Interaction among unitary spike trains: Implications for whole nerve measurements

Interaction among action potential trains was examined in electroneurograms (ENGs) from filaments containing several fibers. In an in vitro preparation, increasing aortic pressure generated trains of action potentials (spikes) from regularly discharging aortic baroreceptors. Multifiber records were electronically sorted, and spikes from each separate baroreceptor were identified. 'True' discharge rates, measured as the instantaneous frequency of the sorted records, were compared with the results of three quantitation approaches commonly used for multifiber or whole nerve activity in mass ENGs: 'spike' counts per unit time, ENG voltage integration, and normalizations of counts or integration. Even in ENGs containing only two active baroreceptors, numerous interactions of the two spike trains occurred. Of 33 paired recordings (66 baroreceptors), 77.3% of the individual baroreceptors had spike trains that could not be clearly discriminated because of interactions with signals from other baroreceptors. Virtually all baroreceptors (96.3%) in nine three-fiber ENGs (27 baroreceptors) showed evidence of missed spikes. Corrected or true spike rates were always greater than activity rates measured by simple counting. Similarity of discharge frequencies in the different baroreceptors was the most important determinant of the rate of voltage signal interaction. Simple count or integration methods underestimated slopes of pressure-response curves and peak activity values by 20-40% compared with sorted instantaneous frequency. Not all coincidence of action potentials in the ENG resulted in simple additions of peak voltages. We conclude that simple counting and integrative techniques substantially underestimate multifiber nerve activity and such quantitation techniques should be used cautiously, particularly in experimental comparisons across groups of animals.