Mice will run for miles (literally!) if there’s a treadmill or wheel in their cage. I could certainly imagine there’s some sort of multiplicative effect, where a substances causes (say) a 1.2x increase in activity, but that’s also not really how caffeine hits me: too much and I’m nonspecifically jittery.

Regardless, that could certainly explain the caffeine-only part, but it’s harder to square with the caffeine*anti-depressant effect.