Physiological activation increases glucose uptake locally in the brain. However, it is not known how high intensity exercise affects regional and global brain glucose uptake. The effect of exercise intensity and exercise capacity on brain glucose uptake was directly measured using positron emission tomography (PET) and [18F]fluoro-deoxy-glucose ([18F]FDG). Fourteen healthy, right-handed men were studied after 35 min of bicycle exercise at exercise intensities corresponding to 30, 55 and 75% of on three separate days. [18F]FDG was injected 10 min after the start of the exercise. Thereafter exercise was continued for another 25 min. PET scanning of the brain was conducted after completion of the exercise. Regional glucose metabolic rate (rGMR) decreased in all measured cortical regions as exercise intensity increased. The mean decrease between the highest and lowest exercise intensity was 32% globally in the brain (38.6+/-4.6 versus 26.1+/-5.0 micromol (100 g)-1 min-1, P<0.001). Lactate availability during exercise tended to correlate negatively with the observed brain glucose uptake. In addition, the decrease in glucose uptake in the dorsal part of the anterior cingulate cortex (37% versus 20%, P<0.05 between 30% and 75% of VO2max) was significantly more pronounced in subjects with higher exercise capacity. These results demonstrate that brain glucose uptake decreases with increase in exercise intensity. Therefore substrates other than glucose, most likely lactate, are utilized by the brain in order to compensate the increased energy needed to maintain neuronal activity during high intensity exercise. Moreover, it seems that exercise training could be related to adaptive metabolic changes locally in the frontal cortical regions.