Puzzle Books and the Brain

Puzzle books do something real in the brain. The harder question is how far that something travels.

A crossword, a Sudoku grid, or a logic book is not a full cognitive training program. It is a narrow, repeatable demand: hold several constraints in mind, retrieve a word or a number, check it against the rest of the grid, and try again. Imaging work and large aging studies agree that this demand is not empty. They also agree that the popular claim — “puzzles keep you ten years younger” — is only partly earned.

What a puzzle actually recruits

Sudoku is the cleanest case, because the rules are fixed and the load is mostly working memory and executive control. A 2026 systematic review of six brain-imaging studies (five fMRI, one fNIRS) found a consistent pattern. Solving a grid recruits the frontoparietal network: the dorsolateral prefrontal cortex, involved in holding and manipulating information, and parietal regions involved in visuospatial working memory. The anterior cingulate also comes online, which fits the constant self-checking the puzzle requires — “does this digit break the row, the column, or the box?”

That is the same circuitry people use for planning, monitoring errors, and keeping several options alive at once. Crossword solving leans more on language: semantic retrieval, spelling, and the ability to switch between a clue and the intersecting answers already filled in. Number puzzles lean more on logic and visuospatial working memory. Neither is a whole-brain workout, but both are real cognitive work, not passive entertainment.

Sharper scores, with a caveat on cause

The largest puzzle-specific dataset comes from the PROTECT platform, run by the University of Exeter and King’s College London. In two linked 2019 papers, researchers asked 19,078 adults aged 50 to 93 how often they did word puzzles and number puzzles, then tested attention, reasoning, and memory.

Frequency tracked with performance. People who did word puzzles regularly scored, on grammatical reasoning, at a level the authors described as equivalent to someone about ten years younger, and on short-term memory about eight years younger. Number-puzzle users showed a similar pattern, on the order of eight years on problem-solving measures. The authors were careful: they could not say the puzzles caused the difference, only that regular use went with better function across those tests.

That caution matters. People who already have sharper brains, more education, or more cognitively active lives are also more likely to buy puzzle books. A study from the University of Aberdeen that followed people for years found that crossword users started with a higher cognitive baseline, but the puzzles did not clearly slow the later rate of decline. Higher starting point and slower slope are different claims. The PROTECT result is the first. It is not, by itself, the second.

Dementia risk: delay more than prevention

Longer studies point to a delay in when problems show up, not a guarantee they never will.

In the Bronx Aging Study, 488 initially intact older adults were followed with repeated memory testing. Among the 101 who later developed dementia, people who did crossword puzzles at baseline reached the point of accelerated memory decline about 2.54 years later than non-puzzlers. Education and other stimulating activities did not add much once puzzles were in the model. There was a twist: once decline started, it was somewhat steeper in the puzzle group. The practical reading is a later onset, not a permanently flatter curve (Pillai et al., 2011).

An earlier Bronx cohort, published in the New England Journal of Medicine in 2003, looked at leisure activities more broadly — reading, writing, board games, crosswords, musical instruments. A one-point rise on the cognitive-activity score was linked to a lower risk of dementia. In their model, older adults who did crosswords four days a week had about a 47 percent lower dementia risk than those who did them once a week (Verghese et al., 2003).

A 2021 Neurology study from Rush University, of 1,978 people with an average age of 80, found that late-life cognitive activity — reading, writing letters, cards, board games, puzzles — was associated with a later diagnosis of Alzheimer’s dementia. The most active group developed dementia at an average age of 94; the least active at 89, a five-year gap. The authors’ point was that it was not too late to start, even in the 80s (Wilson et al., 2021).

Australian data from the ASPREE and ALSOP studies, covering 10,318 adults aged 70 and older, put a smaller number on a related idea. Frequent adult literacy activities and active mental tasks — quizzes, crosswords, cards, chess — were associated with about a 9 to 11 percent lower risk of dementia, a bit more than creative hobbies or passive reading.

These are observational. Healthier, better-educated, more socially engaged people do more puzzles. The studies adjust for some of that, not all of it.

The one trial that surprised people

The strongest direct test is a randomized trial published in NEJM Evidence in 2022, led by teams at Columbia and Duke. One hundred seven adults with mild cognitive impairment, most around age 71, were assigned to 12 weeks of computerized crossword training or computerized cognitive games, plus booster sessions out to 78 weeks.

Crosswords won on the primary outcome. On the ADAS-Cog, a standard measure of cognitive impairment, the games group worsened slightly over 78 weeks and the crossword group improved. The difference was statistically significant. Crosswords also did better on a measure of daily function at 78 weeks, and participants in the crossword arm showed less hippocampal and cortical shrinkage on MRI. The advantage was clearer in people who were further along in mild cognitive impairment (Devanand, Doraiswamy, et al., 2022).

That does not mean a drugstore puzzle book is a treatment. The crosswords were structured, computerized, and dosed like an intervention. It does undercut the assumption that newer “brain games” automatically beat an old paper grid.

Where the claim breaks

Two limits show up again and again.

First, practice makes you better at the thing you practice. A large online experiment by Owen and colleagues, with more than 11,000 participants, found that people improved on the trained tasks and did not reliably transfer those gains to untrained reasoning, memory, or planning. Meta-analyses of cognitive training reach a similar conclusion: near transfer is real, far transfer to everyday life is weak. Getting faster at Thursday’s crossword is not the same as getting better at managing medications or following a conversation in a noisy room.

Second, novelty may matter more than familiarity. In the Synapse Project, Denise Park and colleagues compared older adults who did familiar activities such as puzzles with those who learned a genuinely new, demanding skill, such as digital photography or quilting. Only the new-skill group showed clear gains in memory and signs of neural change. A puzzle book you have done for forty years is still useful stimulation. It is probably not the same stimulus as learning something you cannot yet do.

Mood and routine belong in the picture too. A short daily puzzle can structure attention, give a small sense of completion, and interrupt rumination. Those are real psychological effects. They are not the same as building cognitive reserve, the extra network capacity that seems to let some people function well even as Alzheimer’s pathology accumulates.

A practical reading

The evidence supports a modest claim, not a miracle one.

Regular word and number puzzles are associated with better attention, reasoning, and memory scores in later life, on the order of several years of “younger” performance in cross-sectional data. In people who go on to develop dementia, crossword use has been linked to a delay of roughly two and a half years before accelerated memory decline, and broader late-life cognitive activity to a delay of up to about five years before an Alzheimer’s diagnosis. In one randomized trial in mild cognitive impairment, structured crossword training beat commercial-style cognitive games over a year and a half, including on a brain-volume measure.

None of that proves a puzzle book prevents dementia. The honest mechanism is cognitive reserve plus domain practice: you keep language and executive circuits in use, you get better at the specific task, and you may push the clinical threshold later. Exercise, sleep, blood pressure, hearing, and social contact still carry more weight than the puzzle page.

If the goal is a brain that stays useful, a puzzle book is a reasonable daily habit — cheap, self-paced, and backed by better evidence than most products sold as brain training. It works best as one piece of a larger pattern: something slightly hard, done often, alongside things that are new.

Selected references

• Brooker, H., Wesnes, K. A., Ballard, C., Hampshire, A., Aarsland, D., Khan, Z., Stenton, R., Megalogeni, M., & Corbett, A. (2019). The relationship between the frequency of number-puzzle use and baseline cognitive function in a large online sample of adults aged 50 and over. International Journal of Geriatric Psychiatry.
• Brooker, H., et al. (2019). An online investigation of the relationship between the frequency of word puzzle use and cognitive function in a large sample of older adults. International Journal of Geriatric Psychiatry.
• Devanand, D. P., Goldberg, T. E., Qian, M., Rushia, S. N., Sneed, J. R., Andrews, H. F., Nino, I., Phillips, J., Perea, E. T., Doraiswamy, P. M., et al. (2022). Computerized games versus crosswords training in mild cognitive impairment. NEJM Evidence.
• Owen, A. M., Hampshire, A., Grahn, J. A., Stenton, R., Dajani, S., Burns, A. S., Howard, R. J., & Ballard, C. G. (2010). Putting brain training to the test. Nature.
• Park, D. C., Lodi-Smith, J., Drew, L., Haber, S., Hebrank, A., Bischof, G. N., & Aamodt, W. (2014). The impact of sustained engagement on cognitive function in older adults: The Synapse Project. Psychological Science.
• Pillai, J. A., Hall, C. B., Dickson, D. W., Buschke, H., Lipton, R. B., & Verghese, J. (2011). Association of crossword puzzle participation with memory decline in persons who develop dementia. Journal of the International Neuropsychological Society.
• Verghese, J., Lipton, R. B., Katz, M. J., Hall, C. B., Derby, C. A., Kuslansky, G., Ambrose, A. F., Sliwinski, M., & Buschke, H. (2003). Leisure activities and the risk of dementia in the elderly. New England Journal of Medicine.
• Wilson, R. S., Wang, T., Yu, L., Grodstein, F., Bennett, D. A., & Boyle, P. A. (2021). Cognitive activity and onset age of incident Alzheimer disease dementia. Neurology.
• Wu, Z., et al. (2023). Leisure activities and the risk of dementia. ASPREE / ALSOP analyses.
• Neural correlates of Sudoku play: a systematic review of brain imaging studies. Frontiers in Neuroimaging, 2026.