The original objective of this work was to verify the possibility of using electrical pulsatile cerebral impedance measurements as a diagnostic aid for assessing the brain-death condition in adults; a subordinate target was to validate a simple method for detecting perfusional changes in the brain. To this end, impedance signals were recorded, for a comparative study, from both live subjects and brain-dead patients, using a simple four-electrode arrangement. Rather unexpectedly, pulsatile transcephalic impedance waveforms exhibiting a temporal dependance similar to those of live subjects were detected in artificially ventilated, cerebrally dead, adult subjects; distributions of the time delays between impedance peaks and ECG peaks were also recorded for the two groups (dead and live subjects). These data provided no evidence, at the 1% significance level, against the hypothesis that the two sample groups are drawn from identical populations. The detection of impedance variations from brain-dead patients can be explained by the residual persistence of blood flow through the scalp, by mechanical variations synchronous with the heart beat and by the presence of the oscillating flow and the systolic spikes that precede the final blood flow arrest. The fact that impedance variations can be traced back to a multiplicity of causes, unrelated to the normal unidirectional flow, renders the transcephalic impedance method inappropriate for detecting cerebral perfusion changes in adults. This conclusion is also strengthened by some theoretical results recently derived from a multilayer model of the head.