• computational framework
• computational neuroscience
• finite element modelling
• nerve fiber model
• peripheral neuroprostheses

The development of computational models for the simulation of stimulation and recording from peripheral neuroprostheses has gained attention as a fundamental tool in the design of bidirectional prosthetic applications. We present a hybrid modelling framework in which the solution to the biological volume conduction problem linked to stimulation through charge injection (and, exploiting reciprocity, neural recording) establishes the extracellular driving action for the recruitment of targeted nerve fibers. The volume conduction problem is solved employing finite element modelling and thus requires careful parametric geometrical characterization of the electrode and nerve structures, together with their main electric features. The study of fiber dynamics is carried on with the NEURON simulation framework, individuating appropriate models for motor and sensory myelinated fibers, and unmyelinated fibers. Such models are variations on the classical cable-like Hodgkin-Huxley conductance-based model of nerve fibers and provide reliable results, whose neuroscientific soundness is variably tested in our work. Finally, the hybrid framework is joined to Bensmaia's mechano-neural model to simulate the whole pipeline from mechanical stimulus to neural aggregated response in the form of a microneurogram-like signal. Some preliminary analysis are performed to assess the reliability of this model with a view to providing a complete tool for the development of neural bidirectional prostheses.