Towards selective recording in the peripheral nervous system using cuff electrodes

Neural recording in the peripheral nervous system implies to find a balance between selectivity of the recording and invasiveness of the electrodes used. Cuff electrodes consist of cylindrical structures made of an insulating biocompatible material with metal contacts on the inner face, to be placed around the nerve surface. Among the available categories, they are the least invasive and have shown the best long-term implant stability, as they allow to record electroneurograms without compromising the integrity of the nerve structure. However, the identification of the active fibres inside the nerve is extremely complicated if compared with more invasive alternatives, and requires powerful classification algorithms to interpret the signals, achieving selectivity. Although different designs have been investigated in research, setting the optimal configuration of the contacts is not trivial, and implies a complexity trade-off between the cuff and the processing units. The purpose of this work is to provide a tool for simulating the collection of electroneurograms from the peripheral nerve, thus enabling to test different configurations of cuff electrodes and signal processing algorithms to increase selectivity. The project required to find the best simulation strategy and led to implementing a Multiphysics model from scratch, capable of producing simulated electroneurograms and making them available for processing. The resulting tool allows an integration of the single neurons’ simulations, obtained with specific software, within a Multiphysics model of the nerve coupled with the cuff electrode. The proposed setup allows testing of user-defined electrode designs, exploiting a parametrization of their features. The signals sensed by each contact can be exported and processed, enabling the implementation of algorithms for activity recognition. Performances on the recording of evoked potentials showed a match with experimental results, and the developed strategy leaves room for increasing complexity according to the user’s needs. This will lead, in the long term, to streamline the design of stable and selective peripheral nerve interfaces and to optimise animal tests limiting the use of resources.