Smartigo › Notes

Brain training: does it actually work?

October 10, 2026 · 4 min read

Some statements about brain training show up so often in ads and guides that you start to treat them as settled. "Trains your intelligence." "Slows ageing." Let's look at what the studies actually say, and what they don't.

What's largely undisputed

If you practise a task, you get better at that task. That's true for mental arithmetic, for sudoku, for a memory game. Even the sceptics don't deny it. A large review by Simons and colleagues (2016) concluded that there is broad evidence for improvement on the trained task, less for closely related tasks, and little for distant everyday performance.

So the question isn't whether practice does something. It's where it takes effect.

What's contested: transfer

The dispute is about transfer. Does training a memory game also improve your memory at the supermarket? On one side are research groups around Giovanni Sala and Fernand Gobet, who in several meta-analyses conclude that far transfer is on average zero or very small once you control for study quality and placebo effects. On the other side are papers that do see effects in their data, and they argue that better studies tend to show larger effects. Who is right also depends on how you weigh study quality, control groups and publication bias.

As non-experts we can't settle it. But we can name the careful reading honestly: near transfer is plausible, far transfer is uncertain, and big promises aren't covered.

And "digital dementia"?

The term comes from a German book from 2012 and stuck because it feels so right. That a phone that does everything for us also does something to us is not an unreasonable suspicion, and we've written about it ourselves (Frictionless isn't harmless). But as a claim that screens cause a disease, the thesis is a lot sharper than the data I know of. Studies on screen time and cognitive performance show a mixed picture, and correlations aren't causes. It's conceivable that we lose something when we no longer remember, calculate or look things up ourselves. It's not established that it's as dramatic as the name sounds.

Uncertain doesn't mean unimportant. It means you can stay calm and still do something.

How to recognise a serious offer

• It doesn't promise more than it knows. "Raises your IQ by 10 points" is a warning, not a selling point. Same for anything with "prevents dementia".
• It practises something you really need. Change, prices, routes, deadlines, suspicious messages: that's close to your life, and that's where a gain is most likely.
• It gives you an appetite, not a guilty conscience. A streak that punishes you for missing a day is a business model, not training.
• It's short. If it needs an hour-long programme, it won't survive daily life.

A sober calculation

Suppose the most pessimistic reading is right and there's no far transfer. You will then have invested five minutes a day to become a bit surer with discounts, deadlines and remembering. So you've done something you need anyway, and it hasn't cost much. Suppose the optimistic reading is right. Then something is added. In both cases the question isn't "Is it worth it?" but "Do I enjoy it enough to actually do it?"

That's unspectacular, but it's honest. More on the idea behind it in Five minutes of effort.

A small test

Next time before you open a "brain training" app, ask yourself: what exactly is practised here, and where in my life do I need it? If the answer is "nowhere, but it looks like neuroscience", you've understood everything.

What movement has to do with it

Anyone writing honestly about brain training can't avoid an uncomfortable point: the measures with the best evidence for thinking aren't digital. Regular exercise, enough sleep and social contact come up again and again in reviews among the factors associated with better mental fitness. Caution is in order here too, since association isn't cause, and healthy people may simply move more. But as a starting point it's more honest than any game.

No rejection of thinking tasks follows from this. What follows is that a puzzle is no substitute for a walk, and the other way round. Both together are at least plausibly better than either alone. It's perhaps no coincidence that we built Smartigo as a hiking map: if you think for five minutes and then take a lap around the block, you've done quite a lot right that day.

So when someone asks for the "best" brain training app, the most honest answer is: it depends on whether it keeps you going without kidding you. And the second most honest answer: go outside today.

One more practical thought on dosage. The research on learning and practice favours practising often and briefly over rarely and long. Ten minutes on six days is better for almost any skill than one hour on Sunday. That's not special to thinking games; it holds for vocabulary just as for piano. Whoever takes this to heart doesn't need a programme with a weekly plan and progress bars. A fixed time and a task that is just hard enough will do.