Do brain-training apps work?

Commercial brain-training apps — Lumosity, Elevate, Peak, BrainHQ, CogniFit, and many others — typically promise better memory, sharper focus, or protection against age-related decline. The marketing is confident. The peer-reviewed record is more careful, and in one high-profile case the regulator intervened.

neurogym.club is an informational reference, not a product or clinic. This page summarizes what the literature and public records actually support (as of 2026-07-21). It is not medical advice.

Near transfer vs far transfer — the crux

Cognitive scientists distinguish two kinds of training effect:

• Near transfer: you get better at tasks that closely resemble what you practiced (same rules, same stimuli, same response format).
• Far transfer: improvement shows up on untrained, real-world abilities — school grades, job performance, fluid reasoning outside the app, everyday memory.

Near transfer is common and unsurprising. Practice a reaction-time game and reaction times improve. Far transfer is what most consumers believe they are buying. That is the claim that has been hardest to support at scale.

If a study only shows gains on the trained game or on nearly identical lab tasks, that is near transfer. It is not evidence that general cognition improved.

FTC v. Lumos Labs (Lumosity)

In January 2016, the U.S. Federal Trade Commission announced that Lumos Labs, marketer of Lumosity, would pay $2 million and accept restrictions on advertising claims. The FTC alleged that the company made unfounded claims that Lumosity games would improve performance at work and school, and reduce or delay cognitive impairment associated with age and other health conditions, without adequate scientific substantiation.

Primary source (confirmed 2026-07-21): FTC press release, January 5, 2016.

The settlement is not a scientific meta-analysis. It is a regulatory finding about advertising claims. It remains the standard public cautionary citation for consumer brain-training marketing in the United States.

The 2014 scientific consensus — and the counter-statement

In October 2014, the Stanford Center on Longevity and the Max Planck Institute for Human Development published a public statement signed by a large group of cognitive scientists and neuroscientists. Their core position: the scientific literature does not support claims that software “brain games” improve general cognitive performance in everyday life or prevent cognitive slowing and brain disease in the way marketing often implies.

Full statement (confirmed 2026-07-21): Stanford Center on Longevity / Max Planck consensus.

A second group of researchers published a counter-statement arguing that some cognitive training interventions show meaningful effects and that the first statement understated positive evidence. The existence of two opposing open letters is itself useful: the field was (and in important respects still is) contested. Presenting only one side would misrepresent the dispute.

What both sides largely agree on is narrower: practice improves practiced tasks. The fight is about how far, how durable, and how general those gains go.

Dual n-back: the working-memory celebrity

Dual n-back training was popularized after Jaeggi and colleagues (2008) reported that adaptive dual n-back practice improved measures related to fluid intelligence — a result that, if robust, would be a rare example of far transfer from a working-memory task.

Source: Jaeggi SM, Buschkuehl M, Jonides J, Perrig WJ. Improving fluid intelligence with training on working memory. Proceedings of the National Academy of Sciences. 2008. PubMed / PNAS record (confirmed 2026-07-21).

Replication attempts since then have been mixed. A meta-analysis by Soveri and colleagues (2017) found modest near-transfer effects and limited far-transfer — consistent with improvement on related memory tasks more than broad intelligence gains.

Source: Soveri A, Antfolk J, Karlsson L, Salo B, Laine M. Working memory training revisited: A multi-level meta-analysis of n-back training studies. Psychonomic Bulletin & Review. 2017 (and related n-back meta-analytic work from that program). See also open-access discussions citing this literature (confirmed 2026-07-21).

Fair summary for a non-specialist: dual n-back is an interesting research tool. Enjoying it is not harmful on current evidence. Treating it as a proven path to general intelligence is stronger than the replication record supports.

What is fair to say the apps do?

Honest claims that usually hold:

• They can make you better at the games (near transfer).
• Some programs show improvements on closely related lab tasks.
• Engagement, feedback, and a sense of progress can be real psychological benefits of practice — separate from far-transfer cognition claims.
• Aerobic exercise, resistance training, sleep, and treating hearing loss have stronger, more replicated cognitive evidence bases than commercial app portfolios as a class. See Exercise and cognition and Modifiable dementia risk factors.

Claims that usually do not hold up as advertised:

• Guaranteed improvement in school or work performance.
• Proven delay or prevention of dementia or Alzheimer’s disease via a consumer brain game.
• Far-transfer gains equivalent to what multi-domain lifestyle interventions show in population and trial literature.

Brand names above are examples of the commercial category, not a ranking. We do not endorse or sell any of them.

How to read a “brain training works” headline

1. What was the control group? Active controls (another engaging task) matter more than no-contact controls.
2. What outcomes moved? Trained tasks only, or untrained batteries and real-world endpoints?
3. How long did gains last? Immediate post-test gains can fade.
4. Who funded the study? Industry funding is not automatic invalidation, but it is a reason to look for independent replication.
5. Does the abstract match the methods? Marketing abstracts sometimes overstate transfer.

Industry response and residual uncertainty

After the 2014 open letters, companies and affiliated researchers continued to publish training trials — some with better active controls than early work, some still thin. A fair 2026 reading is not “all cognitive training is fake.” It is: task-specific improvement is easy to show; broad real-world enhancement remains the exception that needs exceptional evidence. Regulatory history (FTC 2016) and the scientific dispute (2014) both push consumers toward that standard, not toward blanket cynicism about every study.

If a brand cites one positive trial, ask whether a pre-registered multi-site replication exists, whether outcomes were far-transfer, and whether the abstract’s verbs match the confidence intervals.

Bottom line

Brain-training apps are entertainment and practice tools. They are not a substitute for the lifestyle and medical factors with the strongest cognitive evidence. For how the brain actually changes with novel learning — without a subscription — see Neuroplasticity explained. For the practices that replicate most consistently, start with Exercise and cognition and Sleep and memory.

Sources on this page were checked against the linked primary or institutional URLs as of 2026-07-21. This page is educational, not medical advice.