Vagus nerve stimulation (active)

What it is

The vagus nerve is the tenth and longest cranial nerve. It runs from the brainstem to the heart, lungs, stomach and gut; the name means the wandering one. It is usually described as the brake of the nervous system, as the main line of the parasympathetic system. That is half the truth.

The other half: 80 to 90 percent of its fibers run upward, from the body to the brain. The nerve is predominantly a reporting line, not a command cable. That is exactly why stimulation works at all. Stimulating the nerve does not primarily talk to the stomach; it sends a signal to nuclei in the brainstem that influence attention, mood and inflammation.

The approved medicine

In 1997 the US FDA approved an implanted stimulation system for hard-to-treat epilepsy: a pacemaker under the collarbone, an electrode around the nerve at the neck. Treatment-resistant depression followed in 2005, and in 2021 a system for rehabilitation after stroke. Non-invasive devices held against the neck are cleared in the US for cluster headache and migraine, backed by placebo-controlled studies.

Alongside this there is the home market: ear clips and electrodes for the outer ear, transcutaneous auricular vagus nerve stimulation. They stimulate the auricular branch of the nerve and are advertised for relaxation, sleep, HRV and stress resilience, usually as a wellness product without approval for any disease.

What was measured in healthy people

The most positive study in healthy people appeared in 2025 in the European Heart Journal. 28 healthy volunteers received, blinded and in crossover, 30 minutes daily for seven days of stimulation at both ears or sham stimulation. Maximal oxygen uptake rose by 3.8 percent, maximal power output by 6 watts, and in blood samples from five participants the inflammatory response was weaker. This is one of the few times that something came out of a blinded measurement in healthy people.

One detail does not fit the calming narrative: heart rate rose by 4 beats per minute. The paper reports the increase in the abstract for peak exercise and in a figure for the resting pulse while lying down. If the simple brake model were right, the pulse would have to fall. A meta-analysis of 18 studies on ear stimulation, by contrast, found on average slightly lower blood pressure and pulse, with large heterogeneity. Something happens, but the mechanism is open.

Home remedies and user reports

Humming, slow exhalation, cold water on the face: that such things shift HRV in the short term is readily measurable. Many users of ear devices report that they become calmer and fall asleep more easily. These are user reports, and the feeling is real. But HRV is a marker, not the goal. That a number moves does not yet mean that someone becomes healthier. Exhaling slowly costs nothing and harms no one, a good starting point before buying any device.

What the studies show

RECOVER 2025 — depression, one year, sham-controlled

493 adults with unipolar depression who had not responded to at least four treatments in the current episode, with an average of 13.3 treatments over their lifetime. All received an implant; by lot it was switched on in 249 and not in 244, double-blind for twelve months. The primary endpoint, the proportion of time with symptom burden on the MADRS scale reduced by at least half, was missed. Several secondary endpoints came out significantly in favor of active stimulation, such as the clinician’s global impression with p equal to 0.004. Shortness of breath occurred more often under stimulation.

Ackland 2025 — endurance in healthy people

Single-center, randomized, double-blind, sham-controlled, crossover design with 28 healthy volunteers and a two-week washout period. The primary endpoint was maximal oxygen uptake. It rose after the active phase by 1.04 milliliters per kilogram per minute, corresponding to 3.8 percent, p equal to 0.005, but not under sham. Endpoint met. Small and short, but properly designed.

Wolf 2021 — HRV under ear stimulation

Bayesian meta-analysis of 16 single-blinded studies in healthy people, active ear stimulation versus sham. The result was strong evidence for the null hypothesis: acute stimulation does not measurably change vagally mediated HRV. HRV is therefore not suitable as reliable proof of effect for ear devices.

Kim 2022 — safety

Systematic review and meta-analysis of 177 studies with 6,322 participants. 55.37 percent of the studies did not mention side effects at all. Where they were reported, the rate was 12.84 per 100,000 person-minute-days of stimulation, with no difference from control.

Where the data stop

The largest and cleanest study missed its prespecified primary endpoint. Prespecified secondary endpoints are worth more than subgroups found after the fact, but they do not replace the primary endpoint. For ear stimulation there are many small randomized studies with questionnaire endpoints and hardly any long-term data. The settings differ greatly between studies, in frequency as well as intensity, and which parameters work best when is open.

Added to this is a methodological problem: at the ear, which is densely packed with nerve endings, a truly ineffective sham stimulation is hard to find. A review reports that active and sham stimulation can influence heart activity similarly. And the fact that more than half of the studies provide no information on side effects is a source of under-reporting. No reported side effects is not the same as no side effects.

Status, approval and legal

Implanted vagus nerve stimulators are medical devices for clearly defined indications, above all hard-to-treat epilepsy and depression, and are implanted by specialized centers. Non-invasive neck devices are cleared in the US for headache disorders. Ear devices for home use are mostly sold as wellness or relaxation products, without an approval for a specific disease behind them. Advertising claims such as no side effects or better HRV are not covered by the data.

Safety

According to the large meta-analysis, ear stimulation is considered safe, with mostly mild, transient complaints: ear pain, headache, tingling, occasionally tiredness. The asterisk is the incomplete reporting in more than half of the studies. Implanted stimulation is a surgical procedure with its own risks; in RECOVER, shortness of breath occurred more often under active stimulation. Because the vagus helps regulate heart rate, anyone with heart rhythm disorders or a cardiac implant should clarify with a doctor beforehand whether a device is an option. Anyone with epilepsy or severe depression belongs with this question in the hands of a specialist, not on an ear clip from the internet.

Sources

• Breit et al., Front Psychiatry 2018 (anatomy and function)
• Conway et al., Brain Stimul 2025 (RECOVER, main publication)
• Rush et al., Brain Stimul 2025 (RECOVER, quality of life)
• Conway et al., Int J Neuropsychopharmacol 2026 (RECOVER, durability)
• Ackland et al., Eur Heart J 2025 (endurance in healthy people)
• Wolf et al., Psychophysiology 2021 (HRV meta-analysis)
• Kim et al., Sci Rep 2022 (safety, meta-analysis)
• Nagai et al., Eur Heart J Open 2025 (blood pressure, meta-analysis)
• FDA, PMA P970003 (epilepsy, 1997)
• FDA, PMA P210007 (Vivistim, stroke rehabilitation, 2021)