Building a better virtual nerve

Vagus nerve stimulation is an approved treatment for many conditions, including epilepsy, depression, and stroke rehabilitation. In this treatment, implantable electrodes are placed on the vagus nerve in the neck and deliver electricity to activate vagal neurons. Accurate computational models of the nerve structure enable researchers to design effective, new therapies with fewer side effects.

Often, these simulations are simplified by extruding the anatomy from a 2D cross section of the nerve to create a 3D model. Marshall et al. instead simulated a “true-3D” vagus nerve by using 3D imaging of fascicles — bundles of fibers — within the nerve, capturing longitudinal changes in structure and their effects on electrical stimulation.

“Along the length of the nerve, fascicles are merging and splitting,” said author Daniel Marshall. “This is occurring at a rate of a merge or split about every half millimeter. An extrusion model from a single cross section is not capturing any of that, and the modeled electrical potentials that affect neural activity may be misinformed.”

The authors incorporated this fascicle splitting and merging into their model; they also reshaped the nerve to conform to the shape of the electrode. They then compared the electrical stimulation responses of the two approaches — extrusion vs. true-3D models — across multiple metrics, including dose-response relationships, fiber-specific thresholds, and spatial selectivity.

They found that the extrusion models agree with the true-3D models, but only under specific conditions.

“Surprisingly, we found that if you use the cross section under the cathodic — that is, the negative stimulation — contact to build the extrusion model, and you model the nerve reshaping, they match quite well,” said Marshall.

The authors hope to guide researchers studying electrical nerve stimulation using models that are both accurate and efficient. They believe their true-3D simulation pipeline and suite of metrics they used to compare models will guide researchers looking to design new peripheral nerve stimulation therapies.

“We’re also working on new modeling approaches that allow us to design new vagus nerve stimulation therapies much faster, using new computational tools and new anatomical imaging data,” said author Nicole Pelot. “We want to design therapies that activate neurons to evoke therapeutic responses and avoid activation of neurons that cause side effects.”

Source: “Computational modeling of human vagus nerve stimulation with three-dimensional fascicular morphology,” by Daniel P. Marshall, Aniruddha R. Upadhye, Ozge N. Buyukcelik, Andrew J. Shoffstall, Warren M. Grill, and Nicole A. Pelot, APL Bioengineering (2026). The article can be accessed at https://doi.org/10.1063/5.0308450 .