A Decade of SPARC and Over $350 Million in Funding: From Neural Mapping to Precision Vagus Nerve Modulation

Introduction

Heartbeat, gastrointestinal motility, and bladder filling are all regulated by the autonomic nervous system. By influencing neural activity through electrical stimulation, researchers hope to develop new interventions for organ dysfunction. However, accurately directing stimulation to disease-relevant nerve fibers and producing reproducible organ responses requires answers to several specific questions. Where do the nerves project? How much do they vary between individuals? Which functions change after stimulation, and how can off-target effects be detected promptly?

In 2026, Bhavesh Patel and colleagues published a review in Bioelectronic Medicine summarizing the major outcomes of the US National Institutes of Health (NIH) Stimulating Peripheral Activity to Relieve Conditions (SPARC) program. Launched in 2015, the program awarded more than $350 million to nearly 100 projects, which generated over 500 publications.

The review describes how neural maps, technological tools, and open resources collectively support precision neuromodulation. It also highlights the challenges that remain in translating foundational research into clinical applications.

Key Challenges in Precision Neuromodulation

The vagus nerve connects the brain with multiple internal organs and is an important target in bioelectronic medicine. However, a single nerve contains fibers with different trajectories and functions, so stimulation may affect both the target organ and other physiological processes. Electrode placement, fascicular organization, and stimulation parameters jointly determine which fibers are recruited.

The cross-species studies summarized in the review identified morphological differences among human, pig, and rat vagus nerves, with greater variability among human samples. Modeling further showed that applying identical parameters can produce markedly different neural responses, both between individuals within a species and across species. Directly reusing or linearly scaling stimulation parameters from animal experiments does not guarantee equivalent neural recruitment in humans. This is one reason SPARC has pursued anatomical mapping, functional recording, and computational modeling together. Understanding both nerve structure and stimulation responses can support more informed choices of electrodes and parameters.

Research Framework and Phase-Specific Objectives

Phase 1 of SPARC ran from 2015 to 2021 and focused on anatomical and functional mapping, technology development, translational research, and data infrastructure. Early mapping efforts primarily covered the heart, lungs, stomach, colon, and bladder, before expanding to other organs, large animals, and disease models.

Phase 2, launched in 2022, placed greater emphasis on the human vagus nerve. The Reconstructing Vagal Anatomy (REVA) project used human cadaver dissection and multiscale imaging to characterize the nerve along its course and at branching points, comparing features such as nerve diameter, fascicle count, and perineurium thickness.

The Vagus Nerve Stimulation (VNS) Endpoints from Standardized Parameters (VESPA) study collects multi-organ physiological readouts in response to standardized stimulation in patients with implanted VNS devices. Such studies help establish relationships between stimulation parameters and physiological responses. However, the review does not provide a common sample size, control design, or set of primary endpoint results for assessing efficacy across indications. Following the program's formal conclusion in 2025, some activities have continued through no-cost extensions, data sharing, and related initiatives.

Advances in Neural Mapping and Technology

One outcome of SPARC is the organization of previously dispersed knowledge about neural connections into searchable maps. Researchers can select a nerve in a three-dimensional whole-body anatomical model and examine the associated neuronal populations and their pathways.

Figure 1. Three-Dimensional Whole-Body Anatomy and Neural Connectivity Maps. The left panel shows the left second lumbar splanchnic nerve selected in a three-dimensional whole-body model. The right panels display the associated neuronal populations traveling through that nerve on female and male connectivity maps. By linking spatial anatomy with connectivity information, the platform supports anatomical queries and hypothesis generation.

Technology development has further incorporated structural information into stimulation design. Automated Simulations to Characterize Electrical Nerve Thresholds (ASCENT) combines sample-specific anatomy, electrode geometry, and stimulation parameters to predict neural activation, helping researchers compare alternative approaches before conducting experiments. Model predictions still need to be checked against experimental recordings.

Another direction integrates stimulation with physiological feedback. The OpenNerve system uses a modular design that supports neural stimulation and the acquisition of multiple types of biological signals. Closed-loop modulation aims to use organ state to determine when stimulation should start, stop, or be adjusted, while monitoring potential off-target effects. System capabilities and long-term benefits in specific diseases require separate validation.

Figure 2. Neural Stimulation and Organ Signal Sensing System. Blue leads connect to neural stimulation electrodes, while red leads acquire signals from internal organs. The implantable pulse generator communicates with an external controller, such as a smartphone, using Bluetooth Low Energy. By integrating stimulation with physiological signal acquisition, the system provides a technical basis for closed-loop modulation. This is a schematic of the system; clinical benefits must be established separately for each indication.

Open resources are another major outcome of the program. As of February 2026, the SPARC Portal had published 414 datasets totaling more than 86 TB, comprising 308 research datasets, 55 computational models, and 51 anatomical models. These resources cover 7 species, 77 anatomical structures, and 26 experimental approaches. They enable different teams to reuse existing experiments and models for cross-species comparisons and the evaluation of stimulation strategies.

Figure 3. Overview of Open Data and Model Resources. The left panel summarizes 414 published datasets totaling more than 86 TB, comprising 308 research datasets, 55 computational models, and 51 anatomical models. They cover 77 anatomical structures, 7 species, and 26 experimental approaches. These statistics are current through February 2026. The right panel shows the geographical distribution of visitors and annual visitor numbers, with statistics through December 2025.

Scope and Limits of Translational Evidence

Translational studies have provided several proofs of concept. For example, an anesthetized feline study cited in the review used bladder sensory feedback recorded from the dorsal root ganglia to trigger closed-loop sacral nerve stimulation. The study reported improved bladder capacity and a reduction of more than 50% in stimulation time compared with open-loop stimulation. These findings support further investigation of feedback-driven modulation. However, they are animal data involving a stimulation target other than the vagus nerve and cannot be interpreted as evidence of clinical efficacy for VNS or transcutaneous auricular vagus nerve stimulation (taVNS).

At a mechanistic level, SPARC's contribution lies in progressively connecting neural structure, fiber recruitment, and organ responses. Anatomical maps support localization, models support prediction, and physiological recordings test whether stimulation produces the intended response. Together, these approaches can help guide subsequent mechanistic experiments and parameter optimization. They do not establish a single therapeutic pathway that applies to all diseases.

The article is not a systematic review of efficacy and does not provide a standardized literature search and screening process or pooled effect estimates. Its examples span different species, devices, and stages of research and cannot be combined into an overall efficacy conclusion. The authors disclosed relationships involving company employment, stock ownership, and patent licensing. The prospects of these technologies should therefore be assessed alongside independent validation, sensor reliability, and the long-term stability of implanted devices.

ZenoWell: Translating Auricular taVNS into Wearable Wellness

ZenoWell applies auricular transcutaneous vagus nerve stimulation (taVNS) in an ear-worn format designed for everyday wellness routines. Its devices deliver electrical stimulation through contacts positioned on the outer ear, targeting regions associated with the auricular branch of the vagus nerve. Depending on the model, users can select modes for sleep preparation, relaxation, meditation, and other routines. Compatible ZenoWell tools also support heart rate and heart rate variability (HRV) tracking, helping users observe physiological trends over time. However, HRV changes alone do not establish direct vagal activation or prove clinical efficacy. ZenoWell's products are separate from the NIH SPARC initiative, and findings from foundational neural mapping and computational modeling should not be interpreted as validation of any specific consumer device. Product-specific controlled studies remain important for establishing efficacy and safety.

Priorities for Further Validation

Future work could combine human anatomical variability with standardized physiological recordings to test whether models can predict individual responses to stimulation and whether parameter selection is reproducible in independent cohorts. Closed-loop systems also require real-world assessments of sensing stability, false triggering, and off-target effects, followed by appropriately controlled human studies to establish sustained improvements in symptoms, function, and quality of life while documenting long-term adverse events and device maintenance requirements. In addition, open data and models should retain clear version histories, experimental conditions, and provenance so that different teams can verify findings and progressively determine which technologies are suitable for clinical use and for which patients.

Reference

Patel B, Pikov V, Damaser MS, et al. Accelerating the development of bioelectronic medicine: outcomes and innovations from the NIH SPARC initiative. Bioelectronic Medicine. 2026;12:20.

DOI: https://doi.org/10.1186/s42234-026-00213-z.

Resource: SPARC Portal (https://sparc.science)