Cross-correlation-based evaluations of the impulse transmission in a nerve: simulation and experimental studies

Bidirectional impulse transmission can be evaluated by cross-correlation analysis when two recording points are simultaneously available on a nerve. This method was tested here on digitized experimental recordings--from the cerebro-buccal connective of Aplysia--and on computer-simulated recordings with predetermined signal and noise content. The data were processed as such, or after being subjected to one of two preliminary treatments aiming at improving the sensitivity and discrimination power of the method. In the first treatment--Positive Value Saving, PVS--digitized values that were larger than the mean level were left unmodified, while the others were replaced by the mean value itself; in the second treatment--Positive Peak Saving, PPS--the values left unmodified were those which were larger than the mean level and represented a relative maximum. PVS tended to eliminate the negative deflections of the extracellular spikes; PPS tended to transform each spike into a single value equal to the spike amplitude. The cross-correlation histograms obtained yielded a clear separation of the impulses travelling in one and the opposite direction of propagation, and provided their subdivision and quantitative estimation according to propagation velocity. In the conditions adopted, spikes comparable in size to the noise range could be revealed. PVS improved sensitivity and discrimination power; PPS provided a very sharp discrimination between impulses with similar propagation velocity, at the expense of a loss of sensitivity.