One Year of Walking Grew the Part of the Brain That Shrinks With Age

The hippocampus, central to memory formation, loses roughly 1 to 2% of its volume per year in late adulthood. That decline is one of the more reliable structural changes in the ageing brain, and it tracks with declining memory performance.

A 2011 trial in PNAS asked whether it could be reversed. The design was straightforward: 120 older adults randomised either to a moderate aerobic walking programme or to a stretching and toning control, for one year, with MRI scans before and after.

The walking group increased anterior hippocampal volume by approximately 2%. The authors described this as effectively reversing one to two years of age-related loss. Spatial memory performance improved alongside it. The control group's hippocampus continued to shrink over the same period, at roughly the expected rate.

One detail does more to establish credibility than the headline number. The researchers also measured the caudate nucleus and the thalamus, and neither changed. If the 2% figure were an artefact of scanner drift, segmentation error or general measurement noise, that noise should have shown up in the other structures too. It did not. The effect was regionally specific, which is what a genuine biological effect looks like and what a methodological artefact usually does not.

The mechanistic account involves exercise-induced increases in brain-derived neurotrophic factor and improved cerebral perfusion, both of which support hippocampal neurogenesis and synaptic maintenance. That account is reasonable but was not directly demonstrated by the imaging itself.

The important caveat is what happened next. Subsequent attempts to replicate the volumetric finding have produced mixed results — some studies have found smaller effects, some none. This is a common trajectory for a striking early trial, and it does not mean the original was wrong. It does mean the honest description is "one well-designed trial with an unusually clean internal control, not yet consistently replicated," rather than "exercise grows your hippocampus."

It is also worth noting the intervention was genuinely modest. This was moderate-intensity walking, three sessions a week, building to roughly 40 minutes per session. Not high-intensity intervals, not resistance training, not anything requiring equipment or unusual commitment. Whatever the true effect size, the dose that produced it was reachable for most people.

The broader point is that structural brain change in later life is not purely a one-way process. The prevailing assumption for a long time was that adult brain volume only declined, and the interesting question was how fast. This trial, whatever its final replication status, contributed to shifting that assumption.

Why this matters for tracking: the useful signal here is aerobic consistency over months, not intensity within a session. If you are logging training, the relevant metric for this particular outcome is the number of moderate aerobic sessions accumulated across a quarter, rather than peak heart rate or pace. The trial ran for a year before it measured anything — a timescale worth matching in how you evaluate your own data.