Acute cardiovascular physical exercise improves cognitive performance, as evidenced by a reduction in reaction time (RT). However, the mechanistic understanding of how this occurs is elusive and has not been rigorously investigated in humans. Here, using positron emission tomography (PET) with [ C]raclopride, in a multi-experiment study we investigated whether acute exercise releases endogenous dopamine (DA) in the brain. We hypothesized that acute exercise augments the brain DA system, and that RT improvement is correlated with this endogenous DA release. The PET study (Experiment 1: n = 16) demonstrated that acute physical exercise released endogenous DA, and that endogenous DA release was correlated with improvements in RT of the Go/No-Go task. Thereafter, using two electrical muscle stimulation (EMS) studies (Experiments 2 and 3: n = 18 and 22 respectively), we investigated what triggers RT improvement. The EMS studies indicated that EMS with moderate arm cranking improved RT, but RT was not improved following EMS alone or EMS combined with no load arm cranking. The novel mechanistic findings from these experiments are: (1) endogenous DA appears to be an important neuromodulator for RT improvement and (2) RT is only altered when exercise is associated with central signals from higher brain centres. Our findings explain how humans rapidly alter their behaviour using neuromodulatory systems and have significant implications for promotion of cognitive health. KEY POINTS: Acute cardiovascular exercise improves cognitive performance, as evidenced by a reduction in reaction time (RT). However, the mechanistic understanding of how this occurs is elusive and has not been rigorously investigated in humans. Using the neurochemical specificity of [ C]raclopride positron emission tomography, we demonstrated that acute supine cycling released endogenous dopamine (DA), and that this release was correlated with improved RT. Additional electrical muscle stimulation studies demonstrated that peripherally driven muscle contractions (i.e. exercise) were insufficient to improve RT. The current study suggests that endogenous DA is an important neuromodulator for RT improvement, and that RT is only altered when exercise is associated with central signals from higher brain centres.