Vagus Nerve Stimulation (VNS)

What it is

Vagus nerve stimulation (VNS) is an implanted neuromodulation therapy in which a pulse generator, placed subcutaneously in the chest, delivers intermittent electrical stimulation to the left cervical vagus nerve through a helical lead. It was FDA-approved for treatment-resistant depression in 2005 (CE-marked in 2001), borrowed from epilepsy, where the antidepressant signal was first noticed as a mood improvement in stimulated seizure patients. Among the interventional modalities in this series, VNS is the one whose evidence base is most genuinely contested — not because it plainly fails, but because its benefit is delayed, cumulative, and most visible on exactly the measures that rigorous acute trials are least able to capture. The result is a treatment that has repeatedly missed primary endpoints in controlled trials while showing durable, clinically meaningful effects in longer-term and open-label data. Holding that tension honestly is the whole point of the VNS story.

How it works

The vagus is overwhelmingly an afferent nerve — roughly 80% of its fibers carry information upward, from viscera to brain — which is what makes it a route into central circuitry. Stimulation of vagal afferents projects first to the nucleus tractus solitarius in the medulla, which then relays to the locus coeruleus (the brain's principal noradrenergic source) and the dorsal raphe (its principal serotonergic source), as well as the parabrachial nucleus, hypothalamus, amygdala, and, via thalamic relays, the insula and prefrontal–cingulate cortex. VNS thus engages mood circuitry "bottom-up," modulating the very monoaminergic source nuclei that antidepressants target pharmacologically and altering activity across the limbic–cortical mood network. Functional imaging shows changes in orbitofrontal, cingulate, and limbic activity with chronic stimulation.

Two further mechanisms situate VNS within the broader themes of this library. First, the vagus is the efferent arm of the cholinergic anti-inflammatory pathway: vagal signaling to splenic macrophages, via α7-nicotinic receptors, suppresses pro-inflammatory cytokine release — a direct link to the inflammation account of depression. Second, locus-coeruleus–driven noradrenergic tone promotes neuroplasticity and BDNF expression; the same plasticity-gating property underlies VNS-paired rehabilitation in stroke. The plausible unifying account is that VNS slowly remodels mood-network function through combined monoaminergic, anti-inflammatory, and neuroplastic effects — which would explain its characteristically gradual onset.

The lead is placed on the left vagus specifically to minimize cardiac effects, since the right vagus more heavily innervates the sinoatrial node. Stimulation is intermittent and chronic — a typical duty cycle delivers stimulation for ~30 seconds every few minutes, around the clock, for years.

The evidence

VNS has a long and instructive evidentiary record, defined by a persistent discordance between acute controlled trials and longer-term outcomes.

The original pivotal acute trial (D-02, 2005) was a 10-week sham-controlled study that was negative on its primary endpoint — VNS did not separate from sham acutely. What secured approval was the contrast between the long-term naturalistic comparison (VNS plus treatment-as-usual outperforming treatment-as-usual alone over a year) and the acute data, suggesting the effect simply takes longer than ten weeks to emerge. The most striking long-term signal came from a five-year observational registry (Aaronson et al., 2017): VNS plus usual care was associated with higher cumulative response and remission than usual care alone, and — notably — with lower all-cause mortality. That mortality finding is provocative but must be read with its design in mind: it is non-randomized and vulnerable to selection and treatment-intensity confounds.

The definitive modern test is the RECOVER trial, the largest randomized study of VNS ever conducted, designed with the Centers for Medicare and Medicaid Services under a Coverage-with-Evidence-Development agreement. The unipolar cohort randomized 493 adults with markedly treatment-resistant depression — a population that had failed, on average, more than a dozen prior treatments — to active or no-stimulation (sham) VNS for a full 12 months of double-blind follow-up.

The result is a near-perfect case study in reading evidence honestly:

• The primary endpoint was not met. Percent of time in MADRS response (≥50% improvement) across months 3–12 was 18.9% with active VNS versus 16.3% with sham — not statistically significant (p = 0.137 against a predetermined threshold of p < 0.023).
• Multiple secondary endpoints significantly favored active VNS. Percent time in response on clinician-rated CGI-I (p = 0.004) and self-rated QIDS-SR (p = 0.049) separated from sham, as did percent time in partial response (≥30% improvement) on CGI-I (p < 0.001) and QIDS-C (p = 0.006). Parallel analyses showed benefits in quality of life and daily function. The dyspnea rate was higher with active stimulation (a known stimulation effect), and no new safety signals emerged; 88% completed the trial.
• The open-label extension following the blinded year characterized durable benefit over 24 months in those who continued — consistent with the long-standing observation that VNS benefit accrues and then holds.

The interpretation is genuinely contested, and reasonable clinicians differ. A strict reading notes that the prespecified primary outcome failed, and that secondary-endpoint positivity in a sponsored trial warrants caution. A more clinically sympathetic reading — articulated in accompanying editorials — emphasizes that in a population this refractory, the convergence of partial-response, functional, and quality-of-life benefits across multiple instruments is meaningful, and that the MADRS percent-time-in-response metric may be an unusually stringent way to detect a slow, partial, durable effect. Both readings are defensible. What is not defensible is quoting either the negative primary endpoint or the positive secondaries in isolation.

Who it's for

VNS is an adjunctive, long-horizon therapy for chronic or recurrent, markedly treatment-resistant depression — typically patients who have failed multiple adequate medication trials and often a course of ECT or other interventions. It is not an acute treatment and is wholly unsuited to anyone needing rapid relief; its value proposition is durability in a relapsing illness, not speed. Candidate selection is genuinely difficult, and recent RECOVER analyses have sought prognostic and prescriptive predictors of response precisely because identifying the right patient remains the central clinical challenge. Given the invasiveness, cost, and slow onset, VNS sits well down the interventional sequence — after ECT, TMS, and pharmacological strategies have been exhausted, and as an alternative to or staging point before DBS.

Procedure, programming, and what to expect

Implantation is a relatively minor outpatient surgical procedure: the generator is placed in a chest pocket and the lead tunneled to the left cervical vagus. After a healing interval, stimulation is titrated gradually over weeks to a tolerable output, balancing efficacy against the stimulation-related side effects below. The defining feature patients must understand up front is latency: meaningful benefit characteristically emerges over six to twelve months or longer, and continues to build thereafter. This delayed, cumulative trajectory is unlike any other antidepressant intervention and demands realistic expectation-setting — a patient anticipating rapid change will discontinue before the therapy has had a chance to work. The generator battery is replaced every several years in a minor procedure.

Side effects and risks

VNS side effects are predominantly stimulation-related and cyclical, occurring during the "on" phase of the duty cycle and frequently attenuating over time and with programming adjustment. The most common is voice alteration or hoarseness (from the lead's proximity to the recurrent laryngeal branch), followed by cough, throat or neck pain, dyspnea, dysphagia, and paresthesias. Most are tolerable and adjustable by reducing output or pulse width. Surgical risks are those of any device implantation — infection, lead malfunction, and, rarely, vocal cord paralysis or intraoperative bradycardia/asystole. Obstructive sleep apnea can worsen. MRI compatibility is constrained by the implant.

Crucially, and in sharp contrast to ECT, VNS is not associated with cognitive impairment — there is no amnestic effect, and if anything a modest pro-cognitive signal has been reported. For a treatment-resistant patient weighing durable maintenance against the cognitive trade-offs of ECT, this is a meaningful point of differentiation.

The non-invasive frontier: transcutaneous VNS

A growing research literature examines transcutaneous auricular VNS (taVNS), which stimulates the auricular branch of the vagus at the cymba conchae of the ear with an external device — no surgery, no implant. taVNS engages the same NTS-centered afferent pathway and has been shown to modulate the default mode network in depressed patients, with early controlled studies suggesting antidepressant effects. taVNS is currently investigational for depression and not a substitute for implanted VNS, but it is an active and appealing frontier precisely because it could deliver vagal neuromodulation without the cost, surgical risk, and irreversibility of implantation.

Where it fits

VNS exemplifies the network-and-plasticity convergence that runs through this entire series: like ECT, TMS, and DBS, it ultimately works by remodeling distributed mood-network function and promoting plasticity, despite entering the system at a completely different node (a peripheral nerve rather than cortex or deep gray matter). Its anti-inflammatory mechanism ties it uniquely among the neurostimulation modalities to the inflammation account, and its noradrenergic engagement to monoaminergic dysfunction and neuroplasticity. On the field's invasiveness-versus-evidence gradient, VNS occupies an awkward position — more invasive than TMS, with an evidence base more equivocal than its invasiveness would seem to demand — which is exactly why honest framing of the RECOVER results matters so much for appropriate use.

Bottom line

The landmark RECOVER trial missed its primary endpoint yet showed significant benefits across multiple secondary, functional, and quality-of-life measures, with durable effects in open-label extension — a result that supports careful use in the right refractory patient while resisting overstatement. It is distinguished from ECT by its lack of cognitive side effects and its durability, and limited by its invasiveness, cost, six-to-twelve-month latency, and contested evidence. For a carefully selected, severely treatment-resistant patient prepared for a long horizon, VNS is a legitimate option; for anyone else, its trade-offs are hard to justify. taVNS may eventually offer the mechanism without the implant, but remains investigational.

Key references

1. Conway CR, Aaronson ST, Sackeim HA, et al. Vagus nerve stimulation in treatment-resistant depression: a one-year, randomized, sham-controlled trial (RECOVER). Brain Stimul. 2025;18(3)
   –689.
2. Aaronson ST, et al. Effects of vagus nerve stimulation on daily function and quality of life in markedly treatment-resistant major depression: a one-year, randomized, sham-controlled trial (RECOVER). Brain Stimul. 2025;18(3)
   –700.
3. (Editorial) Beyond the primary outcome: why the RECOVER trial for vagus nerve stimulation matters in treatment-resistant depression. Brain Stimul. 2025.
4. Aaronson ST, et al. Durability of the benefit of vagus nerve stimulation in markedly treatment-resistant major depression: a RECOVER trial report (24-month open-label extension). 2025.
5. Aaronson ST, Sackeim HA, et al. Prognostic and prescriptive predictors of treatment response to adjunctive VNS therapy: a RECOVER trial report. J Clin Psychiatry. 2025.
6. Aaronson ST, et al. A 5-year observational study of patients with treatment-resistant depression treated with VNS or treatment as usual: comparison of response, remission, and suicidality. Am J Psychiatry. 2017.
7. Rush AJ, et al. Vagus nerve stimulation for treatment-resistant depression: a randomized, controlled acute phase trial (D-02). Biol Psychiatry. 2005.
8. George MS, et al. A one-year comparison of vagus nerve stimulation with treatment as usual for treatment-resistant depression. Biol Psychiatry. 2005.
9. Nahas Z, et al. Two-year outcome of vagus nerve stimulation for treatment-resistant depression. J Clin Psychiatry. 2005.
10. Sackeim HA, et al. Vagus nerve stimulation for treatment-resistant depression: efficacy, side effects, and predictors of outcome. Neuropsychopharmacology. 2001.
11. Carreno FR, Frazer A. Vagal nerve stimulation for treatment-resistant depression. Neurotherapeutics. 2017.
12. Bonaz B, Sinniger V, Pellissier S. Vagus nerve stimulation: a new promising therapeutic tool in inflammatory bowel disease and depression. J Intern Med. 2017.
13. Müller HHO, et al. Vagus nerve stimulation (VNS) and other augmentation strategies for therapy-resistant depression. Front Neurosci. 2018.
14. Fang J, et al. Transcutaneous vagus nerve stimulation modulates default mode network in major depressive disorder. Biol Psychiatry. 2016.
15. Rong P, et al. Effect of transcutaneous auricular vagus nerve stimulation on major depressive disorder. J Affect Disord. 2016.
16. Howland RH. Vagus nerve stimulation. Curr Behav Neurosci Rep. 2014.
17. Cimpianu CL, et al. Vagus nerve stimulation in psychiatry: a systematic review. J Neural Transm. 2017.
18. George MS, et al. Mechanisms of action of vagus nerve stimulation. Neuropsychopharmacology (review). 2007.
19. Tracey KJ. The inflammatory reflex and the cholinergic anti-inflammatory pathway. Nature / J Clin Invest.