Do Brain Training Apps Actually Work? An Honest Look at the Evidence

Part of the guide: MLM Reviews: The Honest Index

You downloaded the app, played the little games on the train for a week, watched your “speed score” climb, and felt sharper. Then one honest question crept in at 11pm: do brain training apps work, or did I just get good at the app? I’ve asked that question with my own money and a stopwatch. After twenty-plus years in this industry building skills that actually pay, I’ve learned to separate the feeling of progress from the proof of it. So here’s the honest version, with the real studies, the $2 million fine nobody mentions, and the one distinction that decides whether any of this changes your life or just your high score.

Let me start where most people start, because I started there too.

A while back I ran a free trial of one of the big brain-training apps. Pretty colors, satisfying dings, a chart that ticked up every session. Within a week my “memory” score looked gifted and I felt like Bradley Cooper in Limitless. Then I sat down to actually learn something hard, a new piece of software for the business, and I was exactly as slow and forgetful as I’d been the week before. Nothing had moved. The only thing I’d genuinely improved at was the game itself.

That gap, between getting better at the app and getting better at your life, is the whole story. Researchers have a name for it, and once you understand it you can never be fooled by a brain-training ad again.

In this guide

1. The two words that decide everything: near transfer and far transfer
2. The $2 million fact the ads will never show you
3. What the scientists themselves actually said
4. So is the whole category junk? No. Here’s the line that matters.
5. How to tell a real protocol from an expensive toy
6. The actionable playbook: how to actually train your brain in 2026
7. The verdict: do brain training apps work?
8. The thing that actually decides whether your mind gets sharper

The two words that decide everything: near transfer and far transfer

Cognitive scientists split the question “does it work” into two very different claims.

Near transfer is real and boring: practice a working-memory game and you get better at working-memory games. Drill a speed task and you get faster at that task. This happens reliably, in every age group, with almost every app. It is also close to worthless, because nobody’s life is bottlenecked by their ability to play a specific phone game.

Far transfer is the claim that makes the money. It says training a game improves something general: your reasoning, your fluid intelligence, your focus at work, your protection against age-related decline. That is the claim on every box. And that is the claim that, for the typical brain-training game, the evidence does not support.

I want to be precise here, because precision is the only thing that makes a review like this worth your time. The 2016 meta-analytic review by Melby-Lervåg and colleagues looked across the working-memory training literature and found no convincing evidence that the gains transfer to general intelligence, verbal ability, or reading. People got better at the trained task and stopped there. The improvement hit a wall at the edge of the game.

The $2 million fact the ads will never show you

Here’s the part that should make you sit up, because it’s not opinion, it’s a government action with a dollar figure attached.

In January 2016 the Federal Trade Commission fined Lumosity, the most famous brain-training brand on earth, $2 million for deceptive advertising. The FTC said Lumos Labs sold consumers on claims it could not back up: that 10 to 15 minutes of games a few times a week would improve performance at work and school, delay age-related decline, and protect against conditions including dementia, stroke, and PTSD. The regulator’s own summary was blunt: the company “preyed on consumers’ fears” about decline and could not prove the science it implied.

There’s a smaller, dirtier detail in that case worth holding onto. The FTC also found Lumosity had solicited some of its glowing testimonials through contests, dangling a free iPad and a lifetime subscription for the right kind of story. So the next time a brain app shows you a tearful review about a transformed memory, remember that someone, somewhere, may have been playing for an iPad.

This is the pain underneath the whole category. You’re a capable, analytical person, and you can feel yourself moving a half-step slower than you did at thirty, blanking on a name mid-sentence, missing a connection a younger colleague catches first. If that mental fog has become your daily normal, that fear is real and it deserves a real answer. What most of the industry sold you instead was a slot machine that pays out in confidence and nothing else.

What the scientists themselves actually said

This isn’t one cranky regulator: in 2014 the Stanford Center on Longevity and the Berlin Max Planck Institute published a consensus statement signed by about 70 of the world’s leading psychologists and neuroscientists. Their conclusion, in plain language: brain games can reliably make you better at the specific task you practice, but there is no compelling evidence that the improvement extends to broad abilities like reasoning and problem solving, and no convincing proof that these games reduce or reverse cognitive decline.

Now, in fairness, because a fair review names the other side, a group of 127 scientists signed a rebuttal pointing to dozens of randomized controlled trials. Some of them had ties to the industry, which you should weigh, but the disagreement is genuine. The science here is contested, not closed. And the 2016 review by Simons and colleagues, which read the entire literature carefully, found something telling: when a company brags about “100 studies,” a close look often reveals a small handful of datasets, recycled across many papers, frequently without active control groups. The volume is an illusion: the number of clean, independent studies is much smaller than the marketing implies.

So if you’ve ever quietly wondered whether you were being naive for buying a brain app, you weren’t. The skepticism in your gut was tracking the actual evidence. Trust that instinct.

So is the whole category junk? No. Here’s the line that matters.

This is where most “brain training is a scam” articles stop, and they’re wrong to. The honest finding is more interesting than the cynical one.

The reason typical brain games fail at far transfer is structural. A whack-a-mole speed game trains one narrow input-output loop. There’s no reason on earth that loop would generalize to a tax problem or a sales conversation, so it doesn’t. The mistake was never “training is impossible.” The mistake was training the wrong thing.

The interesting question, the one serious researchers kept chasing after the consensus statement, is this: is there any kind of training built on the actual underlying machinery of reasoning, rather than on a surface game? Because if you trained the thing intelligence is made of, transfer wouldn’t be a miracle; it would be the expected result.

That’s where Relational Frame Theory comes in, and it’s the most credible thread in this entire field.

Relational training, and the study with the eye-popping number

Relational Frame Theory is a behavioral account of how humans derive relations between things: same as, opposite of, more than, less than, before, after, part of. That sounds abstract until you realize it is literally what an IQ test measures. Analogies, matrices, sequences, “A is to B as C is to what.” Fluid reasoning is, at bottom, relational reasoning, which is exactly what a practical guide to training your brain with RFT sets out to drill.

So a group of researchers built a protocol called SMART, Strengthening Mental Abilities with Relational Training, that drills those relations to fluency, getting steadily harder. And in a 2016 study by Cassidy and colleagues, a sample of children who completed the training showed an average rise of about 23 points on the WISC-IV, moving from a mean IQ near 97 to near 120. Twenty-three points is not a rounding error. It’s the difference between average and gifted.

Now here’s where I refuse to do what Lumosity did. I’m going to tell you the weaknesses of that study, because you deserve the same honesty I’d want.

The 2016 sample was small, around fifteen children, and it did not use a control group. No control group means you can’t fully rule out practice effects or maturation. Follow-up replication work that added control conditions has generally found gains in favor of relational training, but more modest and more mixed than that headline 23. So the honest summary is this: relational training is the most theoretically sound and evidence-backed approach in the brain-training world, the one mechanism with a real reason to transfer, and the precise size of the effect is still being nailed down. That is a far stronger position than any speed game can claim. It is not a guarantee, and anyone who tells you it is should lose your trust the same way Lumosity did.

Notice the internal tension you might be feeling right now. Part of you has always half-believed intelligence is fixed, the number you were born with, and trying to raise it feels a little naive. Another part can feel yourself getting duller and refuses to accept that as the end of the story. Both feelings are honest, and the evidence says the fixed-IQ belief is too pessimistic and the brain-game ads are too optimistic. The truth lives in the narrow, unglamorous middle: train the actual substrate of reasoning, with a structured protocol, consistently, and the right kind of gain is plausible.

How to tell a real protocol from an expensive toy

Whether you ever spend a dollar with me or anyone else, you should be able to walk into this category and judge any app in five minutes, and here’s the checklist I use.

1. Does it train a mechanism or a surface skill? Ask what cognitive process the app claims to build and why that process would generalize. “Faster reaction time” generalizes to nothing. “Fluency in relational reasoning” is the substrate of fluid intelligence. If the answer is just “you’ll get better at our games,” that’s near transfer dressed up.
2. Is the difficulty genuinely adaptive? Real training keeps you at the ragged edge of your ability, somewhere around 80 to 85 percent accuracy, where learning actually happens. A game that lets you feel like a genius every session is selling comfort, not growth.
3. Does it measure you with something other than its own score? A vanity “speed index” that only exists inside the app is a closed loop. You want a baseline assessment and a measure of progress that means something off the screen.
4. Are the claims careful or are they magic? “Scientifically shown to improve reasoning” is a careful claim. “Cure brain fog, reverse aging, unlock your full potential” is the exact language that cost Lumosity two million dollars. Careful wording is a sign the people building it respect the evidence.
5. Is there a dose, like a real protocol has? Medicine has a dose, and so does training. Look for an honest statement of how many sessions per week and over how many weeks before you’d expect to feel anything. Vague “use anytime” framing usually means they have no expected outcome to defend.

The actionable playbook: how to actually train your brain in 2026

Say you’ve decided you want a sharper mind and you want to do it on evidence instead of hope. Here’s the plan I’d hand a friend: six steps, in order, so save this part.

1. Get a real baseline first: before any training, take an objective cognitive assessment and write the number down with the date. You cannot tell if something worked if you never measured where you started. Most people skip this and then argue with a feeling.

2. Pick the mechanism, not the dopamine: choose the program that trains the underlying process, relational reasoning, working memory under real load, not the one with the prettiest whack-a-mole. Re-run the five-point checklist above and let it decide for you.

3. Commit to a dose, not a vibe. Four to five sessions a week, real focus, for at least eight to twelve weeks. The research that shows anything shows it at a real dose. Fifteen minutes whenever you remember will produce a high score and nothing else.

4. Train at the edge: if you’re getting most answers right easily, the difficulty is too low and you’re wasting the session. Push until you’re missing a meaningful share. Discomfort is the receipt for learning.

5. Stack it on sleep, movement, and one hard real-world skill. No app outperforms a full night’s sleep and a brisk walk for raw cognitive function, which is why the boring fundamentals of a daily routine for mental performance beat any single tool. Nothing trains your brain like learning a genuinely difficult real thing, a language, an instrument, a new domain of your business, and if you’re weighing pills against practice, I broke down nootropics versus brain training separately. Use the app as a sharpening stone, not the whole knife.

6. Re-measure at week twelve, honestly: same assessment, new date. If the objective number moved and you feel it in real work, keep going. If it didn’t, you’ve lost twelve weeks and a few dollars instead of a year and your faith. Either way you learned the truth, which is the entire point.

Now read that playbook back and count how much of it any app actually does for you. It picks the exercises, and that’s it. The baseline, the discipline, the dose, the edge, the sleep, the hard real skill, the honest re-measure, all of that is you. The app is one tool inside a system. The system is what produces the result.

The verdict: do brain training apps work?

So, do brain training apps work? The honest answer has two halves: the typical game-style app does not deliver what it advertises. It produces near transfer, reliably makes you better at itself, and there’s strong evidence, a $2 million FTC fine, and 70 scientists on record saying the broad real-world benefits don’t materialize. If you bought one expecting to get smarter at life, your skepticism was correct.

The other half: a structured protocol that trains the actual substrate of reasoning, the relational-frame approach, is a genuinely different animal with real published support and a real mechanism for why it should transfer, even though the exact effect size is still being pinned down. That’s not hype; it’s the most defensible position available in this field, stated with its limits intact.

But here’s the part that matters more than any verdict on any app.

The thing that actually decides whether your mind gets sharper

Whether brain training works for you has almost nothing to do with which app you pick. I’ve watched two people buy the same program in the same week. One treats it like a slot machine, plays for the dings, quits in nine days, and concludes “brain training is a scam.” The other treats it like a protocol, baselines, doses, trains at the edge, re-measures, and twelve weeks later is genuinely faster on hard problems. Same app, opposite outcomes. The variable was never the software; it was whether they ran a system or chased a feeling.

That’s the quiet thing the capable, frustrated reader already suspects. You’re not getting duller because your brain is finished. You’re getting duller because nobody ever handed you a system for training it, only games that pay out in confidence and toys that pay out in nothing. Leaders use systems. Everyone else collects another app and wonders why the high score never shows up in real life.

Feel the difference for a second, because the feeling is the point. The person who runs a real protocol walks into the hard meeting knowing their reasoning is something they forge, not something they were issued at birth and slowly lose. That calm isn’t a personality trait; it’s downstream of a system, a measured baseline, a real dose, an honest re-measure, the same way physical strength is downstream of actually lifting and not of owning a gym membership.

So before you download one more game that only trains itself, do the boring, powerful thing the whole industry hopes you’ll skip. Measure where you actually stand, then train the mechanism, on a dose, at the edge, and check the number in twelve weeks. That’s the difference between a toy and a protocol. It’s the whole reason one changes how you think and the other just changes your screen.

---

Independent analysis: CortexSurge is my own cognitive-training product, so I have a financial interest if you upgrade to a paid plan, and I’ve told you the evidence with its limits intact precisely because you can check every source yourself. This is analysis and opinion, not medical advice, and nothing here is a medical claim about treating or preventing any condition. Cognitive training results vary by person and effort, and no specific outcome is guaranteed. Studies cited are accurate to the best of my research as of June 2026 and the science in this field is actively contested, so read the sources and judge for yourself. Sources: FTC on the Lumosity settlement, Stanford Center on Longevity consensus statement, Melby-Lervåg et al. meta-analysis on working-memory training, and Cassidy et al. (2016) on relational-frame training and IQ.

See also: the tools and systems I actually recommend, and start here if you’re new.