(55) PyFibers: A Python Package for Simulating Model Nerve Fiber Responses to Electrical Stimulation Using NEURON

Introduction: Accurate computational models of the responses of peripheral nerve fibers to applied electrical stimulation and block signals are essential for designing and optimizing neuromodulation therapies. Existing approaches commonly use the freeware NEURON [1], [2] and rely on legacy HOC code, which has a steep learning curve. Further, fiber modeling code is frequently developed anew for each publication, resulting in unnecessary duplication of effort and elevated risk of errors [3]. We addressed these challenges by creating PyFibers, an open-source Python package for building, simulating, and analyzing nerve fiber responses to electrical stimulation in NEURON.

Methods: PyFibers features an object-oriented design with two main modules. First, fiber.py instantiates and connects NEURON sections to form user-defined fibers with distinct geometries and ion channels. Users can select from a library of 13 published fiber models (e.g., MRG [4], Thio [5], Tigerholm [6]) or incorporate custom models. Second, stimulation.py provides tools for delivering extracellular or intracellular stimulation, and for running a configurable bisection search to determine activation or block threshold amplitudes. Recording functions capture transmembrane voltage, transmembrane currents, and gating variables, enabling post-simulation analysis. Additional features include interpolation of extracellular potentials sampled from finite element models, 3D trajectories to represent more realistic nerve pathways, modeling single-fiber action potential recordings, and multi-source stimulation.

Results: PyFibers dramatically simplified the process of implementing and running fiber simulations, reduced the necessary lines of code by more than an order of magnitude. Published models of myelinated and unmyelinated fibers operate under a unified, object-oriented framework, facilitating direct comparisons and minimizing duplication of effort. Validation against existing HOC implementations confirmed that PyFibers reproduces key neural responses, including recovery cycle, conduction velocities, action potential shapes, and activation thresholds. PyFibers simulations also reduced compute time by >50% for threshold searches. We integrated PyFibers into the ASCENT pipeline for modeling peripheral nerve stimulation [7], eliminating almost 8,000 lines of legacy HOC code. Finally, the modular software design allows researchers to create custom fiber models and novel stimulation paradigms with substantially less effort.

Conclusion: PyFibers modernizes peripheral nerve fiber modeling by consolidating diverse fiber models and stimulation paradigms into a single user-friendly Python package. Its modular design, open-source availability, validation, and documentation make it a powerful resource for researchers seeking to refine neuromodulation strategies. By reducing the technical barriers to computational modeling, PyFibers will foster community-driven innovation and facilitate transparent sharing of both code and data, ultimately advancing the field of neural interfaces and improving therapeutic outcomes.