Best exercise for the ageing brain depends on the goal

Mild cognitive impairment sits between normal ageing and dementia, and people with it face an increased risk of progressing to dementia — with no drug that reliably halts that slide. Exercise is one of the most promising non-pharmacological tools for protecting cognition, but 'exercise' covers everything from jogging to tai chi to lifting weights. Until now it has been unclear whether these modalities are interchangeable for brain health, or whether different types of training benefit different cognitive domains. This review set out to compare modalities head-to-head statistically and to identify what an optimal exercise prescription for MCI might look like.

The authors conducted a systematic review and frequentist random-effects network meta-analysis, following PRISMA guidelines and preregistered on PROSPERO. They identified randomised controlled trials of exercise in people with mild cognitive impairment, covering aerobic exercise, resistance exercise, mind-body exercise, dance, combined aerobic-plus-resistance training, and multicomponent programmes, compared against passive or active controls. Fifty-five trials involving 4,603 participants were included. Standardised mean differences were calculated from change-from-baseline scores, modalities were ranked using SUCRA probabilities, and prespecified subgroup analyses examined prescription characteristics such as intensity, frequency and duration for multicomponent exercise.

For global cognitive function, every exercise modality beat passive control, with multicomponent exercise ranking highest (SUCRA 83.3%). The picture split by domain: for memory-related outcomes, resistance, mind-body, dance and multicomponent exercise were superior to passive control, with mind-body exercise ranking top (SUCRA 83.9%); for executive function, aerobic, resistance, dance, combined and multicomponent exercise all showed significant benefits, with resistance training ranking top (SUCRA 73.7%). Subgroup analyses suggested moderate-intensity multicomponent exercise done at least five times per week was associated with greater global cognitive gains. Interventions lasting 13–24 weeks showed the largest effect estimates, although duration was not a statistically significant moderator.

This synthesis confirms the broad message that exercise helps cognition in MCI, but it complicates the one-size-fits-all view: the best modality appears to depend on which cognitive domain you care about. The domain-specific pattern — mind-body practices for memory, resistance work for executive function, mixed programmes for global cognition — supports individualised exercise selection rather than a single universal prescription. Because SUCRA rankings rest largely on indirect comparisons across heterogeneous trials, they indicate probability of being best rather than proof of superiority. Direct head-to-head trials between modalities remain the key gap.

For anyone thinking about long-term brain health, this is a reason to believe the type of exercise you choose is not irrelevant — and that variety may be the smartest hedge, given multicomponent programmes ranked best for overall cognition. It also strengthens the case that resistance training belongs in the brain-health conversation, not just the muscle one, and that practices like tai chi or yoga are more than relaxation. These findings come from people already diagnosed with MCI, so they describe that population rather than prescribing anything for healthy readers.