Aerobic, Resistance, or Multicomponent: Does the Match Matter?

The exercise evidence now has to answer a comparison question: if cognitive outcomes differ, does the type of exercise matter? This chapter compares aerobic, resistance, and multicomponent programs against the outcomes they were used to measure. The goal is not to crown one universal winner, but to show how a modality can become more or less relevant depending on the cognitive target, the kind of evidence behind the comparison, and whether a person can actually sustain the program.

A ranking changes with the cognitive outcome

The network meta-analysis produced a different order for global cognition and executive function. For global cognition, multicomponent exercise had the highest SUCRA value at 88.5%, followed by aerobic exercise at 58.3% and resistance exercise at 48.5%. For executive function, aerobic exercise ranked first at 90.8%, followed by multicomponent exercise at 50.7% and resistance exercise at 38.1%.

The simplest interpretation is also the most useful: a program that combines several exercise modes may be a promising fit for broad cognitive performance, while aerobic exercise may be especially promising for executive control. The order changes because the outcome changes. Global cognition is a broad composite assembled from multiple abilities; executive function emphasizes control, inhibition, flexibility, and problem solving. It would be surprising if the same intervention order were guaranteed for every possible cognitive task.

The ranking is consistent with the mechanism model, but it does not prove the mechanism. A multicomponent program may offer several kinds of physiological and movement demands that align with a broad composite. Aerobic exercise may support the vascular and prefrontal processes relevant to executive control. These are plausible interpretations of the pattern, not a demonstration that one component caused the result. The comparative finding is strongest when stated as a conditional association: in this network, these modes had different probabilities of being most effective for these outcomes.

This is a more informative answer than “all exercise is equal,” but it is not a reason to turn the order into a prescription. The studies used different participants, durations, intensities, tests, and controls. A rank can summarize a fragmented evidence base while still leaving uncertainty about how the programs would compare if the same participants completed them under the same conditions.

What SUCRA can—and cannot—tell us

SUCRA is short for the surface under the cumulative ranking curve. In plain language, it summarizes how likely an intervention is to occupy a favorable position within the network’s estimated ordering. A higher value indicates a higher probability of being among the more effective options for that outcome in the analyzed comparison. It is a ranking statistic, not a direct measurement of how much a participant improved.

That distinction prevents three common mistakes. First, a higher SUCRA value does not necessarily mean a large or clinically important difference from the next option. Two interventions can have different rank probabilities even when their estimated effects overlap substantially. Second, a lower rank does not mean that the intervention had no benefit against control.