Cerebral metabolic rate of oxygen (CMRO(2)) can be assessed quantitatively using (15)O(2) and positron emission tomography. Determining the arterial input function is considered critical with regards to the separation of the metabolic product of (15)O(2) (RW) from a measured whole blood. A mathematical formula based on physiologic model has been proposed to predict RW. This study was intended to verify the adequacy of that model and a simplified procedure applying that model for wide range of species and physiologic conditions. The formula consists of four parameters, including of a production rate of RW (k) corresponding to the total body oxidative metabolism (BMRO(2)). Experiments were performed on 6 monkeys, 3 pigs, 12 rats, and 231 clinical patients, among which the monkeys were studied at varied physiologic conditions. The formula reproduced the observed RW. Greater k values were observed in smaller animals, whereas other parameters did not differ amongst species. The simulation showed CMRO(2) sensitive only to k, but not to others, suggesting that validity of determination of only k from a single blood sample. Also, k was correlated with BMRO(2), suggesting that k can be determined from BMRO(2). The present model and simplified procedure can be used to assess CMRO(2) for a wide range of conditions and species.