(© Monkey Business - stock.adobe.com)
In A Nutshell
- Exercise on its own had no significant effect on slowing cognitive decline in adults over 50, despite long being promoted as a brain-health staple.
- In middle age (roughly 55–65), doing a wide variety of different activities, not any single one, was linked to slower cognitive decline, with a protective effect nearly as large as the harm caused by smoking.
- After 65, mentally demanding activities, such as reading, writing, puzzles, and using a computer, emerged as the strongest individual predictor of a slower mental decline.
- Social habits matter differently depending on age: staying in close contact with friends and family benefits the brain most around midlife, while joining organized groups like clubs or religious services appears more protective in later life.
For decades, doctors and public health campaigns have pushed exercise as one of the best things aging adults can do for their brains. New research out of Georgetown University suggests that message, while well-intentioned, is missing something important, and for millions of Americans trying to outrun dementia, that gap matters.
A study published in Innovation in Aging found that physical activity on its own had no significant effect on the rate of cognitive decline in middle-aged and older adults. When researchers tracked more than 20,000 Americans over more than a decade, logging their exercise habits alongside their mental sharpness, going to the gym simply didn’t move the needle. What did matter was whether people were reading, doing word games, staying mentally engaged, and mixing up their daily routines with a variety of different activities.
Recent estimates suggest that 42% of Americans may develop dementia in their lifetime. For anyone counting on a daily walk or a weekly yoga class to protect their memory, this research points toward a more targeted set of habits.
Why Exercise Falls Short in Preventing Cognitive Decline
To investigate what actually predicts cognitive decline over time, researchers at Georgetown and the University of California, Riverside drew on two large national surveys. The Health and Retirement Study (HRS) followed more than 20,800 adults aged 50 and older, interviewing them up to seven times between 2008 and 2020. A second dataset, the Midlife in the United States Study (MIDUS), tracked roughly 2,700 adults ranging from their 30s to their mid-80s across two rounds of cognitive testing about 9.6 years apart.
Participants reported on four categories of activity: cognitive tasks (reading, writing, word games, using a computer), physical activity at various intensity levels, contact with family and friends, and participation in organized social groups like clubs or religious services. Researchers tracked cognitive test scores forward in time, controlling carefully for education, smoking history, diabetes, depression, and dozens of other variables known to affect both activity levels and brain health. Importantly, activities were measured first and cognitive changes were tracked afterward, so the team could see which habits actually came before slower decline rather than simply traveling alongside it.
Physical activity consistently came up short. Whether participants exercised lightly, moderately, or vigorously, their workout habits did not significantly predict how cognitive scores would change in the years ahead. The likely reason: the brain benefits of physical activity may be mostly locked in much earlier in life. Someone who stayed active through their 30s and 40s has probably already built whatever cardiovascular and neurological advantages exercise can offer. Starting or ramping up exercise later in life may not be enough, on its own, to slow cognitive decline once aging-related changes are already underway. The paper also points out that physical inactivity earlier in life can set off a chain reaction, contributing to diabetes and cardiovascular disease, which in turn impair brain function over time. By late life, some of those long-term health effects may be harder to offset with exercise alone.
Exercise is still worth doing. It matters for heart health and physical function, and the researchers found it may offer some cognitive benefit as part of a varied routine. As a standalone strategy for slowing brain aging in mid-to-late life, though, the evidence here didn’t support it.
What Mental and Social Activities Actually Protect the Aging Brain
What did show up as protective: mentally demanding habits, certain kinds of social engagement, and the practice of doing many different types of activities.
Adults in the HRS sample who spread their time across a diverse mix of activities showed slower cognitive decline during midlife, roughly between ages 55 and 65. The size of that benefit was notable. As the researchers wrote, “the positive effect of engaging in many different activities on the rate of cognitive decline during midlife was nearly as large as the negative effect of smoking.” Since smoking is one of the most established drivers of faster cognitive decline, that’s a meaningful comparison.
The diversity advantage faded with age, though it didn’t disappear. By the oldest ages studied, people who had maintained varied routines were still cognitively better off than those who hadn’t, but the gap narrowed. Past a certain point, what specific activities people chose mattered more than how varied their lineup was.
For adults over 65, frequent cognitive activity emerged as the strongest individual predictor of slower decline. Reading regularly, writing, playing cards or chess, working crossword puzzles, using a computer, these mentally demanding habits were tied to markedly better cognitive trajectories late in life. The researchers found the cognitive benefit of frequent mental activity grew larger with age and was ultimately comparable in size to the cognitive toll of having diabetes, which gives a sense of just how meaningful the effect was.
Social activity mattered too, but how it mattered shifted with age. Regular contact with friends and family was linked to slower cognitive decline roughly between ages 55 and 65. Later in life, organized group participation, such as clubs, religious services, or community meetings, showed a stronger association with maintaining cognitive function. One-on-one socializing appears to matter more at 60; structured group activity may count for more at 75.
One important caveat: the MIDUS dataset did not replicate these patterns. None of the activity measures showed a significant link to cognitive decline in that sample, which the researchers attribute primarily to its smaller size and shorter follow-up window.
Taken together, the findings make a fairly direct case. Keeping the mind actively engaged through reading, writing, puzzles, and mentally challenging games appears to be the most reliable late-life investment in cognitive health. And in middle age, building a genuinely varied routine across different kinds of activities may be just as valuable as any single habit, including a trip to the gym.
Disclaimer: This article is based on observational research and is intended for informational purposes only. It should not be interpreted as medical advice. The findings reflect associations identified in specific study populations and may not apply to all individuals. Factors such as genetics, overall health, and lifestyle vary widely from person to person. Readers are encouraged to consult a qualified healthcare provider before making changes to their physical activity, diet, or daily routine based on this or any other research.
Paper Notes
Limitations
Several caveats apply. In the MIDUS sample, the absence of significant prospective findings likely reflects that dataset’s shorter follow-up period (one follow-up assessment roughly nine years later, compared with up to six follow-ups every two years in HRS) and a sample more than 30 times smaller. Detecting small differences in cognitive decline rates requires considerable tracking time and sample depth that MIDUS may not have had.
Activity data in both surveys were self-reported, which means participants with worse cognition, who are more likely to misremember or underreport, could have introduced measurement error. Individuals with severe cognitive impairment were also often excluded from analysis, either because they were institutionalized or because proxy respondents were not administered cognitive tests, making the study better suited to capturing early cognitive decline than later-stage dementia.
Despite the prospective design, some bias likely remains, particularly around cognitive activities. People who are cognitively sharper to begin with may be more inclined to pursue mentally demanding habits, making it difficult to cleanly separate the effect of the activity from the underlying cognitive health that motivated it. The researchers acknowledge their results on cognitive activities are not fully consistent with a meta-analysis of randomized controlled trials, which found no significant effect of cognitive training on test performance.
Funding and Disclosures
This work was supported by the National Institute on Aging (grant numbers P01 AG020166, U19AG051426, and U19AG051426-06A1) and the Graduate School of Arts and Sciences at Georgetown University. The authors declared no conflicts of interest.
Publication Details
Authors: Dana A. Glei, PhD (Center for Population and Health, Georgetown University); Chioun Lee, PhD (Department of Sociology, University of California, Riverside); Casey K. Brown, PhD (Department of Psychology, Georgetown University); Maxine Weinstein, PhD (Center for Population and Health, Georgetown University) | Journal: Innovation in Aging, 2026, Volume 10, Issue 2 | Title: “Does type or diversity of activities delay aging-related cognitive decline?” | DOI: 10.1093/geroni/igaf133 | Published: Advance access February 5, 2026
