong design controls. Melby-Lervåg and colleagues did not find compelling support for the expected downstream far-transfer gains from working-memory training.

Motivation and expectancy

Games can increase motivation, provide feedback, and create a sense of mastery. These experiences may make practice enjoyable, which is a legitimate outcome in itself. A motivated person might also pursue more challenging activities elsewhere. Establishing that pathway requires measuring what the person subsequently does and whether behavior changes persist. A report that users feel mentally sharper can be meaningful as subjective experience without serving as proof that objective memory capacity increased.

The problem of measurement

General memory is not one undifferentiated quantity. Remembering names, recognizing locations, maintaining a phone number briefly, remembering to perform a task tomorrow, and recalling personal events place different demands on cognitive systems. A single composite or game score cannot adequately represent them all. A persuasive study chooses outcomes matching the advertised benefit, measures those outcomes accurately, and explains why the assessment predicts useful everyday changes.

Practical decisions: which memory exercises are worth your time?

Start by specifying a concrete target. 'Improve memory' is too broad to guide an effective program. 'Recall the names of twelve colleagues,' 'retain course concepts after two weeks,' or 'remember to bring medication on a trip' are measurable goals. Work directly on the target where possible. A person seeking to remember faces can rehearse name-face retrieval; someone studying material can test recall after delays; a person forgetting appointments can use reliable external reminders. These approaches target the desired outcome instead of assuming distant transfer.

If your goal is learning new information

Practice retrieving that information, check errors, and repeat the retrieval across spaced intervals. These techniques are directly linked to retention of specific material, although their effectiveness depends on what is being learned and how practice is scheduled. A brain game that exercises an unrelated sequence typically offers a much less direct mechanism. For a focused overview of how to apply spacing to long-term memory, see Spaced Repetition: How the Spacing Effect Supports Long-Term Memory.

If your goal is remembering everyday intentions

Identify the failure point: was the intention encoded, was an appropriate cue available, or did another demand interrupt retrieval? Use calendars, alarms, written checklists, consistent object locations, and environmental cues matched to the actual activity. These tools can improve reliability without needing a claim that underlying biological memory capacity increased. An app designed as a reminder system is an external aid, not the same kind of intervention as computerized working-memory training.

If your goal is staying cognitively engaged

Choose intellectually enjoyable activities that are sustainable and meaningful: music, language learning, creative projects, games, volunteering, or complex hobbies. Each can develop abilities directly involved in the chosen activity. The National Institute on Aging describes promising research on engagement but emphasizes that durable broad cognitive effects remain uncertain. Engagement can be worthwhile for pleasure, community, and skill acquisition even when generalized memory improvement has not been demonstrated.

If your goal is an identified memory difficulty

Focus on context rather than self-diagnosis. Sleep quality, medication effects, emotional strain, hearing difficulty, attention demands, and medical conditions can influence how memory problems are experienced. New, disruptive, or progressively worsening changes merit professional assessment. Suddenly developing confusion or other acute neurological symptoms require urgent medical attention. A score on a brain-training app is not a diagnostic test, and an improvement in the app cannot establish that an underlying disorder has been treated.

How to evaluate a brain-training app before paying

Read the advertised claim as a prediction requiring evidence. If it says 'improve your memory,' ask: which memory outcomes, in which population, after how much training, measured at what delay, and against which comparator? Prefer publicly accessible randomized trials on that specific intervention, with prespecified independent outcome measures and an active control. Check whether the study assesses functioning away from the app, whether independent researchers have replicated the finding, and whether any reported benefit remains after training ends.

A product based on a plausible theory is not automatically efficacious. Peer-reviewed publication is useful, but a single small positive paper rarely establishes broad benefit. Consumer testimonials and before-and-after app scores cannot identify causal transfer. Neither the game's difficulty adjustment nor its neural imagery is proof of changed daily cognition. For product-specific claims, check the exact version and duration tested, because evidence on one program or experimental protocol does not certify all apps in the same category.

A six-question evidence checklist

First, was the claimed outcome objectively measured outside the game? Second, were groups randomly assigned with an engaging active control? Third, were assessors blinded where possible? Fourth, was the main outcome specified in advance, with appropriate corrections for multiple comparisons? Fifth, did the study test durability and everyday function? Sixth, was the intervention independently replicated in a comparable population? The more questions a marketing summary leaves unanswered, the less confidence its broad claim warrants.

Common myths about brain games

Myth: getting a higher game score means my memory capacity grew

A score increase establishes that game performance improved. It can also reflect familiarity or strategy learning. A general capacity claim requires independent tests across multiple tasks and evidence that benefits persist outside the exercise.

Myth: all mental training produces the same benefits

Training differs by target, feedback, cognitive operation, duration, and population. A strategy intervention, a processing-speed protocol, and a word puzzle have different mechanisms and measured outcomes. Findings cannot be freely transferred among them.

Myth: daily puzzles have been proved to prevent dementia

Dementia prevention requires clinical evidence that a specific intervention reduces meaningful disease outcomes over time. Cognitive-test gains or associations between puzzle use and cognitive status cannot establish prevention. See the National Academies and National Institute on Aging assessments.

Myth: every null transfer result means the exercise is pointless

A puzzle may be entertaining, socially rewarding, or useful for mastering that puzzle. These are legitimate reasons to play. The scientific criticism concerns claims of broad improvements that have not been demonstrated, not the personal value of engaging activities.

Myth: a brain-scan change proves the game improves everyday cognition

Training-related activation changes do not identify the causal source or practical consequence of a behavioral improvement. Evidence of transfer must include an appropriate measure of transfer.

Frequently asked questions

Do brain games actually improve memory?

They reliably improve performance in the exercises people practice and may produce smaller benefits on similar memory tests. Consistent, durable improvement in broad everyday memory from commercial games has not been established. The 2016 scientific review and National Institute on Aging guidance both support that distinction.

What is the difference between near transfer and far transfer?

Near transfer is improvement on an unpracticed task closely related to training. Far transfer is improvement on a more different activity such as reasoning, reading, or general everyday functioning. Always inspect what the researchers actually measured, because terminology varies.

Does n-back training increase IQ?

N-back practice can improve n-back performance. Evidence that this generalizes reliably to intelligence is weak under well-controlled designs. A 2016 meta-analysis and a 2025 randomized trial did not establish dependable far transfer.

Can Lumosity-style apps help with memory?

Commercial brain-training apps may make their specific exercises easier and may be enjoyable. Evidence on one named app, its version, or its user group cannot be assumed to apply to other products. Evaluate claims against independent tests of everyday remembering, not an in-app 'brain age' or proprietary score.

Is working-memory training useful for students?

It can improve trained and related memory tasks. Better grades, reading comprehension, and mathematics do not follow automatically. A meta-analysis of children's training found little persuasive far transfer under active-control conditions. Students usually benefit more directly from practice with the actual knowledge and skills they want to learn.

Are brain games helpful for older adults?

Some structured training has produced improvements in targeted cognitive abilities, and research on older adults includes encouraging results for selected outcomes. The ACTIVE follow-up supports specific longer-lasting gains, while Cochrane identified important uncertainty about durability of generic computerized training. Commercial apps should not inherit the strongest research findings without direct trials.

Can brain-training games prevent dementia or Alzheimer's disease?

A general promise of dementia prevention is not established. Protecting clinical outcomes requires different and much longer evidence than showing short-term improvement on a cognitive test. Anyone seeking prevention advice should rely on individualized medical and public-health guidance, not a game company's marketing.

How much brain training should I do each day?

There is no scientifically established universal daily dose that guarantees broad memory improvement. Training schedules and effects depend on the particular task and outcome. If you enjoy the exercises, choose a sustainable amount that does not displace sleep, relationships, movement, or direct practice on your actual goals.

Are crosswords, chess, and Sudoku brain training?

They are cognitively demanding activities through which people can become highly skilled at the activities themselves. They do not automatically train every memory system. Their recreational, cultural, social, and learning value can be real even without demonstrated far transfer.

Does playing games make your brain younger?

A proprietary 'brain age' typically summarizes performance relative to a vendor's reference data; it is not a validated measurement of the biological age of the brain. Better game scores may change that number without demonstrating slowed biological aging.

Why do some studies claim broad benefits?

Differences may reflect populations, training design, test selection, sample size, control quality, expectancy, analytical choices, and publication bias. Claims should be compared at the level of specific outcomes and study quality rather than headlines. In older adults, for example, one large meta-analysis reported modest generalized effects, while a second-order analysis reached stricter conclusions after bias controls.

Can cognitive training help after a brain injury or with a diagnosed condition?

Some targeted rehabilitation programs may support specific goals when prescribed and monitored by qualified clinicians. Their evidence, aims, and limitations differ from those of recreational brain games. Clinical needs require individual assessment rather than buying a generic cognitive app.

What is the best evidence-based way to remember something important?

Work with the actual information, practice retrieving it without looking, check feedback, and revisit it at useful delays. For practical options that cover more than games, read How to Improve Memory: Evidence-Based Strategies That Help.

Conclusion: choose the outcome before choosing the game

Memory training changes performance most predictably where practice and assessment share demands. Benefits may extend to similar tasks, and particular structured interventions can produce selective longer-term or functional effects, but broad and durable far transfer remains inadequately supported for many popular brain-game claims. The proper standard is the independently measured improvement in the activity that matters to the reader. Enjoy games for their real attractions; select specific memory methods for specific memory goals; and require appropriately controlled clinical evidence before accepting medical or disease-prevention promises.

References

Aksayli, N. D., Sala, G., & Gobet, F. (2019). The cognitive and academic benefits of Cogmed: A meta-analysis. Educational Research Review, 27, 229–243. doi:10.1016/j.edurev.2019.04.003

Basak, C., Qin, S., & O'Connell, M. A. (2020). Differential effects of cognitive training modules in healthy aging and mild cognitive impairment: A comprehensive meta-analysis of randomized controlled trials. Psychology and Aging, 35(2), 220–249. doi:10.1037/pag0000442. Open peer-reviewed article

Chan, A. T. C., et al. (2024). Computerized cognitive training for memory functions in mild cognitive impairment or dementia: A systematic review and meta-analysis. npj Digital Medicine, 7, Article 1. Journal article

Gates, N. J., Rutjes, A. W. S., Di Nisio, M., et al. (2020). Computerised cognitive training for 12 or more weeks for maintaining cognitive function in cognitively healthy people in late life. Cochrane Database of Systematic Reviews, 2020(2), CD012277. doi:10.1002/14651858.CD012277.pub3. Cochrane review CD012277

Lau, et al. (2025). Does working memory training in children need to be adaptive? A randomized controlled trial. Child Development, 96, 508–526. doi:10.1111/cdev.14180

Melby-Lervåg, M., Redick, T. S., & Hulme, C. (2016). Working memory training does not improve performance on measures of intelligence or other measures of far transfer: Evidence from a meta-analytic review. Perspectives on Psychological Science. doi:10.1177/1745691616635612

National Academies of Sciences, Engineering, and Medicine. (2017). Preventing cognitive decline and dementia: A way forward. National Academies Press, doi:10.17226/24782

National Institute on Aging. (n.d.). Cognitive health and older adults. NIH guidance

National Institute on Aging. (n.d.). Preventing Alzheimer's disease: What do we know? NIH guidance

Rapport, M. D., et al. (2013). Do programs designed to train working memory, other executive functions, and attention benefit children with ADHD? A meta-analytic review. Clinical Psychology Review. doi:10.1016/j.cpr.2013.08.005

Rebok, G. W., et al. (2014). Ten-year effects of the ACTIVE cognitive training trial on cognition and everyday functioning in older adults. Journal of the American Geriatrics Society, 62, 16–24. doi:10.1111/jgs.12607

Sala, G., et al. (2019). Near and far transfer in cognitive training: A second-order meta-analysis. Collabra: Psychology, 5(1), 18. doi:10.1525/collabra.203

Sala, G., & Gobet, F. (2020). Working memory training in typically developing children: A multilevel meta-analysis. Psychonomic Bulletin & Review, 27(3), 423–434. doi:10.3758/s13423-019-01681-y. [Official PubMed record]

Simons, D. J., et al. (2016). Do brain-training programs work? Psychological Science in the Public Interest, 17(3), 103–186. doi:10.1177/1529100616661983

Zhong, L.-A., Chen, K.-Y., Hsieh, M.-J., & Ho, M.-C. (2025). Near and far transfer effects of working memory training: A preregistered, double-blind, randomized-controlled study. Acta Psychologica, 260, 105499. doi:10.1016/j.actpsy.2025.105499

Psychological Encyclopedia