The "do brain training apps actually work?" question has a clean published answer once you separate the two things it asks. Near transfer — improvement on the trained task itself — is robust and uncontested. Far transfer — improvement on untrained tasks or in real-world cognitive performance — is the contested claim, and it's where the brain-training category got into trouble. This post walks through the primary sources, the regulatory record, and what changed in 2026.
The 2014 Stanford-organised consensus statement
The cleanest summary of the scientific position on commercial brain-training comes from a 2014 consensus statement signed by over seventy cognitive scientists, organised through Stanford's Center on Longevity and the Max Planck Institute for Human Development. The statement's core conclusion: "there is little evidence that playing brain games improves underlying broad cognitive abilities, or that it enables one to better navigate a complex realm of everyday life." The full statement is hosted at Stanford Longevity; Science magazine's coverage is the most-cited secondary source.
What the statement actually targeted: the marketing promise of broad transfer to real-world ability. The framing matters because the consensus did not say brain games were worthless. It said the marketing claims outran the evidence.
The 2016 Simons et al. comprehensive review
Two years later, Daniel Simons and colleagues published a comprehensive review in Psychological Science in the Public Interest titled "Do 'Brain-Training' Programs Work?" (PMC full text; DOI 10.1177/1529100616661983). The review synthesised the published evidence base for commercial brain-training products and reached three main conclusions:
- Practice on a specific cognitive game tends to produce improvement on that specific game (near transfer is robust).
- Evidence for far transfer — generalisation to untrained cognitive abilities or real-world performance — is weak for most commercial products.
- The strongest evidence for any commercial product is the speed-of-processing programme from the NIH-funded ACTIVE Trial (now part of BrainHQ).
The review is exhaustive — over 100 pages, hundreds of citations — and remains the standard reference for the state of the field. The 2026 picture has not meaningfully changed.
The FTC's 2016 enforcement action against Lumosity
In January 2016 the FTC announced a $2 million settlement with Lumos Labs, the maker of Lumosity, over "deceptive advertising" (FTC press release). The complaint identified specific marketing claims as problematic:
- That Lumosity would help users perform better at school and work
- That it would delay age-related cognitive decline
- That it would reduce the effects of Alzheimer's disease
- That it would reduce cognitive impairment from health conditions including ADHD and PTSD
The settlement required Lumosity to notify auto-renewed subscribers, offer cancellation paths, and pay the $2 million. The FTC's public statement noted that the brain-training industry had "preyed on consumers' fears about age-related cognitive decline."
What changed after 2016: brain-training marketing became more careful. The category did not collapse, but the broad-cognition-uplift pitch quietly retreated. By 2026, the major apps — Lumosity, Elevate, Peak — make narrower claims than they did in 2014, and the FTC has not pursued comparable action against the post-settlement category.
The NIH ACTIVE Trial — the strongest exception
The Advanced Cognitive Training for Independent and Vital Elderly (ACTIVE) Trial is the strongest evidence for any commercial brain-training product. Funded by the National Institute on Aging and the National Institute of Nursing Research, the trial enrolled 2,832 healthy older adults across six U.S. sites and randomised them to three training arms — memory, reasoning, and speed of processing — plus a no-contact control.
Ten-year follow-up data (Rebok et al., 2014, PMC4055506) reported the speed-of-processing group had significantly less difficulty with instrumental activities of daily living (preparing meals, managing finances, getting around) compared to controls. The speed-of-processing programme used in ACTIVE became the core of BrainHQ's commercial product, and ACTIVE remains the most-cited counter-example to the broad-skepticism position.
Two important caveats:
- The effect is specific to processing-speed training in older adults — it does not generalise to other brain-training programmes
- Subsequent re-analyses (Edwards et al., 2017 on dementia risk) have been contested, and the strength of the long-tail claims continues to be debated
For most users, the right read is: BrainHQ has more evidence behind its specific claims than other commercial brain-training products. It does not mean every brain-training app has equivalent backing.
What the deliberate-practice literature adds
Outside the commercial-brain-training conversation, the deliberate-practice tradition — Ericsson, Krampe & Tesch-Römer 1993, with the 2019 partial replication by MacNamara & Maitra in Royal Society Open Science — supports a narrower but more durable result: structured, deliberate, repeated practice tends to maintain and improve the specific skill being practised. Domain-specific. Diminishing returns.
This is the framework Senwitt sits inside, deliberately. The product's promise is "practise the skills, keep using the skills" — which is what near transfer actually supports — rather than "broad cognitive uplift," which is what far transfer would require and what the evidence has not produced.
So do they work?
The honest 2026 answer in three parts:
Yes — for the specific cognitive moves the app actually drills. If you practise the Stroop task daily for six weeks, you'll get better at the Stroop task. That's near transfer, and it's real.
Mostly no — for the broad cognitive-uplift marketing claims that defined the category in the 2010s. The Stanford consensus, the Simons review, and the FTC settlement all converge on the same point: those claims outran the evidence.
Yes, for one specific case — BrainHQ's speed-of-processing programme in older adults, where the ACTIVE Trial provides the strongest commercial-product evidence in the category. Even here the effect is narrow and contested.
For 2026 users picking an app, the rational frame is: pick a product whose claims fit the evidence. Senwitt's scope-of-evidence page lays out exactly what we do and don't claim. The brain exercise vs brain training research page draws the category-level distinction that the post-2016 honest products are built around.
What changed in 2026
Three notable shifts in 2026 specifically:
The AI era reframed the question. The cognitive-offloading conversation around AI tools — drafting, summarising, calculating, reasoning all moving outside the user's head — gave the broader category a new framing that doesn't depend on the contested brain-training claims. Daily practice as a counterweight to AI offloading is a narrower, more defensible pitch.
The MIT Media Lab cognitive-debt preprint (Kosmyna et al., 2025, arXiv 2506.08872) gave the AI-offloading argument an EEG-level empirical hook, with the Stanković 2026 critique flagging methodology concerns. We cite both together for intellectual honesty.
The BCG/HBR 2026 AI brain-fry coverage put the cognitive-fatigue side of AI use into the mainstream conversation, broadening the audience for daily-practice tools beyond the cognitive-aging niche the brain-training category used to occupy.
