Daily Steps and Higher Amyloid: What a Harvard Study Found

Older couple holding hands while walking along a tree-lined park path on a winter day
Photo by Duc Tinh Ngo on Pexels

Written by

–

Drafted with AI assistance and reviewed by Kurt Greiner. How we use AI.

You may have seen the headline that just 3,000 steps a day could help slow Alzheimer’s. The real finding is more interesting, and more useful. In a study of 296 older adults from the Harvard Aging Brain Study, all with healthy thinking at the start, people who walked more showed slower buildup of a brain protein called tau and slower decline in memory and daily function over about nine years. That association held only for people who already had higher levels of amyloid, an early Alzheimer’s marker. The association leveled off at roughly 5,000 to 7,500 steps a day. Observational data cannot show how much activity is needed, so the study does not say whether 10,000 steps adds anything. This was an observational study, so it shows an association, not proof that walking caused the difference.

What the researchers did

The team used data from the Harvard Aging Brain Study. All 296 participants were older adults with normal thinking at baseline. Each wore a pedometer at the start, and the researchers then followed their thinking and daily function with yearly tests for up to 14 years, with a median follow-up of 9 years. Researchers call the stage where amyloid is elevated but thinking is still normal preclinical Alzheimer’s disease. A subset of 172 people also had repeated PET scans that track tau in the brain.[1]

They sorted people into four step groups: 3,000 or fewer a day (“inactive”), 3,001 to 5,000 (“low”), 5,001 to 7,500 (“moderate”) and 7,501 or more (“active”). They then asked whether step counts were linked to changes in amyloid, tau, memory and thinking, and everyday function.[1]

What they found

  • Walking was not linked to less amyloid. The association was not related to lower amyloid levels at the start or over time.[1]
  • It was linked to slower tau buildup. Among people with higher amyloid, more steps went with slower accumulation of tau in the inferior temporal region, a part of the cortex the study measured with PET. The researchers’ mediation analysis found that slower tau buildup statistically accounted for the link between walking and slower memory-and-thinking decline.[1]
  • Thinking and function declined more slowly. In people with higher amyloid, the more active groups had slower decline. Estimated memory-and-thinking (PACC5) z-scores declined by 2.5 over 9 years in the inactive group, versus 1.5 in the low group and 1.1 in the moderate group. That works out to 40% and 54% less decline than the inactive group for the low and moderate groups. Everyday function, measured on a clinical rating scale, worsened more slowly in the moderate and active groups (45% and 51% slower). The low-activity estimate (34%) did not reach statistical significance.[1]
  • The dose-response curve flattens. The largest step-for-step difference in the estimates was between the inactive and low groups. The moderate group showed slower decline again, and the 7,501-plus group looked about the same as the moderate group. The authors suggest 5,001 to 7,500 steps may be a more reachable goal than 10,000 for older people who are sedentary.[1]
  • A model-based timeline. Using a PACC5 z-score of -1.5, the authors’ threshold for impaired cognition, the model estimated that people with higher amyloid would reach it after 6.5 years if inactive, 9.6 years in the low group and 13.6 years in the moderate group. These are estimates from a statistical model, not observed outcomes.[1]

Where the “3,000 steps” headline comes from

The 3,000 figure is the top of the “inactive” group, not a magic dose. In people with higher amyloid, the 3,001 to 5,000 group had slower tau buildup and less estimated decline in memory and thinking than the inactive group.[1] We think that is where the headline’s number came from; the study did not test that. The study can’t show that starting small changes outcomes for any one person. It isn’t a finding that 3,000 steps protects you.

What this study can’t tell you

  • It can’t prove cause and effect. This is an observational study. The authors say they can’t fully rule out reverse causality, meaning that people in the earliest stages of brain change may simply walk less. They call for randomized trials.[1]
  • Steps were counted once. The pedometer reading came from baseline only. It didn’t capture the intensity or history of activity, or things like swimming and strength training that don’t register as steps.[1]
  • The group is narrow. The cohort is highly educated and mostly non-Hispanic white, and it excluded people with symptomatic cerebrovascular disease or poorly controlled vascular risk factors. The results may not carry over to everyone.[1]
  • The link showed up in a specific group. The association appeared in people with higher amyloid, not across the board. Nobody in the study had dementia at the start, so it says nothing about treating the disease once it’s established.[1]

How to use this

  • Find your baseline first. Check your phone or watch for a typical week. If you’re near or under 3,000 steps a day, you’re in the group where the study saw the largest step-for-step difference, and that pattern showed up in people with higher amyloid. You can’t know your own amyloid level from a step counter, so treat this as a reason to start, not a prediction for you.
  • Aim for the next rung first. Moving from under 3,000 to 3,000 to 5,000 is a realistic first target. The study’s pattern leveled off at 5,001 to 7,500 steps a day. It is an observational pattern, not a dose the study tested, so use it as a reference point, not a proven target.
  • Break it up. Three 10-minute walks a day, one after each meal, come to roughly 3,000 steps at a typical pace. That is our estimate, and the study did not test this schedule.
  • Talk to your doctor first if walking is hard. If you have heart disease, balance problems, joint pain or another condition, ask what a safe ramp-up looks like. If you’re worried about memory, a clinician is the right person to ask, not a step counter.

We covered walking speed separately in why your walking speed might be a brain-health signal. That study looked at how fast you walk. This one asks how much you walk, and where the pattern levels off.

Walking won’t guarantee anything about your brain, and this study doesn’t claim it will. What it does give you is a modest, concrete target that is easy to start this week and that older adults with higher amyloid appeared to benefit from.

References

[1] Yau WYW, Kirn DR, Rabin JS, Properzi MJ, Schultz AP, Shirzadi Z, Palmgren K, Matos P, Maa C, Pruzin JJ, Schultz SA, Buckley RF, Rentz DM, Johnson KA, Sperling RA, Chhatwal JP. Physical activity as a modifiable risk factor in preclinical Alzheimer’s disease. Nature Medicine, 2025;31(12):4075-4083. Population: 296 cognitively unimpaired older adults from the Harvard Aging Brain Study with pedometer-measured steps at baseline, yearly cognitive and functional assessments for up to 14 years (median 9 years), and tau PET in a subset (n = 172). DOI: https://doi.org/10.1038/s41591-025-03955-6. Finding as used: higher physical activity was associated with slower cognitive and functional decline in individuals with elevated baseline amyloid, not with lower amyloid burden but with slower amyloid-related inferior temporal tau accumulation, which significantly mediated the association with slower cognitive decline; dose-response associations plateaued at 5,001 to 7,500 steps per day; in individuals with elevated amyloid, estimated 9-year PACC5 decline was 2.5 (inactive, 3,000 or fewer steps), 1.5 (3,001-5,000), 1.1 (5,001-7,500) and 1.2 (7,501 or more) points, with 40%, 54% and 51% less decline than inactive, and CDR-SOB progression 34% (not statistically significant), 45% and 51% slower; modeled time to a PACC5 z-score of -1.5, the authors’ threshold for impaired cognition, was 6.5, 9.6, 13.6 and 12.7 years; limitations stated by the authors: observational design with possible reverse causality, baseline-only pedometer measurement without intensity, duration or non-step activity, and a highly educated, predominantly non-Hispanic white cohort that excluded symptomatic cerebrovascular disease and poorly controlled vascular risk. ↑︎ a b c d e f g h i j k l

Kurt Greiner
Kurt Greiner

Kurt is a digital strategist and IT professional blending emerging technology with practical application, exploring how individuals can leverage modern tools to navigate the second half of life with purpose.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

This website provides wellness information for educational purposes only. It is not medical advice. Consult a healthcare professional before making any health-related decisions or changes.

Bettering Me Wellness 40+ logo
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.