Skip to content

Exercise and executive function in Alzheimer's disease: modeled responses varied by training type

A dose-response analysis explored exercise volume and cognitive outcomes, but its abstract does not provide enough detail to turn modeled peaks into an individual prescription.

By 100HP editorialAbstract-based explanation checked

Based on the published abstract. The full paper may contain additional methods, results and limitations.

The 30-second takeaway

The analysis found a nonlinear relationship between exercise volume and executive function in people with Alzheimer's disease. Modeled response patterns differed across training types, while resistance training did not show consistent or statistically significant results. The abstract does not establish a universally optimal exercise amount or demonstrate that one type is superior for an individual.

Keep in mind

This abstract-based account lacks study counts, participant totals, follow-up durations, and uncertainty intervals around modeled effects. These gaps limit assessment of how precise or broadly applicable the apparent dose-response patterns are.

The question was about dose-response patterns

Researchers modeled how executive-function outcomes varied with exercise volume instead of treating all exercise programs as interchangeable. The reported nonlinear pattern means the modeled response did not simply increase in a constant proportion with exercise volume.

Different significance results do not prove superiority

Aerobic and integrated training showed favorable modeled patterns, whereas resistance results were inconsistent or nonsignificant. That contrast alone does not establish a statistically supported difference between exercise types, and nonsignificance does not prove resistance training has no effect.

CHECK THE ORIGINAL

The original publication

Optimizing exercise dosage for executive function in alzheimer's disease: A Bayesian dose-response meta-analysis of randomized trials.

Sui S, Wang M
Arch Gerontol Geriatr · 2025

PubMed ID
40882468
Record checked

AI-assisted research and writing. This explains one selected publication; it is not a complete review of everything known. Our approach.