According to classical cable theory, a magnetic coil (MC) should excite a linear nerve fibre in a homogeneous medium at the negative-going first spatial derivative of the induced electric field. This prediction was tested by MC stimulation of mammalian phrenic and amphibian sciatic nerve and branches in vitro, immersed in Ringer solution within a trough, and identifying the sites of excitation by recording responses of similar latency to local electrical stimulation. Subsequently, the identified sites of excitation were compared with measurements of the induced electric field and its calculated first spatial derivative. A special hardware device was used to selectively reverse MC current direction and to generate predominantly monophasic- or polyphasic-induced pulse profiles whose initial phases were identical in polarity, shape and amplitude. When using the amphibian nerve preparation, a complication was excitation at low threshold points related to cut branches. 2. Reversal of monophasic current resulted in latency shifts corresponding approximately to the distance between induced cathode and anode. The location of each site of excitation was at, or very near, the negative-going first spatial derivative peaks of the induced electric field measured parallel to the straight nerve. Significantly, excitation of the nerve did not occur at the peak of the induced electric field above the centre of the 'figure of eight' MC junction. 3. A polyphasic pulse excited the nerve at both sites, by the negative-going first phase at one location, and approximately 150 microseconds later, by the reversed negative-going second phase at the other location. Polyphasic and monophasic pulses elicited responses with similar latency when the induced current flowed towards the recording electrode. 4. Straddling a nerve with non-coding solid lucite cylinders created a localized spatial narrowing and increase in the induced electric field, resulting in a lowered threshold of excitation. The corresponding closer spacing between first spatial derivative peaks was exhibited by a significant reduction in latency shift when MC current direction was reversed. 5. When a nerve is bent and the induced current is directed along the nerve towards the bend, the threshold of excitation is reduced there. Increasing the angle of the bend from 0 deg to more than 90 deg graded the decrease in threshold. 6. In a straight nerve the threshold was lowest when current was directed towards the cut end.(ABSTRACT TRUNCATED AT 400 WORDS)
