The guide / Part 11

Do brain games and crossword puzzles really help your memory?

By The Prolongeva Editorial Team Written from the published research cited in each article Updated October 2026

Affiliate disclosure: we are an independent site and may earn a commission if you buy Neuro Serge through links on this site, at no extra cost to you.

You have seen the apps, the daily puzzle and the ads promising a sharper mind in minutes a day. If you are in your 50s and have started to notice small lapses, it is a fair question: does any of it actually help? We opened the studies. The answer is more careful than the marketing.

What do brain games clearly improve?

One thing is not in dispute: practice makes you better at the game. In a 2016 review in Psychological Science in the Public Interest, Simons and colleagues examined the peer-reviewed studies cited by leading brain-training companies. They found extensive evidence that training improves performance on the trained tasks.

That is real, but it is not the same as a better memory for names, appointments or where you left your keys. The question that matters is whether the skill travels.

Do the gains carry over to real life?

This is where the evidence gets thin. Two other sources we opened point the same cautious way, though they are not identical.

A Cochrane review (Gates and colleagues, 2020) looked at computerised training lasting 12 weeks or more in cognitively healthy adults aged 65 and over. It identified eight randomised trials with 1,183 participants in total. Versus active controls, its slight improvement in global cognition at the end of the 12-week intervention (SMD -0.31, 95% CI -0.57 to -0.05) came from only 2 studies and 232 participants, and was rated low-quality evidence. Also versus active controls, it found little or no difference in episodic memory (4 studies, 439 participants) and working memory (3 studies, 392 participants). Versus inactive controls, it reported a slight improvement in episodic memory (1 study, 150 participants, low-quality evidence). It called the benefit of uncertain clinical importance and found no evidence the global effect lasted 12 months later.

A 2020 meta-analysis in Scientific Reports (Bonnechere, Langley and Sahakian) was a little more positive. It pooled 16 studies with 1,543 participants over 60 and found small to moderate gains in some areas, such as processing speed, and none for attention or visuospatial ability. It also noted that only 3 of the 16 studies used intention-to-treat analysis and that some of the studies had relatively small sample sizes.

What did the big ACTIVE trial find?

ACTIVE is the study people usually mean. It was a randomised trial across six US cities with 2,832 older adults (mean age 73.6 at the start). Groups received ten sessions of memory, reasoning or speed-of-processing training, with booster sessions at 11 and 35 months.

At the 10-year follow-up (Rebok and colleagues, Journal of the American Geriatrics Society, 2014), all three trained groups reported less difficulty with everyday activities than controls. At a mean age of 82, about 60% of trained participants versus 50% of controls were at or above their baseline level of self-reported daily function. Reasoning and speed training, but not memory training, improved the targeted cognitive abilities for 10 years.

Note two cautions. The original ACTIVE report (Ball and colleagues, JAMA, 2002) describes ten-session group training delivered in the trial setting, not a phone app. And the daily-function result was self-reported.

In our reading, this is an interesting result and not a settled one. It is one trial, one type of training, a subgroup defined by who completed boosters, and a diagnosis recorded in insurance claims. We found nothing in it that tests a commercial game.

What about crosswords and puzzles?

The best crossword evidence we found is in people who already have mild cognitive impairment (MCI). The background section of a 2025 paper (Wang and colleagues, International Journal of Clinical Trials) states that in a recently completed trial, crossword puzzles were superior to computerised cognitive training on specific cognitive measures and functional outcomes, with findings correlating to decreased brain atrophy over the study period. The paper describes the follow-up trial, COGIT-2: 240 people with MCI, 78 weeks, high-dose versus low-dose home crosswords against a health education control.

We only read the abstract, which gives no numbers, so we cannot report the size of the effect. And it does not show that crosswords help healthy adults, or that they prevent dementia. Enjoying one is a fine reason to do it; whether it protects memory in people without MCI, we found no evidence either way.

What else is worth doing?

The studies above do not tell us what else protects memory, and we will not borrow their authority for other claims. Our other guides cover those topics with their own sources:

• Sleep and memory after 45 — how sleep relates to memory, and when to ask about it.
• 7 daily habits that protect your memory — everyday habits and what the evidence says about them.
• Midlife forgetfulness — what is commonly reported and what is not.

In our reading, a reasonable approach is to treat games as a hobby you enjoy, not as a treatment, and to be wary of any ad that promises more than the trial evidence shows.

When should you see a doctor?

Puzzles cannot tell you whether a change in your memory matters. See a doctor if lapses are new, getting worse over months, noticed by people close to you, or paired with getting lost, repeating questions or trouble with familiar tasks. Many causes are worth checking, and a clinician can judge your case in a way an article cannot.

For how doctors separate common age-related changes from early dementia, see normal aging vs. early dementia.

This article is general information, not medical advice. It summarises a small number of studies and cannot judge your situation. If you are worried about your memory, see a doctor.