Methods: Seventy-one (50 males) healthy controls (age 21 to 81) had one or two positron emission tomography (PET) scans while inhaling [15O]O2. Cerebral oxygen consumption (CMRO2) was calculated voxel-wise and mean values from cortical grey matter were extracted, using individual masks from magnetic resonance images. Subjects were controls in seven previous studies and a mixed-effects REML regression were performed to account for differences between studies. Results: Mean CMRO2 was 172 mmol O2/100g/minute (95% confidence interval (CI95) 165, 179) at age 50 and the slope of the regression was -0.14 mmol O2/100g/minute per year (CI95 -0.49, 0.21) with p=0.442. Conclusion: The data shows CMRO2 to be constant per unit of surviving cortical tissue during the age span investigated. Earlier studies have shown that atrophy in older subjects affects PET measurements. The present method corrects for atrophy by using grey matter masks from individual subjects, so although atrophy is present in older subjects, the masks used to extract mean values were smaller in older subjects. Of course, the total oxidative metabolism is reduced in proportion to the atrophy but no decline is seen in remaining tissue. The study of oxygen consumption in healthy brain aging is particularly important as subjects with neurodegenerative diseases show decreases in the same parameter.