We developed a new non-invasive technique; one-point sampling method, for quantitative measurement of regional cerebral blood flow (rCBF) with N-isopropyl-p-[123I]iodoamphetamine and SPECT. Although the continuous withdrawal of arterial blood and octanol treatment of the blood are required in the conventional microsphere method, the new technique does not require these two procedures. The total activity of 123I-IMP obtained by the continuous withdrawal of arterial blood is inferred by the activity of 123I-IMP obtained by the one point arterial sample using a regression line. To determine when one point sampling time was optimum for inferring integral input function of the continuous withdrawal and whether the treatment of sampled blood for octanol fraction was required, we examined a correlation between the total activity of arterial blood withdrawn from 0 to 5 min after the injection and the activity of one point sample obtained at time t, and calculated a regression line. As the results, the minimum % error for the inference using the regression line was obtained at 6 min after the 123I-IMP injection, moreover, the octanol treatment was not required. Then examining an effect on the values of rCBF when the sampling time was deviated from 6 min, we could correct the values in approximately 3% error when the sample was obtained at 6 +/- 1 min after the injection. The one-point sampling method provides accurate and relatively non-invasive measurement of rCBF without octanol extraction of arterial blood.