Impedance changes were measured during the compound action potential (CAP) in isolated crab nerves with 50 kHz or direct applied current (DC), to suggest an optimal frequency for the imaging of neuronal depolarisation by electrical impedance tomography (EIT). With DC, a predominant apparent impedance decrease of 0.2-0.7 per cent of the resting baseline was observed, which had a time course similar to that of the action potential. Control recordings suggested that these changes could not be attributed to technical errors resulting from the measurement method. A component in some measurements was attributable to effects of the measuring current on the latency of the CAP, but the mechanism underlying the remainder of the changes was unclear. No changes greater than +/- 0.005 per cent were seen with measurement at 50 kHz. This suggests that research into noninvasive rapid imaging of nervous activity in the brain with EIT would be more profitably performed with measurements at frequencies lower than the 50 kHz currently used for safety reasons.