The use of rheoencephalography (REG) in the clinical practice to evaluate cerebral blood flow is conditional on the finding of a method for removing the extracranial interference caused by the scalp blood flow. To remove this undesirable influence, digital processing based on statistics could be an effective technique if the appropriate data model were applied. This paper focuses on the analysis of the spatiotemporal features of the extracranial REG component, by comparing its morphology and phase shift at several scalp sites. For this purpose, a numerical model of the scalp was employed to assess tissue impedance changes caused by the inflow of a stepwise blood pulse wave. These results were compared with the experimental impedance waveforms recorded on six pairs of adjacent electrodes. The correlation coefficients between each pair of impedance recordings of each subject were always greater than 0.942, showing a mean value of 0.986. This result suggests that the extracranial REG component can be considered as morphologically invariant. On the other hand, negligible phase shifts were observed when mean electrode distances, measured in the blood flow direction, were relatively small, although temporal corrections in the data model would be advisable for longer distances.