Opinion Mar 14, 2025 · 11 min read

The Myth of "Brain Training" Games

Most commercial brain training apps don't transfer to real-world skills. Here's what the research actually says — and what genuinely works instead.

$2M
FTC fine on Lumosity
75
Neuroscientists against app claims
150min
Min exercise per week
6–12wk
To see real cognitive gains

The core problem: near-transfer vs. far-transfer

The billion-dollar brain training industry has been plagued by inflated claims for decades. Virtually every cognitive training program makes you better at that specific task. The scientific question is whether those gains transfer to other tasks or real-world performance.

Near-transfer

Improvements on tasks similar to the training task. Always observed. Getting faster at Lumosity's speed game = faster at that game. Not useful.

Far-transfer

Improvements that generalize to different cognitive domains or real-world tasks. Rarely observed from brain training apps. This is what you actually want.

FTC enforcement action

In 2016, the FTC fined Lumosity $2 million for deceptive advertising — specifically for claiming its games improved performance on tasks like school, work, and dementia prevention without adequate scientific evidence. A 2014 letter signed by 75 neuroscientists stated the claims "go beyond what the current science can support."

Cognitive interventions: evidence comparison

Far-transfer evidence strength for the most commonly recommended cognitive interventions, rated on a 0–10 scale based on the number and quality of RCTs.

Far-transfer evidence strength (0 = none, 10 = very strong)

Aerobic exercise 9.2 / 10
Sleep optimization 8.8 / 10
Learning a complex new skill 7.1 / 10
Action video games 5.9 / 10
Dual n-back training 3.8 / 10
Crossword puzzles 1.8 / 10
Commercial brain training apps 1.2 / 10

Ratings reflect consensus of independent meta-analyses, not any single study. Far-transfer = improvement in domains outside the training task itself.

What actually produces far-transfer gains

Aerobic exercise

Strong evidence

Of all cognitive interventions, physical exercise has the strongest evidence for broad cognitive benefits. It increases BDNF (brain-derived neurotrophic factor — critical for synaptic plasticity), increases hippocampal volume, and improves executive function, memory, and processing speed across multiple domains.

+1–2%
Hippocampal growth/yr
6–12wk
To see measurable gains
150min
Minimum per week
−35%
Cognitive decline risk

Learning a complex new skill

Good evidence

Learning to play an instrument, a second language, or a genuinely complex skill (programming, juggling, chess) produces structural brain changes and broad cognitive benefits. The key requirement: genuine novelty and complexity. The skill must challenge you continuously — not something you become comfortable with.

Requirements: the skill must be novel (brand new to you), complex (not mastered in days), and sustained (months to years). A 20-minute Duolingo streak does not qualify.

Action video games

Moderate evidence

Playing fast-paced action games (FPS, RTS) for 20–50 hours has been shown to improve visual attention, reaction time, and task-switching in multiple controlled trials. Crucially, the effect generalizes across diverse perceptual tasks better than most "brain training" programs — because the games require genuine real-time decision-making under pressure, not rote button-press practice.

Dual n-back training

Contested evidence

Dual n-back is the most scientifically studied working memory task. Early meta-analyses suggested fluid intelligence improvements. Later large-scale replications were more mixed. Current scientific consensus: it reliably improves dual n-back performance, moderately improves working memory, but fluid intelligence gains are uncertain and likely smaller than originally reported. Still better than most commercial apps.

What does not transfer (near-transfer only)

Training type What it improves Far-transfer Verdict
Lumosity / Elevate Scores on their own tests Minimal Not worth the subscription
Crossword puzzles Vocabulary, crossword skill Very weak Enjoyable but not training
Sudoku / puzzle games That specific puzzle type Very weak Not cognitive training
Jigsaw puzzles Visuospatial familiarity Weak Enjoyable, minor benefit
Passive learning (podcasts) Familiarity, not retention None Replace with active recall

How to build a real cognitive training program

Most people want a quick fix. The evidence points to boring, sustainable habits — not apps. Here is the protocol the research actually supports:

1

Foundation: aerobic exercise

150 minutes per week of moderate aerobic exercise (or 75 minutes vigorous). This is non-negotiable — no other intervention comes close for broad cognitive benefit. Running, cycling, swimming, rowing. Zone 2 heart rate or higher.

2

Sleep: 7–9 hours, consistently

Cognitive performance degrades faster than subjective fatigue. You will not notice how impaired you are on 5–6 hours per night. Measure it: test your reaction time and memory on different sleep durations and you'll see the data clearly.

3

Skill learning: pick one hard new thing

Choose one genuinely novel complex skill — a musical instrument, a language, a programming language, chess. Spend at least 30 minutes three times per week. Must be something you cannot yet do, and that challenges you continuously.

4

Tracking: use standardized tests

Use independent measurement tools (like Human Benchmark) — not performance on the training task itself. Test quarterly. Look for improvement in reaction time, memory span, and processing speed. If you're not improving, adjust inputs.

Track your cognitive baseline

Take all tests now to establish your baseline. Retest monthly to measure whether your program is producing real, measurable change.

See all 12 tests

Related articles