Numerical Analysis of Electric Field Generated by a Nerve Cuff Electrode.

A 3-D finite element model of an unifascicular nerve was developed to study the potential distribution from a nerve cuff electrode [1]. A schematic model nerve is shown in the figure. The z-axis runs in the direction of the axons in the nerve trunk. The 3-D domain representing the nerve, surrounding tissue and the electrode were assumed purely resistive. The nerve model consisted of 28 layers with 256 volume elements along the axis of the nerve. A finer mesh was used in the region inside and close to the electrode.
The x-y and the x-z planes were considered to be two planes of symmetry. Therefore, only one-fourth of the problem was modeled using 7769 nodes and 1680 volume elements.
The electrical potential distribution was obtained by solving the Laplace Equation within a 3-D domain. The solution was obtained by minimizing a functional given by the equation at the bottom of the Figure.

Geometry of the Nerve

The geometry of the represented nerve is shown in the figure. It was modeled as a three-mm. diameter, unifascicular anisotropic structure with
s-longitudinal = 1.0 S/m and s-transverse = 0.1 S/m, surrounded by a 30 µm perineurium with sigma = 0.00063 S/m.
Two layers of encapsulation tissue, a 150 µm loose layer with s = 1.0 S/m and a 50 µm tight layer with s = 0.0659 S/m (so that the effective sigma = 0.22/m) and a 50 µm layer of 0.9% saline (sigma =2.0 S/m) were considered to between the nerve and the electrode.
The electrode contacts were 30 µm thick with s = 10^4 S/m and with surface area of 1.57 sq. mm. They were embedded on the inner surface of a 1 mm thick insulated nerve cuff (sigma = 10^-6 S/m). The surrounding body fluid was represented by a 20 mm layer of saline.