Do number puzzles help memory? evidence from studies

The solver studies a 4×4 grid with one empty square and makes no guess about what goes in it. The row holding the empty square gets her attention first, and it already contains a 1, a 3, and a 4. In the column with the empty square, she then finds a 2 and a 3. She turns next to the 2×2 block holding the empty square, which has a 4 and a 1 in it. Since only one number fits the square, she writes a 2.

The habit of keeping a few numbers in mind and discarding the ones that cannot fit is the reason many say that number puzzles help memory. Many readers ask whether this practice really improves memory, and whether any benefit lasts for years or ends when the puzzle is finished.

The outcomes of the studies described below differ, and all of them are observational. For each study, the sections give what was measured, who took part and what the authors conclude the findings do and do not show.

What the 2019 exeter analysis measured

In 2019, the International Journal of Geriatric Psychiatry published a study from the University of Exeter. Its data came from 19,078 healthy volunteers aged 50 to 93, all of whom belonged to the online PROTECT study. Participants reported how often they solved number puzzles, and the researchers set those answers beside the volunteers' results on a series of computer-based tasks.

Attention, reasoning and memory were what the tasks assessed. Participants who took part in word and number puzzles more regularly performed better across those tasks, the authors reported. People who reported different habits were compared with one another, and the same people were not measured before and after they began puzzles.

How the differences changed with age was also described by the authors. The findings put people who do word puzzles at the level of someone 10 years younger on grammatical reasoning tests and 8 years younger on short-term memory tests. In the study sample, these numbers describe how test scores relate to puzzle habits. Dr. Helen Brooker, the lead researcher, said the team could not conclude that playing these puzzles automatically lowers dementia risk.

Why an association is not proof of cause

The lead researcher, Dr. Helen Brooker, said the team could not state that playing these puzzles necessarily reduces dementia risk. Regular use helping brains work better for longer was an earlier finding, and she said this work supports it. A claim about cause is what the first statement makes, and the data in this analysis do not extend that far.

An independent expert who spoke to the Science Media Centre about the 2019 Exeter papers made a similar point. The data show only a link at one point in time. The expert said they could not exclude the idea that people who are more intellectually capable might simply enjoy puzzles more in the first place. Under that explanation, higher test scores and frequent puzzle use would appear together even when puzzles had no effect on memory.

A survey that asks people to describe their own habits and then records the answers can show that those who say they do puzzles also tend to score better on tasks of attention, reasoning, and memory. What the participants were like before they started is not shown by the data, and neither is whether puzzles caused any change. People may also have reported their habits imperfectly. Associations are what this type of survey can show, and it gives no sign of which factor came first.

What the 2018 BMJ study and later trials add

The BMJ published a study in 2018 that followed 498 volunteers, about 64 years old when it began, for 15 years. Later in life, people who stayed intellectually engaged did better on cognitive tests, and problem-solving had the strongest link to better performance among the activities examined. Engaged participants scored higher, yet they lost ability over time at the same rate as less engaged participants, so greater engagement did not change the pattern of cognitive decline with age.

A 2020 study used functional near‑infrared spectroscopy to monitor activity in the prefrontal cortex while 33 healthy adults solved a number grid in two steps, following heuristic rules. The authors proposed that this puzzle could serve as cognitive remediation training for certain neuropsychiatric disorders. The study recorded brain activity during the solving process rather than measuring memory test scores over time.

The University of Hong Kong conducted a trial registered as NCT04913857. It tested a program named SMART that uses number grids. The trial involved participants who had mild cognitive impairment and whose Montreal Cognitive Assessment scores ranged from 19 to 26. The study began in September 2020 and finished in December 2023. Its design allows researchers to determine whether the program has an effect on memory, something that earlier studies could not assess.

NumFall is made by the publisher of this site. Any health or brain-training points here are general information, not medical advice.