Vagus nerve stimulation — VNS — is now an established neuromodulation treatment for drug-resistant epilepsy. Its history, however, began long before implantable pulse generators existed.
The idea developed through several very different stages: nineteenth-century attempts to alter cerebral blood flow, physiological experiments showing that stimulation of the vagus could change brain activity, animal seizure studies, the creation of an implantable system, pilot human implants in 1988 and large clinical trials that eventually led to regulatory approval.
This page follows that history. The site’s separate modern VNS page explains how contemporary VNS is used in epilepsy care.
The vagus nerve was not originally approached as an epilepsy pathway
The vagus is a mixed cranial nerve with extensive connections between the brainstem and organs of the chest and abdomen.
Its involvement in heart rate, respiration and visceral regulation was recognised long before its potential role in neuromodulation was considered.
Nineteenth-century physicians did not yet understand the vagus in modern network terms. Some epilepsy treatments involving the neck were instead based on theories about cerebral circulation.
James Leonard Corning and the 1880s
One of the strangest precursors of modern VNS came from New York neurologist James Leonard Corning.
Corning believed that seizures were related to excessive cerebral blood flow or ‘venous hyperaemia’, an idea that was influential in parts of nineteenth-century medicine.
He developed mechanical devices including a ‘carotid fork’ and a neck-worn ‘carotid truss’ designed to compress the carotid arteries and reduce blood flow to the brain.
He later combined carotid compression with transcutaneous electrical stimulation of structures in the neck, including what he believed to be the vagus nerve.
These experiments are sometimes described as the earliest therapeutic use of electrical vagal stimulation for epilepsy.
Corning was not doing modern VNS
The historical connection should not be overstated.
Corning’s treatment was based on an incorrect theory of epilepsy.
His electrical stimulation was delivered externally rather than through an implanted cervical electrode.
The intervention also affected the carotid sheath and likely stimulated structures other than the vagus nerve.
Reported adverse effects included bradycardia, dizziness and syncope.
The technique was not widely accepted and disappeared from routine practice.
Its importance is therefore historical rather than technological: it shows that physicians had experimented with electrically manipulating the vagal region more than a century before modern VNS.
Early twentieth-century physiology asks a different question
During the twentieth century, research moved away from Corning’s cerebral-blood-flow theory.
Physiologists became interested in whether vagal afferent activity could influence electrical activity within the brain itself.
This was crucial because most fibres in the vagus carry information toward the brain rather than away from it.
If stimulation of those fibres could alter central nervous-system activity, the nerve could potentially provide an indirect route for influencing brain networks without operating inside the skull.
1938: Bailey and Bremer
In 1938, Percival Bailey and Frédéric Bremer studied the effects of vagus nerve stimulation in cats.
Their experiments showed that vagal stimulation could alter cortical electrical activity.
This was an important conceptual change.
The vagus was no longer merely a nerve controlling visceral function. Electrical signals travelling through it could influence cerebral activity.
The work did not establish VNS as an epilepsy treatment, but it supplied physiological evidence that stimulation of the nerve could reach and modify the brain.
Later animal physiology maps the response
Other investigators continued studying vagal stimulation in animals during the following decades.
Experiments examined cortical synchronisation and desynchronisation, brainstem responses and the effects of stimulating different vagal fibres.
By the 1960s, researchers including Chase and colleagues were investigating how cervical vagal stimulation altered cortical and subcortical electrical activity.
The results were complex rather than uniformly suppressive.
Different stimulation parameters and different fibre populations could produce different patterns of cerebral response.
The key therapeutic step came from seizure models
The leap from physiology to epilepsy treatment came when investigators began asking whether vagal stimulation could suppress seizures rather than merely alter EEG rhythms.
Jacob Zabara became the key figure in this transition.
Working at Temple University in Philadelphia, Zabara investigated cervical vagus nerve stimulation in animal seizure models during the 1980s.
Jacob Zabara and the canine experiments
Zabara found that repetitive electrical stimulation of the cervical vagus nerve could interrupt or suppress experimentally induced motor seizures in dogs.
The effect could occur rapidly.
Importantly, the antiseizure effect persisted when the nerve was cut below the stimulation site in experimental work, supporting the idea that the therapeutic signal travelled centrally through vagal afferent fibres rather than depending primarily on peripheral organ effects.
This gave modern VNS its central therapeutic hypothesis: stimulate a peripheral nerve whose afferent pathways influence widespread brain networks.
1985: the idea becomes a proposed epilepsy therapy
By the mid-1980s, Zabara was proposing repetitive vagal stimulation as a potential treatment for epilepsy.
Early animal reports helped persuade researchers and engineers that the approach might be clinically testable.
The problem was no longer simply whether vagal stimulation could influence the brain.
The next challenge was whether it could be delivered chronically, safely and reproducibly in humans.
Why an implantable device was necessary
A practical long-term treatment required much more than an electrode placed temporarily on the nerve.
The system needed a pulse generator, an implanted lead, a reliable way to attach the lead to the cervical vagus nerve, programmable stimulation and a battery capable of functioning for years.
The design drew conceptually from implanted cardiac pacemakers but had to solve different anatomical and neurostimulation problems.
This became the basis of the NeuroCybernetic Prosthesis system developed by Cyberonics.
Cyberonics and the NeuroCybernetic Prosthesis
Cyberonics was founded in the late 1980s to develop an implantable vagus nerve stimulation system.
The device consisted of a pulse generator implanted in the chest and a lead tunnelled beneath the skin to electrodes wrapped around the cervical vagus nerve.
The system delivered intermittent electrical stimulation according to programmed settings.
Unlike resective epilepsy surgery, it did not attempt to remove the brain region producing seizures.
It represented a different surgical philosophy: modify seizure networks through neuromodulation.
Why the left vagus nerve became standard
Clinical VNS was developed primarily using the left cervical vagus nerve.
This choice reflected concern about cardiac effects because vagal branches on the right have historically been considered more closely associated with sinoatrial-node control.
The anatomical physiology is more complex than that simple rule suggests, but left-sided implantation became established during development and clinical trials.
The early implanted system was therefore designed around left cervical stimulation.
1988: the first human implant
The first human implantation of the modern VNS Therapy technology took place in 1988.
Neurologist James Kiffin Penry and neurosurgeon William Bell implanted the system at Bowman Gray School of Medicine in the United States.
Early pilot studies involved a very small number of people with severe drug-resistant epilepsy.
The aim at this stage was primarily to establish whether chronic vagal stimulation was technically feasible and tolerable.
The first patients produced an unexpectedly encouraging signal
Early reports described four implanted patients.
Two became seizure-free during the observation period, one experienced a substantial reduction in seizure frequency and one showed no meaningful seizure improvement.
Those numbers were far too small to establish efficacy.
But they were encouraging enough to justify larger controlled studies.
The key historical point is that VNS moved from an animal model to chronic human implantation in only a few years.
A treatment without a visible brain target
This was unusual within epilepsy surgery.
Resective procedures depended on identifying tissue that could be removed safely.
VNS did not require localisation of one resectable epileptogenic zone.
That made it particularly attractive for people whose seizures arose from multiple regions, could not be localised adequately, involved eloquent cortex or had persisted after previous surgery.
The operation itself was performed in the neck and chest rather than inside the skull.
The early clinical trials
During the late 1980s and early 1990s, pilot studies were followed by multicentre controlled trials.
Researchers compared higher-intensity stimulation with lower-intensity stimulation while attempting to keep participants and evaluators blinded to treatment assignment.
The trials showed that active VNS could reduce seizure frequency in a proportion of patients with medically refractory focal seizures.
The treatment was not a cure for most people.
Instead, it established a pattern that remains central to neuromodulation: meaningful seizure reduction in some patients, variable benefit between individuals and improvement that may accumulate over time.
1994: European approval
By June 1994, the VNS Therapy System had been approved for epilepsy treatment in countries of the European Union under the European regulatory framework then in force.
This was an important milestone because it moved the technology from experimental trials into clinical use outside research protocols.
Thousands of patient-years of experience would soon begin accumulating.
1997: United States FDA approval
On 16 July 1997, the US Food and Drug Administration approved the NeuroCybernetic Prosthesis VNS system for epilepsy.
The original US indication was adjunctive therapy for reducing seizure frequency in adults and adolescents over 12 years of age with medically refractory partial-onset seizures.
The approval followed pivotal clinical trials and marked the point at which implanted VNS became an established regulatory category rather than an experimental neurostimulation procedure.
Why the 1997 approval mattered historically
VNS was the first implanted peripheral nerve stimulation therapy to achieve widespread regulatory acceptance specifically for epilepsy.
It also established neuromodulation as a major treatment pathway alongside medication, diet and resective surgery.
The goal was not to excise the epileptogenic tissue.
The goal was to alter the probability, spread or expression of seizures through repeated electrical stimulation.
This opened a new chapter in epilepsy treatment.
The term ‘adjunctive’ is important
From the beginning, regulatory language described VNS as adjunctive therapy.
That reflected the population studied.
Most patients remained on antiseizure medicines and had epilepsy that had not responded adequately to medication.
VNS was therefore introduced as an additional treatment rather than as a routine replacement for medication or resective surgery.
The device behaved differently from a drug
Early investigators noticed that the pattern of benefit did not always resemble medication response.
Some people improved gradually over months or years rather than immediately after stimulation was activated.
This raised questions about whether VNS caused longer-term network or neurochemical changes beyond the momentary effect of each electrical pulse.
The mechanism remained uncertain, and in many respects still is.
Historically, clinical adoption therefore preceded a complete mechanistic explanation.
The magnet introduced patient-controlled stimulation
The VNS system incorporated a magnet that could be passed over the generator to trigger additional stimulation.
This created an unusual relationship between patient and implanted device.
Some people who experienced warning symptoms or auras could attempt to activate an extra stimulation burst at seizure onset.
This did not work reliably for everyone, but it was an early form of patient-interactive neuromodulation.
Voice change became a recognisable signature
Because the vagus nerve supplies branches involved in laryngeal function, stimulation commonly produced transient voice alteration, throat sensation or cough during active stimulation.
These effects became characteristic of early VNS trials.
They also created difficulties for blinding in controlled studies because participants could sometimes tell when stronger stimulation was occurring.
This methodological issue became part of the history of evaluating implanted neurostimulation treatments.
VNS was not developed from a complete mechanism
One striking feature of VNS history is how much of the therapy was established empirically.
The vagus carries widespread afferent projections into the brainstem, including pathways involving the nucleus tractus solitarius and onward connections to broader neuromodulatory systems.
But the precise antiseizure mechanism was not known when VNS entered clinical practice.
Animal evidence, physiological plausibility and controlled clinical benefit were enough to support use before the full mechanism was understood.
From partial seizures to broader clinical use
The original US approval specifically referred to partial-onset seizures in people aged 12 years and older.
Clinical practice and research later expanded into broader epilepsy populations, including younger children and people with generalised or mixed seizure disorders in some jurisdictions and settings.
Regulatory indications also changed over time.
In 2017, for example, the FDA indication was expanded to include patients aged four years and older with refractory partial-onset seizures.
Terminology later shifted from ‘partial onset’ toward the modern term ‘focal onset’.
A platform rather than one frozen device
The original NeuroCybernetic Prosthesis did not remain technologically static.
Successive generations changed generator size, battery characteristics, programming and sensing capabilities.
Later systems introduced features designed to detect rapid heart-rate increases that can accompany some seizures and automatically deliver extra stimulation.
The modern device is therefore descended from the 1988 implant but is not technically identical to it.
VNS influenced later neuromodulation
The significance of VNS extends beyond the vagus nerve.
It demonstrated that chronic electrical stimulation could be used safely and repeatedly as long-term epilepsy therapy without removing brain tissue.
That helped establish a wider therapeutic space in which deep brain stimulation and responsive neurostimulation could later develop.
Those technologies target different structures and use different control strategies, but all belong to the broader shift toward modulating epileptic networks electrically.
It also changed what ‘epilepsy surgery’ could mean
Historically, epilepsy surgery had meant opening the skull to remove, cut or disconnect brain tissue.
VNS complicated that definition.
Implantation required surgery, but the therapeutic target was a peripheral nerve in the neck.
The brain itself was not resected or directly implanted.
VNS therefore sat between neurosurgery, neurology and biomedical engineering and helped create the modern category of device-based epilepsy therapy.
Why VNS persisted when many experimental treatments disappeared
Many historical epilepsy treatments produced promising early reports and then vanished.
VNS survived because its development passed through controlled trials, regulatory review, long-term follow-up and repeated technical refinement.
Its benefits were modest for many patients but reproducible enough to justify continued use.
It also addressed a real therapeutic gap: people with drug-resistant epilepsy who were not suitable candidates for curative resection.
The central historical lesson
The history of VNS did not begin with a modern understanding of epilepsy networks.
It began with crude nineteenth-century ideas about blood flow, moved through basic physiology showing that the vagus could influence the brain, and became clinically credible only after animal seizure experiments demonstrated an antiseizure effect.
Jacob Zabara’s canine studies supplied the decisive modern hypothesis.
Penry, Bell and their colleagues then brought that hypothesis into human treatment in 1988.
Controlled trials and regulatory approvals during the 1990s transformed an unconventional experiment into an established therapy.
The most important change was conceptual.
For the first time in modern epilepsy treatment, an implanted device could reduce seizures by electrically modulating a neural pathway outside the brain rather than removing the tissue where seizures arose.
That idea helped launch the era of chronic neuromodulation in epilepsy.
Sources and further reading
Lanska DJ. J. L. Corning and vagal nerve stimulation for seizures in the 1880s. Neurology. 2002.
The physiology, anatomy and stimulation of the vagus nerve in epilepsy. Review.
Zabara J. Inhibition of experimental seizures in canines by repetitive vagal stimulation. Epilepsia. 1992.
Ben-Menachem E. Modern management of epilepsy: vagus nerve stimulation. Baillière’s Clinical Neurology. 1996.
McLachlan RS. Vagus nerve stimulation for intractable epilepsy: a review. Journal of Clinical Neurophysiology. 1997.
Afra P et al. Evolution of the Vagus Nerve Stimulation Therapy System technology for drug-resistant epilepsy. Frontiers in Medical Technology. 2021.
US Food and Drug Administration. VNS Therapy System, PMA P970003. Original approval decision: 16 July 1997.
US FDA. VNS Therapy System marketing history and regulatory documentation.
Historical evidence checked: September 2026.