The biophysical basis of accommodation in human peripheral nerve fibres

When a hyperpolarizing or sub-threshold depolarizing current is applied to a nerve, the electrotonic changes in membrane potential induced are reflected in the changes in stimulus current required to excite the axons. These threshold changes, known as threshold electrotonus, have a stereotyped form in normal subjects, but have been found to show distinctive forms in patients with certain neurological pathologies. The aim of this study was to investigate the use of threshold electrotonus to determine the nature of the underlying differences between normal and abnormal nerve which give rise to these differences in form. Threshold electrotonus was first studied in rat nerves in vitro. The stereotyped normal form was confirmed, as was its conversion by potassium channel block into a highly abnormal form ('ALS Type II') previously found in some patients with amyotrophic lateral sclerosis. A technique was developed to compare electrotonus and threshold electrotonus in the same nerves, and no evidence of the previously hypothesised membrane bistability in ALS Type II nerves found. A new method of recording threshold electrotonus, using two pulse durations, was proposed to take account of differences in strength-duration properties. When applied to normal volunteers, this method demonstrated previously unseen differences in accommodation between human motor and sensory axons. A new computer simulation of threshold electrotonus was developed in C++, to test the ability of a simplified equivalent circuit of myelinated axons to account for normal and abnormal recordings, and to test the hypothesis that fitting techniques might reveal the underlying differences in membrane properties. It was first shown that the model could satisfactorily reproduce threshold electrotonus waveforms, and that a fitting procedure could identify changes in the parametes of the equivalent circuit producing test waveforms. This fitting procedure was then applied to 2 series of new threshold electrotonus recordings recorded with two pulse durations: the changes in normal motor axons caused by ischaemia and release of ischaemia, and the differences between motor and sensory axons. The Powell fitting procedure correctly identified the dominant role of membrane depolarization in ischaemia and hyperpolarization following ischaemia, but could not satisfactorily account for the motor-sensory differences, although they were more pronounced with the new method. The same or similar forms of threshold electrotonus could be produced from quite different selections for the nerve properties in the simulation. It was concluded that threshold electrotonus measurements on their own, even using two pulse widths, contain too little information to unambiguously determine changes in nerve membrane properties. The nerve testing protocol should therefore be extended to gather additional independent data (e.g. the recovery cycle following an action potential). The fitting procedure could be improved by using Bayesian methods to incorporate more a priori information, and by the use of techniques to avoid local minima