Identification of multifiber neural spike trains by interval prediction and morphology

Interpretation of neurograms from multifiber preparations is usually limited to forming a grand running average. In so doing, individual nerve activity is lost. This information is important in studying nervous system activity. We have developed a method to find the rates of constituent nerve fibers using linear one-step prediction and cross correlation between the spike template and instantaneous spikes. Some size discrimination is employed, as is a memory to mark spikes which are "accounted for." Thus we are able to parse out the several rates which make up an aggregate spike train. Our results show that some afferent baroreceptor fibers may increase rate, while others may decrease rate, in response to holding carotid sinus pressure at a high level for 10 minutes.