What the research actually says about puzzles and the brain

Sudoku has been on the back pages of newspapers since 2004, Wordle landed in late 2021, and crosswords have been a fixture of the New York Times since 1942 — which in puzzle years makes them roughly your great-grandparent. Across all three, and across every brain-training app that's come and gone in the meantime, the same question keeps surfacing: is this actually doing anything for me?

The honest answer is the one cognitive scientists have been giving for about twenty years now, even as they keep tweaking the details. Doing puzzles makes you better at doing puzzles. Whether it does anything beyond that — whether it sharpens your memory at work, slows age-related cognitive change, or wards off dementia — is a smaller, slipperier story than the apps would like you to believe.

This article exists because we'd rather not relitigate it on every other piece on the site. So: the long version, plainly, with the research where the research helps and the limits where the research has them.

Near transfer and far transfer

Cognitive scientists, when they look at all this, tend to split the question into two halves. Near transfer is whether practising a task makes you better at that task itself, while far transfer is whether practising a task makes you better at other, unrelated cognitive tasks: working memory, processing speed, the kinds of decisions you make on a day when nothing in front of you looks remotely like a Sudoku grid.

Near transfer is uncontroversial — practise anything for long enough and you'll get better at it. Sudoku solvers spot patterns faster than non-solvers, crossword regulars have wider vocabularies than non-regulars, and the people who play violin are probably better than you at the violin. None of this is in dispute.

Far transfer is the live question, and for the better part of two decades the evidence has been disappointing. The largest single study to look hard at this was the BBC-backed Brain Training Britain trial, published in Nature in 2010 with about eleven thousand participants: adults trained on cognitive games for six weeks, got better at the games themselves, and showed no measurable gain on anything else¹. A 2014 meta-analysis by Lampit and colleagues, pulling 52 studies together, came back with the same pattern — small effects, mostly within the trained-on tasks, weaker and less consistent the further from the training you looked². Newer reviews haven't really moved the picture, and the consensus settles somewhere in the middle: puzzles aren't doing nothing, but they're doing less than the marketing claims, and the carry-over to your everyday brain is modest at best. We've written a companion piece on the narrow-transfer problem for readers who want to dig further.

What about ageing?

This is where the conversation gets careful, because the stakes are higher and the evidence is fuzzier — and where, frankly, most of the dodgy claims live.

The cognitive reserve hypothesis — the idea that mentally and socially active lives are associated with better cognitive outcomes in later life — is well established as a population-level observation. People who do more mentally engaging activities, on average, hold cognitive function longer than people who do fewer, which is the finding most often cited when puzzle apps want to imply they prevent dementia.

What that finding does not establish is causation: it doesn't show that puzzle-doing specifically causes the protection. People who do puzzles also tend to be the kinds of people who read, maintain social ties, exercise, and watch their cholesterol, and any one of those is a stronger candidate explanation than the puzzle. Cognitive reserve is real; "Sudoku gives you cognitive reserve" is a claim no high-quality randomised trial currently supports.

The same caveats apply to dementia risk: observational studies show associations between mentally active lifestyles and lower dementia incidence, but intervention studies — where researchers actually assign people to do puzzles versus not — have not produced strong evidence that puzzles change the trajectory.

So: regular cognitive engagement, broadly construed, is associated with better outcomes, and a daily puzzle is part of regular cognitive engagement. Whether that puzzle is doing meaningful work, or just signalling the kind of life you live, is unprovable with the studies we have. If you're reading this for yourself and have noticed change in your own thinking lately, what changes in puzzling after 60 is the more direct take. Or if you're reading on someone else's behalf, Sudoku for an older relative is the practical companion to this one.

What's reasonable to say

A few things, plainly.

Doing Sudoku will make you better at Sudoku, and that's real and measurable. A daily puzzle is a defensible small ritual: fifteen minutes of focused logical attention is more demanding than fifteen minutes of phone scrolling, and probably less corrosive to your mood. For older readers worried about cognitive change, a daily puzzle is a low-stakes, low-cost activity that is at worst harmless and at best one of many small contributions to a mentally engaged life; it is not a treatment, nor a substitute for talking to a clinician about real concerns.

A few things we don't say, because the research doesn't support them.

"Train your brain."

"Boost cognition by some specific percentage."

"Prevent dementia."

"Sharpen memory and focus."

Each of those phrases is doing more rhetorical work than the underlying evidence justifies, and we'd rather lose the marginal search ranking than write a sentence we'd be embarrassed to defend in a quiet conversation with a neuropsychologist.

So why do this at all?

Because Sudoku is interesting — the pattern recognition is satisfying, and the arc from a half-empty grid to a finished one feels like something, a small well-shaped thing in a day that often doesn't have many of those. People have always made small daily rituals out of small daily challenges, and a puzzle is a clean, low-cost version of that.

If a fifteen-minute solve replaces fifteen minutes of doomscrolling, your day is probably nudged in a small good direction — not because the puzzle did something profound to your brain, but because you spent fifteen minutes in a calmer, more focused mode than the alternative. That's reason enough on its own, without the puzzle having to be medicine.

And if you're reading through this and wondering whether something specific to your own cognitive picture matters — whether something is changing, whether you should be worried — that's a conversation for a clinician, not a puzzle site. We can write you something honest about Sudoku. We can't write you something honest about you.