Axonal projections of single bulbospinal inspiratory neurons revealed by spike-triggered averaging and antidromic activation

Activity was recorded extracellularly from 26 inspiratory bulbospinal neurons in anesthetized, paralyzed, artificially ventilated cats. All but one were located in the ventral respiratory group. A neuron was classified as either I alpha or I beta by comparing its firing pattern during inspiratory cycles with lung inflation to its pattern when lung inflation was withheld during the central inspiratory phase (2, 14, 15). In this study, the projection and conduction velocity of these axons were determined using two methods: antidromic activation (AA) of the bulbospinal neurons and spike-triggered averaging (STA) of the extracellular field potentials. These methods have been compared directly because the same electrode was used both for stimulating the axon of the bulbospinal neuron and recording its axonal potential in the same location. Axonal projections from these neurons were mapped in the contralateral spinal cord with a mobile electrode by determining where the lowest stimulus threshold occurs for AA and greatest axonal potential can be recorded with STA. The locations of these axons were in the ventral and lateral funiculi. Each method determined a similar location for an axon. Positions of 10 axons were determined at both the third (C3) and fourth (C4) cervical segments. Single axons maintained their positions in either descending tract from rostral C3 to mid-C4. In five of six cases where two "neighboring" medullary units were characterized, the axons of each pair projected together within 350 micron of each other in the cervical spinal cord. Estimates of mean axonal conduction velocity (CV) from antidromic activation from a single stimulus site, "single-point AA," were as much as 42% less than corresponding estimates from STA extracellular field potentials at that point (P less than 0.001). Such single-point estimates were less than determinations that were calculated from the difference in conduction time and the difference in conduction distance from two points in the spinal cord. These two-point determinations averaged 55.4 +/- 13.1 m/s (using AA) and 53.3 +/- 13.1 (using STA) for 10 neurons. These values were not significantly (P greater than 0.2) different from each other and are greater than most earlier reports, which used the single-point AA method. Either method, AA or STA, can be used to determine axonal position and CV. The advantages and disadvantages of each method are discussed.(ABSTRACT TRUNCATED AT 400 WORDS)