Can strength training make your brain look younger?

Does Lifting Weights Slow Brain Ageing? A Trial in 309 Adults Aged 62–70

2026-10-19 · By Anthony Falcone

Verdict: Keep it, with a caveat. Probably a little, but keep the headline in proportion. Both heavy and moderate strength training lowered an MRI-based brain age estimate by about 1.4 years after a year of training and about 2 years a year later, while the no-exercise group didn't change significantly. The measure is rough and the extra benefit over doing nothing was smaller than those figures suggest.

What the researchers did

"Brain age" has become a popular way to talk about longevity. The idea: train a computer model on brain scans from many people of known age, then use it to estimate how old a new person's brain looks. The difference between that estimate and your real age is called the brain age gap. A negative gap means the brain looks younger than expected.

This study used the Live Active Successful Aging (LISA) trial in Copenhagen, Denmark. It randomly assigned 309 healthy people aged 62–70, all approaching retirement age, to one of three groups for a year. The heavy-training group did three supervised sessions a week at a training centre. The moderate group did one supervised session a week plus two unsupervised sessions at home. Both programmes mixed resistance exercises with functional training aimed at strength, endurance and balance, set individually from each person's baseline tests and increased gradually. The control group was asked to keep their usual habits, which meant less than an hour of strenuous activity a week.

Everyone had a resting-state brain scan (functional MRI, which maps how brain regions activate together while you lie still) at the start, after the one-year programme, and again one year after it ended. The researchers built their brain clock from the scans of 2,433 other healthy adults, using 6,670 connections between brain regions, and then applied it to the trial participants. The people analysing the scans didn't know who was in which group.

What they found

Both training groups ended up with a lower brain age gap. In the heavy group it fell by 1.42 years after the year of training and by 1.85 years at the two-year scan. In the moderate group it fell by 1.39 and 2.26 years. The control group also drifted down, by 0.66 and 0.82 years, but those changes were not statistically significant. The overall test of whether the groups changed differently over time was significant (p = 0.031), although the size of that group-by-time effect was small.

Notice what that means: the paper's headline numbers are each group compared with itself, not with the control group. A rough subtraction gives the extra change over doing nothing: about 0.8 years after a year of training (1.42 − 0.66 = 0.76 for heavy, 1.39 − 0.66 = 0.73 for moderate) and about 1 to 1.4 years at two years (1.85 − 0.82 = 1.03, 2.26 − 0.82 = 1.44). That is our arithmetic, not a test the paper ran.

The moderate, mostly home-based programme did as well as three supervised sessions a week on brain age. Only the heavy group, though, showed a separate change on the scans: stronger connectivity in a large area at the front of the brain, a region involved in attention, executive control and working memory, compared with the control group. Moderate training showed no such difference.

The effect showed up only when the clock looked at the whole brain. Versions of the clock limited to single networks, such as the motor or default mode network, found nothing significant. And the link with fitness was weak: across everyone, bigger gains in leg strength went with a slightly larger drop in brain age (correlation of −0.12), but within the groups that held only for moderate training.

How much to trust a brain clock

The model's average error when estimating one person's brain age was about 8.7 years (8.652 in the paper). That doesn't make the group averages useless, because errors partly cancel out across a hundred people, but it does mean a change of one or two years is a small signal measured with a noisy tool. A brain clock reading is also not the same as better memory or a lower dementia risk; this paper didn't test either.

The authors describe the 1–2 year reduction as modest and in line with what other brain age studies find for lifestyle factors, and say its clinical meaning needs testing with longer follow-up.

What this means for you

Strength training in your 60s is worth doing for many reasons, and this trial adds a hint that it helps how the brain ages, without needing a gym every day.
Moderate counts: one supervised session plus two home sessions a week matched three supervised sessions on brain age in this trial.
Treat "brain age" numbers, from a study or a commercial test, as rough estimates, not a score to chase.
If you have a heart condition, joint problems or another medical condition, talk to your doctor before starting a strength programme.

How much to trust it

The authors' own: participants were healthy older adults from a high-income European country, so the results may not apply to other populations or to people with health conditions; the clinical meaning of a 1–2 year change still needs longer follow-up; network-level clocks found no effect; and the links with muscle strength were small and inconsistent.
Brain age here is an estimate from a model with an average error of about 8.7 years per person. The paper didn't measure memory or thinking, so it can't say whether people's brains actually worked better.
Not everyone was in every comparison. The one-year results include 73–75 people per group, out of 101–105 randomised, and the paper doesn't explain the gap.
The headline reductions compare each group with its own starting point. The control group also improved slightly, so the extra benefit of training is smaller, roughly 0.7 to 1.4 years by our arithmetic.
The paper has small inconsistencies between its text and its table and figure (for example 1.84 vs 1.85 years); we've used the table values. Its demographics table also lists ages from 60 to 71, while the methods say 62–70. It also refers to the main trial paper for the exact exercises, sets and loads, so we can't give them here.

The study

Gonzalez-Gomez R, Demnitz N, Coronel C, Gates AT, Kjaer M, Siebner HR, Boraxbekk CJ, Ibanez AM. Randomized controlled trial of resistance exercise and brain aging clocks. GeroScience, 2026. doi.org/10.1007/s11357-026-02141-x

Our Healthy Ageing 40+ protocol builds strength in with two sessions a week, gym or home, alongside easy cardio with one interval session, protein targets for over-40s, a daily 5–10 minute balance routine, and a bone, sleep and health-check guide. See the protocol