Do puzzles train your brain? What the science actually says
Near transfer, far transfer, cognitive reserve and what the population studies really show. The honest version, without the brain-training marketing.
There are two versions of this story and both are bad. One says puzzles «prevent Alzheimer's» and «make your brain ten years younger». The other says they are useless entertainment and that the whole brain-training sector is smoke.
The reality is in between and it is more interesting than either. Let us take it in parts, and carefully: this is a field where far more has been sold than has been shown.
What is well established: you improve at what you train
If you do sudokus every day, within a few weeks you will be clearly better at sudokus. That is not trivial and it is not a trick: it is real, measurable learning, and it holds up reasonably over time.
The same goes for word searches (visual search), anagrams (manipulating letters in working memory) or mental arithmetic. You train a specific skill and that skill improves. This is the part nobody disputes.
It is called near transfer: the improvement carries over to tasks very similar to the trained one. Practising sudokus will make you somewhat better at other number puzzles with similar logic.
What is disputed: whether it helps in daily life
Here the problem starts. The question that really matters is whether training with puzzles improves memory, attention or reasoning in general, in a way you notice when remembering where you left your keys or following a difficult conversation.
That is called far transfer, and the evidence is considerably weaker than brain-training advertising has suggested for years. Systematic reviews of the literature tend to find that improvements stay close to the trained task, and that effects on general capacities, when they appear, are small and do not always replicate.
At one point this went far enough to end up in court: in the United States, the Federal Trade Commission penalised brain-training companies for deceptive advertising, precisely for claiming benefits to everyday performance and cognitive decline risk that their own studies did not support.
The honest conclusion: doing sudokus probably will not make you smarter in general. It will make you better at sudokus, and that is already something.
What population studies suggest (with important caveats)
On the other hand, there is a consistent observation: people who lead a mentally active life — reading, learning things, playing, keeping up social relationships, cognitively demanding work — tend to show less cognitive decline and fewer dementia diagnoses than those who do not.
The problem is that those studies are observational, and you cannot conclude causality from them, for two reasons worth understanding:
- Reverse causality. Cognitive decline begins silently years before diagnosis. It may be that people stop doing crosswords because the decline has started, not that they decline from not doing them.
- Confounding factors. Someone who does puzzles daily also tends to have more years of education, better socioeconomic circumstances, more social life and better access to healthcare. Separating the puzzle's effect from the rest is very hard.
The concept usually invoked here is cognitive reserve: the idea that a lifetime of mental activity builds a kind of buffer allowing more brain damage to be tolerated before symptoms appear. It is a reasonable and supported hypothesis, but it points to a lifestyle sustained over decades, not to twenty minutes of sudoku a day.
What does have more evidence behind it
If the aim is looking after your head long term, there are habits with considerably more solid backing than puzzles:
- Regular physical activity. Probably the one with the best accumulated evidence, and it improves sleep and mood too.
- Sleeping well and enough. Sleep is when memory consolidates; chronic deprivation shows up in everything.
- Controlling blood pressure, blood sugar and cholesterol. Vascular and brain health go closely together.
- Treating hearing loss. It is one of the modifiable risk factors that has risen most in recent reviews, probably because the isolation it causes cuts stimulation drastically.
- Not smoking and moderating alcohol.
- An active social life. Conversation is one of the most cognitively demanding tasks there is.
- Keeping on learning new things. Learning a language or an instrument looks more promising than repeating an already mastered task, because it demands constant acquisition rather than mere execution.
If you had to choose between half an hour of crosswords and half an hour of brisk walking, the evidence points to the walk.
So what are puzzles good for?
For quite a lot, if you ask the right thing of them:
- They are enjoyable. And that is not a consolation prize: adherence is everything. A habit you like you keep for years; one that bores you you drop in a fortnight.
- They give structure and a sense of achievement. Finishing something every day, at a suitable difficulty, is a good antidote to listlessness and works very well as a routine.
- They keep up the habit of concentrating. In an environment designed to fragment attention, spending ten unbroken minutes on a closed problem has value in itself.
- They are a social excuse. The shared crossword, the game with someone, comparing results: that is where social life slips in, and social life does have evidence behind it.
- They replace something worse. Ten minutes of puzzle in a waiting room competes with ten minutes of infinite scrolling. There is no doubt which of the two leaves you feeling better.
How to think about it without kidding yourself
The reasonable approach is to treat puzzles as what they are: an enjoyable activity that trains specific skills and sustains a daily habit of attention. Neither a pill against dementia nor a waste of time.
And if you are going to do them, two ideas that do fit what is known:
- Vary them. Repeating the same puzzle means you automate it, and an automated task demands far less. Mixing different types keeps the demand up.
- Raise the difficulty. Learning happens at the edge of what you can do. If you finish every day without effort, it has stopped being training and is just habit.
Those two things, incidentally, are the reason MenteBrillante has fifteen different puzzles rather than one, five different challenges every day, and a difficulty that rises as you solve. It plays free in the browser.
If you want to start with the practical side, we have specific guides to sudoku and hard sudoku techniques.