The aim of this study was to assess regional cerebral blood flow (rCBF) in normal subjects at rest using 99Tcm-HMPAO single photon emission tomography (SPET). Analysis of reproducibility and repeatability was performed both before and after normalization of flow data. Six healthy volunteers were examined, three times each, according to a routine rCBF protocol. A computerized brain atlas was used to evaluate flow data in eight selected regions. The overall reproducibility of rCBF was evaluated from two scans performed at an average interval of 3 months. Repeatability was evaluated from two scans, 3 h apart and without re-injection of 99Tcm-HMPAO. For the normalized (relative) flow data, the reproducibility was +/- 1.3% and the repeatability +/- 2.2% (i.e. methodological errors dominate). For the non-normalized flow data, the corresponding values were +/- 14.8% and +/- 5.9%. rCBF SPET with 99Tcm-HMPAO is highly reproducible provided that the flow data are normalized. The variation in flow between individuals at one point in time and 3 months later was less than +/- 5% for all brain regions.