Does the exercise-cognition dose-response differ between older adults with cognitive impairment and those who are cognitively normal?

The dose-response curve bends, and the bend sits at different doses in different groups

In older adults with mild cognitive impairment or dementia, more exercise helps cognition only up to a point, and the curve is not a straight line. A network meta-analysis of 68 trials covering 6,544 participants estimated that about 336 Metabolic Equivalent of Task minutes per week — a MET-minute being a way of counting exercise volume, so that brisk walking for a minute costs more than sitting for a minute — was the minimum needed for any cognitive benefit. Roughly 934 MET-minutes per week was the minimum for an improvement large enough to be clinically meaningful [1].

A separate network meta-analysis of 44 trials in adults aged 50 and over, with 4,793 participants, also found a nonlinear association between overall exercise and cognition, but its numbers were different: it found no minimum threshold at all, estimated 724 MET-minutes per week for clinically relevant change, and reported that benefits became less clear above about 1,200 MET-minutes per week [3]. That analysis did not specify whether its participants had cognitive impairment or normal cognition, which is exactly the gap that matters here.

Why the comparison is hard to make: the trials were built around impaired groups

The dose-response studies in this set were designed around people who already have cognitive problems. One focuses on mild cognitive impairment or dementia and explicitly notes that the dose-response varied substantially between those two groups, with some exercise types producing significant effects at lower doses than others [1]. Another examines mind-body exercise specifically in older adults with mild cognitive impairment [2]. A third looks at exercise effects on BDNF — a protein involved in brain plasticity — and depressive symptoms in middle-aged and older adults with mild cognitive impairment [4]. None of these includes a cognitively normal comparison group, so none can tell us whether the curve would shift for someone without impairment.

The one review that does include both groups looked at rhythmic physical activity and global cognition in older adults with and without mild cognitive impairment. It found beneficial effects in five of eleven trials, but the protocols varied so widely in frequency, duration and type of activity that the authors attributed the mixed results to that heterogeneity, and they did not analyze dose-response at all [5]. So even where normal and impaired adults appear in the same review, the dose question was not asked.

What could make the curve differ between the two groups

The impaired-group evidence hints at where a difference might come from. Within mild cognitive impairment and dementia, the dose-response relationship was not uniform: the same exercise volume did not produce the same cognitive payoff across severity levels, and the best-performing modality was aerobic exercise combined with resistance training, which showed superior efficacy even at modest activity levels [1]. If severity alone shifts the curve inside the impaired range, it is plausible that moving from impaired to normal cognition would shift it further — but that is an inference from the pattern, not something any of these studies tested.

Exercise type also changes the dose picture independently of who is exercising. The broader older-adult analysis found that the dose-response depended on exercise modality, that clinically important effects may occur at lower doses for many types of exercise, and that resistance exercise outperformed other modalities [3]. The impaired-group analysis similarly found modality-specific thresholds, with some types effective at lower doses [1]. Because modality and dose interact, any comparison between impaired and normal groups would need to hold exercise type constant — something none of the available studies did across both groups.

What this means if you are trying to apply a number

For someone with mild cognitive impairment or dementia, the most defensible reading of this evidence is that a combined aerobic-plus-resistance program is the strongest choice, that benefits can appear at fairly modest activity levels, and that the dose needed for a clinically meaningful change is higher than the dose needed for any change at all [1]. The exact MET-minute figures should be treated as estimates from pooled trials with substantial variability, not as a personal prescription.

For a cognitively normal older adult, the closest available anchor comes from the mixed older-adult analysis: a nonlinear curve, a clinically relevant dose around 724 MET-minutes per week, and diminishing clarity above roughly 1,200 [3]. Whether that curve is the same one that applies to people with impairment, or a shifted version of it, remains untested in this evidence — the direct comparison has not been made.

Sources used in this answer

Related research questions