Walking 5,000 Steps a Day Linked to Up to Seven Extra Years of Normal Cognition in People With Brain Amyloid
RapAdmin
Researchers at Mass General Brigham followed 296 cognitively normal older adults for up to 14 years. They measured daily steps with a pedometer at the start and tracked brain amyloid and tau with PET scans alongside yearly cognitive testing. In people who already carried elevated amyloid, the early hallmark of Alzheimer’s disease, higher step counts were associated with slower buildup of tau in the inferior temporal cortex and slower decline in memory and daily function. Walking was not linked to amyloid levels at all. Statistical mediation analysis suggested that slower tau accumulation accounted for most of the cognitive association. The relationship was curved: the largest differences appeared between sedentary people and those taking 3,000 to 7,500 steps daily, with little added association beyond that. Because the study is observational, it cannot prove that walking causes these effects.
For years, the case for exercise as Alzheimer’s prevention has rested on large population studies that asked people how active they were and then counted who developed dementia. Those studies had two weaknesses. People misremember how much they move, and dementia diagnoses tell you nothing about what is happening inside the brain years earlier. A new study in Nature Medicine tries to close both gaps.
The team, led by neurologists Wai-Ying Wendy Yau and Jasmeer Chhatwal at Mass General Brigham, drew on the Harvard Aging Brain Study. Participants wore a waistband pedometer for a week at enrollment, then returned annually for cognitive testing for a median of about nine years. Many also had repeated brain scans measuring amyloid plaques and tau tangles, the two proteins that define Alzheimer’s disease.
The first finding was a null one, and it matters. Step counts had no relationship with how much amyloid people had or how fast it accumulated. Walking does not appear to touch the plaque problem.
What walking did track with was tau. Among people whose scans showed elevated amyloid, those who walked more accumulated tau more slowly in the inferior temporal cortex. That is a region where tau spreads as people move from silent disease toward symptoms. Tau is the protein most closely tied to memory loss, and in this analysis the slower tau buildup statistically explained about 84% of the link between walking and slower cognitive decline.
The practical picture is striking. Among amyloid-positive participants, the model estimated that sedentary people (3,000 steps or fewer per day) would cross a threshold for cognitive impairment about 6.5 years after enrollment. Those taking 3,001 to 5,000 steps would reach it at about 9.6 years, and those taking 5,001 to 7,500 at about 13.6 years. Walking more than 7,500 steps showed no extra cognitive advantage in this group. For people without elevated amyloid, step count made no measurable difference to cognition over the study period, likely because they were declining very little to begin with.
This shifts how researchers think exercise might work against Alzheimer’s. Rather than preventing the disease from starting, physical activity may weaken the link between amyloid and the downstream tau cascade. That is a different job from the one anti-amyloid drugs like lecanemab perform, which is why the authors suggest testing the two together.
The caveats are serious. Activity was measured once, for one week, at the start. The most sedentary group was about six years older than the most active group and somewhat more likely to carry amyloid. Only around 21 amyloid-positive people sat in that sedentary reference group, and fewer still had tau scans. People already sliding toward Alzheimer’s may simply move less, which would flip the cause and effect. The authors ran several checks against this, and the results held, but only a randomized trial can settle it.
Still, the message for sedentary older adults is unusually concrete. The target that tracked with protection was not 10,000 steps. It was a brisk afternoon walk’s worth.
Actionable Insights
The clearest take-home is for people who are mostly sedentary. In this study, amyloid-positive adults walking 3,001 to 5,000 steps a day showed about 40% less cognitive decline over nine years than those walking 3,000 or fewer. They also reached the impairment threshold about 3 years later. At 5,001 to 7,500 steps the gap widened to roughly 54% less decline and an estimated 7-year delay, though that figure extends near the edge of the follow-up data. Above 7,500 steps, there was no further cognitive gain in this cohort. Daily function, measured by a clinician rating scale, followed the same pattern, with 34% to 51% slower worsening across the active groups.
Three cautions apply. First, these are associations between groups of people, not results of assigning people to walk, so the true causal effect is probably smaller. Second, the benefit appeared only in people with elevated brain amyloid; blood tests for Alzheimer’s markers now exist and can be discussed with a physician. Third, walking was not associated with any change in amyloid, so it is not a substitute for amyloid-targeting treatment. A reasonable, low-risk action is to move from under 3,000 steps toward 5,000 to 7,500, and to make some of those steps brisk.
Context and Source
- Open Access Paper: Physical activity as a modifiable risk factor in preclinical Alzheimer’s disease (November, 2025)
- Institution: Department of Neurology, Mass General Brigham and Harvard Medical School, Boston, with collaborators at Sunnybrook Research Institute and University of Toronto, Banner Alzheimer’s Institute, and University of Melbourne
- Country: United States (with contributing authors in Canada and Australia)
- Journal: Nature Medicine
- Impact evaluation: The impact score of this journal is 52.5 (2025 Journal Impact Factor, Clarivate JCR released June 2026), evaluated against a typical high-end range of 0 to 60+ for top general science and medicine journals, therefore this is an Elite impact journal.









