Exercise and cognitive aging

Exercise is the most thoroughly evidenced modifiable factor for preserving cognitive function as you age. Two decades of randomized controlled trials show it produces measurable structural changes in the brain, improves performance across multiple cognitive domains, and, when combined with other lifestyle interventions, demonstrably slows cognitive decline in older adults at improved risk.

What the trials show

A 2018 meta-analysis of 39 RCTs covering 333 effect sizes found that exercise improved cognitive function in adults 50 and older with an overall effect size of g = 0.29 (95% CI 0.17-0.41, p < 0.01) [1]. Sessions of 45-60 minutes at moderate-to-vigorous intensity produced the strongest results. Aerobic exercise, resistance training, multicomponent programs, and tai chi all showed benefit, no single modality clearly dominated [1].

An earlier meta-analysis by Colcombe and Kramer, drawing on 18 RCTs from 1966-2001, found roughly 0.5 SD improvement across cognitive tasks, with executive-control processes, planning, inhibition, task-switching, showing the largest gains [2]. Combined aerobic and strength training outperformed aerobic exercise alone [2]. This matters because executive function is precisely the cognitive domain most vulnerable to normal aging.

Exercise grows the hippocampus

The most compelling structural evidence comes from a one-year RCT by Erickson and colleagues (n = 120 older adults) comparing moderate aerobic walking against a stretching control. One year of aerobic exercise increased hippocampal volume by approximately 2%, reversing an estimated 1-2 years of age-related atrophy [3]. The control group lost roughly 1.4% of hippocampal volume over the same period, putting the between-group difference at 2-3%.

The hippocampal volume gain correlated with improved spatial memory and with increased serum BDNF (brain-derived neurotrophic factor), pointing to BDNF-driven neuroplasticity as a plausible mechanism [3]. The caudate nucleus and thalamus showed no change, this was not a global "exercise inflates brain volume" effect, but a regionally specific one. That said, the volume increase does not confirm neurogenesis per se; angiogenesis, glial changes, or increased synaptic density are alternative structural explanations.

Slowing decline in at-risk adults: the FINGER trial

The FINGER trial (n = 1,260 older adults in Finland; two years) embedded exercise within a multimodal program including dietary counseling, cognitive training, and vascular risk monitoring, compared against general health advice. The intervention produced a statistically significant slowing of cognitive decline: NTB composite between-group difference of 0.022 per year (95% CI 0.002-0.042, p = 0.030) [4]. Executive function and processing speed showed the largest domain-specific benefits.

FINGER was the first RCT to demonstrate that a multidomain lifestyle intervention can improve cognitive trajectories in at-risk older adults [4]. The World-Wide FINGERS network is replicating this design across different countries; results are pending.

The dementia prevention picture

The 2020 Lancet Dementia Commission (Livingston et al.) identified physical inactivity as one of 12 modifiable risk factors collectively accounting for roughly 40% of dementia cases worldwide [5]. Verhoeven's analysis notes that this is a synthesised population-level estimate, not a direct experimental finding, the commission treats the observational literature as roughly causal, which requires caution. The FINGER trial [4] and the exercise meta-analyses [1, 2] provide the closest causal anchors for that population-level framing.

How exercise likely protects the brain

Three complementary pathways emerge from the evidence. First, exercise upregulates BDNF, supporting hippocampal neuroplasticity, the Erickson trial's BDNF-volume correlation is the strongest human primary evidence for this [3]. Second, aerobic exercise improves endothelial function and cerebral blood flow, which plausibly drives executive function gains, though direct measurement of this mechanism in the key RCTs cited here is limited. Third, exercise reduces blood pressure, insulin resistance, and adiposity, all independently linked to dementia risk [5], suggesting that some cognitive benefit is vascular in origin, not purely neurotrophic.

Evidence-informed dose parameters

Studies point toward 45-60 minute sessions at moderate-to-vigorous intensity [1], with combined aerobic and resistance training preferred over aerobic exercise alone [2]. Measurable hippocampal volume change required at least one year of consistent training [3]. Meta-analyses did not specify precise weekly frequency; standard public-health guidelines of 150 minutes per week of moderate activity serve as the practical anchor.

Where the evidence is contested

The FINGER trial's NTB difference of 0.022 per year is statistically significant, but its clinical meaningfulness remains debated. Both groups improved over two years, likely reflecting practice effects, which compress the net between-group signal [4]. Because exercise was embedded in a multimodal bundle, this design cannot isolate exercise's specific contribution.

On mechanism, the Erickson 2011 hippocampal volume finding is solid, yet what underlies that volume change, neurogenesis, angiogenesis, or synaptic remodeling, remains unresolved [3].

A broader confounding concern applies across the observational studies feeding into commission-level estimates. Cognitively sharper people may be more physically active ("healthy adherer" confounding), so causation cannot be fully disentangled from correlation. RCT evidence mitigates but does not eliminate this concern. Trial durations of two years or less are typical, and whether structural and cognitive gains persist after training stops remains untested at adequate scale.

References

1. Northey JM, Cherbuin N et al. 2018 (BJSM) — 39-RCT meta-analysis; 333 effect sizes; n=4,884 (50+ y); g=0.29 (95% CI 0.17–0.41, p<0.01); aerobic,… · PubMed 28438770
2. Colcombe S, Kramer AF. 2003 (Psychol Sci) — 18 RCTs, 1966–2001; overall ~0.5 SD across cognitive tasks; executive-control processes showed largest… · PubMed 12661673
3. Erickson KI, Voss MW, Kramer AF et al. 2011 (PNAS) — n=120 older adults; 1-yr aerobic walking vs stretching control; hippocampal volume +2%… · PubMed 21282661
4. Ngandu T, Solomon A, Kivipelto M et al. 2015 (Lancet) — FINGER RCT; n=1,260; 2yr; Finland; multidomain (diet + exercise + cognitive training +… · PubMed 25771249
5. Livingston G, Huntley J, Sommerlad A et al. 2020 (Lancet) — Dementia Prevention, Intervention and Care Commission; physical inactivity as 1 of 12… · PubMed 32738937