Multi-electrode nerve cuff recording-model analysis of the effects of finite cuff length

The effect of finite cuff length on the signals recorded by electrodes at different positions along the nerve was analysed in a model study. Relations were derived using a one-dimensional model. These were evaluated in a more realistic axially symmetric 3D model. This evaluation indicated that the cuff appeared shorter because of edge effects at the beginning and end of the cuff. The method for velocity selective filtering introduced by Donaldson was subsequently analysed. In this method, velocity selective filtering is achieved by summing the signals of subsequent tripoles after applying time shifts tuned to a certain conduction velocity. It was also found that the optimum electrode distance for a given cuff length for maximum summed RMS of symmetrical tripoles in the cuff is larger than when evaluating peak-peak amplitudes of single fibre action potentials. Velocity selective filtering yields better selectivity when using symmetrical tripoles, but may yield larger signal RMS when using the wider asymmetrical tripoles, potentially allowing for shorter cuffs. It is speculated that application of a multi-electrode reference may improve velocity selectivity for asymmetrical tripoles.