Comparative evaluation of methods for quantification of neural activity

Electrograms of single- and few-fibre baro- and chemoreceptor preparations from the carotid sinus nerve of rabbits were recorded monophasically at different levels of activation, as well as complete absence of potentials during temporary cold block of the preparation. The neural activity of every neurogram was determined by a PDP-12 computer using three different methods: (i) Determining all maxima and minima in the electrogram and distinguishing action potentials by comparing the amplitude changes to discriminator values derived from previous analysis of noise; (ii) counting transitions of the electrogram across a constant-voltage threshold just above the baseline noise; (iii) integrating the electrogram with or without rectification with reference to the electrical zero level. The results provided evidence that, in single- as in few-fibre preparations if methods (ii) in part, and (iii) generally, yield erroneous values and, consequently, distorted activity response curves due to measurement of baseline or superimposed noise, loss of summated potentials, baseline shifts and/or the presence of different types of potentials in the recording. Most satisfactory results were obtained by method (i). However, since even the latter becomes incorrect if the neural frequency exceeds several hundred to about 1000 impulses/s, none of these methods can be considered for exact quantification of neural activity in multi-fibre preparations or whole nerves.