Separation between sensory and motor fibers in the human cervical vagus nerve
DOI: Pending DOI
Synopsis
Current Vagus Nerve Stimulation (VNS) is non-selective, causing side effects that can limit its therapeutic efficacy. To improve neuromodulation precision, this study mapped the organization of different vagal fascicle types in 20 human vagus nerves (VN) on the macro and micro anatomical scale. Using micro-CT (µCT) and immunohistochemistry (IHC) staining, the spatial organization and separation of vagal fibers derived from the nodose ganglion and the fascicles that bypass it was uncovered. Unveiling the distinct anatomical arrangement of the cervical could be the key guiding future neuromodulation of the vagus nerve to reduce off-target effects, enhancing VNS safety and efficacy through improved selective stimulation.
Background
Currently used VNS devices deliver stimulation via cuff electrodes that activate vagal fibers in a relatively non-selective manner, which sometimes results in off-target effects that hinder therapeutic efficacy. Side-effects from activation of vagal fibers innervating the larynx and pharynx limit the maximal stimulus intensity and may be a contributing factor to a lack of response to VNS [1,2]. Controlled activation of sensory versus motor functions of the vagus nerve (VN) by function-selective stimulation could tailor therapies to brain or peripheral organ diseases and improve VNS efficacy and safety. VN anatomy and organization of vagal fibers have been extensively studied in rodent [3] and swine models [4]; however, the microscopic organization of afferent and efferent fibers in the human VN remains largely unexplored. Spatially selective VNS targeting specific nerve fascicles were shown to elicit function-selective responses in swine [5,6] and could in principle represent a viable strategy for selective VNS in humans. However, it is unknown whether afferent and efferent fiber trajectories occupy separate areas within the vagal trunk of humans. To address this gap, we mapped the lateral and longitudinal anatomical organization of afferent and efferent fibers inside the trunk of 20 left and right human VNs.
Methods
Twenty vagus nerves were extracted from twelve cadavers, spanning the jugular foramen (JF) to the laryngeal prominence (LP) with sutures placed at JF, LP, and along the nerve's ventral side. Nerves were scanned using micro-CT (µCT) and a 3D reconstruction of the fascicles within the nerve was generated with a 3D U-Net model. Topological graphs encoding fascicular structure and geometric features were extracted and visualized in 3D Slicer. Fascicle trajectories were tracked using these graphs in conjunction with IHC images at the nodose ganglion. Fascicles emerging from the nodose were colored red (N), and bypassing fascicles were colored green (xN). Percent contributions of both fascicle types were generated with an area-based, graded mixing model that predicted fascicle composition in downstream segments of nerves. In ten nerves, three levels were used for quantitative IHC analysis to estimate and validate fiber compositions of N, xN and mixed fascicles: below the Nodose Ganglion, at the upper cervical, and near the LP (middle cervical). IHC staining and high-resolution imaging was performed for neurofilament (NF), myelin basic protein (MBP), choline acetyltransferase (ChAT), and tyrosine hydroxylase (TH). Automated segmentation quantified fibers into four categories: myelinated motor (MM; NF+, MBP+, ChAT+), myelinated sensory (MS; NF+, MBP+, ChAT-), unmyelinated motor (UM; NF+, MBP-, TH+), and unmyelinated sensory (US; NF+, MBP-, TH-).
Results
Even though N and xN fascicles remain separate close to the NG, majority of the fascicles are mixed by the LP. In some cases, fascicles remain mostly N or xN as they travel distally through the LP. In 84 N and 29 xN fascicles from 10 nerves, there is a clear difference in the percent area of MM and MS fibers: N fascicles have a mean of 0.3 ± 0.1% MM and 89.2 ± 0.5% MS fibers, whereas xN fascicles have 17.3 ± 2.4% MM and 72.8 ± 2.4% MS fibers. Unmyelinated fibers occupy similar areas in N and xN fascicles. A logistic regression (R2 = 0.8834, RMSE = 3.4195%) was fit to the percent area of MM versus MS area in below nodose fascicles, and mean values for each category were plotted on the line to define µCT fiber estimates. The Minimum Description Length (MDL) principle was applied to identify thresholds for percent area composition of the below nodose fascicles and divide the regression line into three categories: maximum sensory, maximum motor, and mixed. Tracking MM and MS fiber compositions in downstream fascicles, we found that, in 14 of the 20 nerves, over 50% of the total fascicle area was mixed by the level of the LP. At the same time, at 2 cm centimeters above the LP, in only 4 of the 20 nerves total fascicle area exceeded 50% mixed. To quantify the average cross-sectional separation between MS and MM fibers at different levels, we aligned anatomically (medial-lateral, rostral-ventral) all fascicles from 18 nerves at those levels and calculated average MS and MM areas across nerve area. We found that MS and MM fibers occupy distinct quadrants in both L and R nerves. The sensory and motor “hotspots” are clearest at 3 centimeters above the LP, and the MS-MM asymmetry is maintained through the LP.
Discussion
Our findings indicate that in the human vagus nerve there is spatial separation between maximum sensory and motor fibers in both the longitudinal and lateral directions above the LP. Placement of VNS devices two centimeters above the level of the LP would in principle permit spatially selective activation of sensory vs. motor functions of the VN. Based on anatomical organization of vagal fibers, steering stimulation with selective electrodes towards the dorsal-medial region of the left or right cervical VN would in principle produce sensory over motor responses.
Disclosures
Work funded by NIH-SPARC Reconstructing Vagal Anatomy Project; contract #75N98022C00019
References
1. González, H. F. J., Yengo-Kahn, A., & Englot, D. J. (2019). Vagus Nerve Stimulation for the Treatment of Epilepsy. Neurosurgery clinics of North America, 30(2), 219–230.
2. Thompson, N., Mastitskaya, S., & Holder, D. (2019). Avoiding off-target effects in electrical stimulation of the cervical vagus nerve: Neuroanatomical tracing techniques to study fascicular anatomy of the vagus nerve. Journal of neuroscience methods, 325, 108325. https://doi.org/10.1016/j.jneumeth.2019.108325
3. Kupari, J., Häring, M., Agirre, E., Castelo-Branco, G., & Ernfors, P. (2019). An Atlas of Vagal Sensory Neurons and Their Molecular Specialization. Cell reports, 27(8), 2508–2523.e4. https://doi.org/10.1016/j.celrep.2019.04.096
4. Jayaprakash, N., Song, W., Toth, V., Vardhan, A., Levy, T., Tomaio, J., Qanud, K., Mughrabi, I., Chang, Y. C., Rob, M., Daytz, A., Abbas, A., Nassrallah, Z., Volpe, B. T., Tracey, K. J., Al-Abed, Y., Datta-Chaudhuri, T., Miller, L., Barbe, M. F., Lee, S. C., … Zanos, S. (2023). Organ- and function-specific anatomical organization of vagal fibers supports fascicular vagus nerve stimulation. Brain stimulation, 16(2), 484–506. https://doi.org/10.1016/j.brs.2023.02.003
5. Rossetti, N., Song, W., Schnepel, P., Jayaprakash, N., Koutsouras, D. A., Fichman, M., Wong, J., Levy, T., Elgohary, M., Qanud, K., Giannotti, A., Barbe, M. F., Chen, F. L., Langereis, G., Datta-Chaudhuri, T., Mihajlović, V., & Zanos, S. (2025). Control of spatiotemporal activation of organ-specific fibers in the swine vagus nerve by intermittent interferential current stimulation. Nature Communications, 16(1). https://doi.org/10.1038/s41467-025-59595-4
6. Blanz, S. L., Musselman, E. D., Settell, M. L., Knudsen, B. E., Nicolai, E. N., Trevathan, J. K., Verner, R., Begnaud, J., Skubal, A., Suminski, A. J., Williams, J., Shoffstall, A. J., Grill, W. M., Pelot, N. A., & Ludwig, K. A. (2023). Spatially selective stimulation of the pig vagus nerve to modulate target effect versus side effect. Journal of Neural Engineering, 20(1), 016051–016051. https://doi.org/10.1088/1741-2552/acb3fd