Evaluating the Effect of Fibre Parameters on Neural Simulations Computed Using ASCENT

The use of computational models is a key part of modern research. For neural stimulation technique development simulations are particularly important, enabling the initial tri-alling of novel methods without using animal models. However, the fidelity of the simulation parameters is crucial to ensuring that its results are predictive of a real-world implementation. ASCENT, developed by the Grill Lab, is a pipeline combines COMSOL with NEURON to simplify the process of simulating specific nerve and cuff configurations. This paper aims to investigate ASCENT, and to highlight areas of importance for other researchers hoping to use it as part of their own work. In particular, this paper is focussed on the effect of three simulation parameters; modelled fibre count, the ratio of fibres within each fascicle, and random number seed values; on the fascicle-level selectivity for a stimulation pulse. The rationale for this was a) to study the effects of sample/cuff independent variables, and b) to determine the usefulness of ASCENT as a simulation platform for selective stimulation methods, which are of current interest.This study finds that the fibre diameter seed and fibre count both have a significant effect on the calculated values of fascicle selectivity. However, the fibre location seed and fascicle fibre ratio have a negligible effect. Based on this, recommendations are made for optimising neural simulations done using ASCENT.