Vagus Nerve for Anxiety: What Actually Works, What Is Hype

If you have spent any time on health social media, you have been told to hum, gargle, splash your face with cold water, massage your ear or buy a clip-on device to “reset your vagus nerve” and switch off anxiety. Some of that rests on real physiology. Most of it rests on nothing. This guide separates the two: what the vagus nerve is, what the research actually shows about anxiety and the measure everyone uses as a stand-in for “vagal tone”, which techniques have controlled trials behind them, which ones have none, and when anxiety is a job for a clinician rather than a breathing drill.

This is general information from published research, not medical advice. If your anxiety is persistent, is getting worse, or comes with thoughts of harming yourself, see the “when to get help” section first.

The short version

• Slow breathing is the one “vagus” technique with a large evidence base. A meta-analysis of 223 studies found vagally mediated heart rate variability rose during slow breathing, right after a session and after multi-session training. Breathwork trials cut self-reported stress and anxiety by a small to moderate amount.
• Exercise and mindfulness do as much or more for anxiety, without the vagus branding. Resistance training, aerobic exercise and 8-week mindfulness courses each reduce anxiety with effect sizes roughly in the 0.3 to 0.5 range in meta-analyses; one trial found a mindfulness course matched a first-line antidepressant.
• Sleep and alcohol move your HRV more than any hack. One short night modestly lowers vagal HRV and sharply shifts the balance toward sympathetic drive; alcohol cut HRV in the first hours of sleep in a dose-dependent way across 4,098 people; a moderate dose, roughly two to four drinks for an average adult, cut RMSSD by about 6 ms.
• Cold showers did not change anxiety in the one large trial we found. Cold on the face slows your heart through a real reflex; that is not the same as treating anxiety.
• Humming, gargling, ear massage, foot massage, clip-on stimulators and “vagus reset” programs have either no trials or null results. The famous humming study measured nitric oxide in ten people’s noses. The ear-clip devices did nothing measurable to vagal HRV across 16 sham-controlled studies.
• Do not massage the sides of your neck to “stimulate the vagus”. Pressing on the carotid sinus is a clinical maneuver cardiologists do with an ECG running because it can stop the heart for several seconds.

What the vagus nerve is

The vagus is the tenth cranial nerve. It leaves the brainstem, runs down each side of the neck and branches to the heart, lungs, esophagus, stomach and most of the abdominal organs; reference texts describe it as the longest nerve in the body (StatPearls, Cranial Nerve 10). It is the main nerve of the parasympathetic system, the “rest and digest” side that slows the heart and drives digestion. Two facts matter for everything that follows.

First, it is mostly a sensory nerve. Reviews put the ratio of body-to-brain fibers against brain-to-body fibers at roughly 80:20 or even 9 to 1 (van Weperen and Vaseghi, Frontiers in Neuroscience, 2023; Wang and colleagues, Frontiers in Physiology, 2020). Most of what the vagus does is report upward on the state of your organs. “Stimulating your vagus” is a much vaguer idea than the hacks make it sound.

Second, nobody measures your vagal tone. What gets measured is heart rate variability (HRV), the beat-to-beat variation in the timing of your heartbeats. Two HRV numbers, RMSSD and high-frequency power, track the vagal influence on the heart and are used as an index of “cardiac vagal tone” (Laborde, Mosley and Thayer, Frontiers in Psychology, 2017). An index is not a nerve reading. Keep that in mind whenever a product promises to “raise your vagal tone”: it means a number on a wearable, measured at the heart, under a lot of other influences.

One more caveat. A lot of the online content about “vagal safety” and “ventral vagal states” comes from polyvagal theory. That framework is popular in coaching and therapy circles and contested in the research literature; a 2023 paper in Biological Psychology argues that each of its five founding premises is likely refuted (Grossman, Biological Psychology, 2023). You do not need the theory for any of the practical advice below.

Anxiety and the vagus: what the research shows

There is a real signal here, and it is smaller than the marketing. A meta-analysis of 36 studies, 2,086 people with anxiety disorders against 2,294 controls, found lower resting HRV in the anxiety groups: a Hedges’ g of -0.29 for high-frequency HRV, which the authors call a small-to-moderate effect (Chalmers and colleagues, Frontiers in Psychiatry, 2014). So low HRV travels with anxiety on average. That is an association, not a diagnosis and not a treatment target.

The gap between the two shows up in the trials. In the best-known breathing study, mood improved while HRV did not separate between groups. In a meta-analysis of 19 mindfulness trials, programs that reduce anxiety in other reviews showed “insufficient evidence” of raising vagally mediated HRV (g = 0.19 after one outlier was removed, confidence interval crossing zero) (Brown and colleagues, Psychosomatic Medicine, 2021). People got less anxious without their vagal index moving. The honest reading: what helps anxiety helps anxiety, and whether it works “through the vagus” is mostly unproven. That changes nothing about what to do; it changes a lot about what to pay for.

What has real evidence

Slow breathing, about five to six breaths a minute

This is the one technique in the vagus canon that is well studied. A 2022 meta-analysis pooled 223 studies and found vagally mediated HRV increased during slow breathing (172 studies), immediately after a single session (16 studies) and after multi-session training (49 studies); the authors call it “a low-cost, low-tech technique” with few adverse effects expected (Laborde and colleagues, Neuroscience and Biobehavioral Reviews, 2022). A separate systematic review of slow breathing (under 10 breaths per minute) linked it to “increased comfort, relaxation” and reduced anxiety, though only 15 studies met its inclusion bar (Zaccaro and colleagues, Frontiers in Human Neuroscience, 2018).

How much does it help how you feel? A 2023 meta-analysis of randomized trials found breathwork reduced self-reported stress by a Hedges’ g of -0.35 (12 trials, 785 adults), anxiety by -0.32 and depressive symptoms by -0.40, and noted that most studies were at moderate risk of bias (Fincham and colleagues, Scientific Reports, 2023). Real, small to moderate, from mostly middling trials. That is a fair description of almost everything in this post.

The dose that keeps showing up is about six breaths per minute. HRV biofeedback researchers put the peak around 0.1 Hz, “six breaths per minute”, refined to about 5.5, varying a little between people (Lehrer and Gevirtz, Frontiers in Psychology, 2014). One simple way to land near that rate: breathe in for about four seconds and out for about six, which gives six breaths a minute, for five minutes. A remote randomized trial compared three five-minute daily breathing patterns with mindfulness meditation for one month and found breathwork, “particularly cyclic sighing” (a double inhale followed by one long exhale), improved mood and lowered breathing rate more than meditation did (Balban and colleagues, Cell Reports Medicine, 2023). The same trial is the one where HRV did not differ between groups.

Exercise

Nobody sells exercise as a vagus hack, which is odd, because it moves the same markers and has better anxiety data than any of the hacks. A 2025 meta-analysis of 34 randomized trials (1,434 participants) found long-term training, aerobic or resistance, shifted autonomic balance (a lower LF/HF ratio), with larger effects in programs of eight weeks or more (Zhang, Bi and Luo, Frontiers in Cardiovascular Medicine, 2025). On anxiety itself: resistance training reduced symptoms across 16 randomized trials (922 participants) with an overall effect of 0.31, and 0.50 in healthy adults (Gordon and colleagues, Sports Medicine, 2017); in a review of 15 studies, exercise beat a waiting list in people with clinical anxiety (SMD -0.41 in the five trials that could be pooled), and in the trials that compared intensities, harder sessions did more than easy ones (SMD -0.38) (Aylett, Small and Bower, BMC Health Services Research, 2018). Across 40 studies of people with chronic illness, sessions of at least 30 minutes worked best (Herring, O’Connor and Dishman, Archives of Internal Medicine, 2010). Those effect sizes sit in the same range as breathwork and meditation. The difference is that exercise also does everything else it does.

Mindfulness and meditation

A JAMA Internal Medicine review of 47 trials found moderate evidence that mindfulness programs reduce anxiety (effect size 0.38 at eight weeks) and depression (0.30), and “no evidence that meditation programs were better than any active treatment” (Goyal and colleagues, JAMA Internal Medicine, 2014). The strongest single trial came later: 276 adults with anxiety disorders were randomized to an 8-week mindfulness-based stress reduction course or escitalopram, a first-line SSRI. The course was non-inferior to the drug on the clinician severity scale, and 15 percent of the mindfulness group reported study-related adverse events against 79 percent on the medication (Hoge and colleagues, JAMA Psychiatry, 2023). Eight weeks of structured practice, not a two-minute hack.

Sleep, and what you drink

If your goal is the number on the wearable, this is where the big swings are. Across 11 randomized trials, one night of sleep deprivation lowered RMSSD (SMD -0.24) and sharply raised the LF/HF ratio (SMD 1.47) (Zhang and colleagues, Frontiers in Neurology, 2025). In 4,098 Finnish employees wearing HRV monitors over 12,411 nights, alcohol cut RMSSD in the first three hours of sleep by 2.0 ms after a low dose, 5.7 ms after a moderate dose and 12.9 ms after a high dose, with heart rate up 1.4, 4.0 and 8.7 beats per minute (Pietila and colleagues, JMIR Mental Health, 2018). A couple of drinks will move your early-night HRV more than any ear massage can. Our sleep guide covers the sleep side in detail.

Cold water: real reflex, not a treatment

Cold on the face triggers the diving reflex: trigeminal nerves in the face signal the brainstem, the vagus slows the heart, breathing pauses and peripheral vessels constrict. Cold applied to the face is the usual way to trigger the reflex, and doctors sometimes use it to interrupt a racing heart rhythm (StatPearls, Physiology, Diving Reflex). That is why a cold splash feels like it does something. It is a heart-rate reflex.

Whether it treats anxiety is a different question, and the answer so far is no. The one controlled study we found of cold stimulation and HRV used a lab thermode on the neck and cheek of 61 healthy adults, found a small HRV shift at the neck, and never measured anxiety or stress at all (Jungmann and colleagues, JMIR Formative Research, 2018). The one large cold-shower trial, 3,018 adults randomized to 30 to 90 seconds of cold at the end of a warm shower for 30 days, found 29 percent less self-reported sickness absence, and anxiety, measured as a secondary outcome, did not differ between groups (Buijze and colleagues, PLOS ONE, 2016). Take cold showers if you like them. Do not take them for your anxiety and expect the research to back you. Our cold showers guide goes deeper on what they do and do not do.

Mostly hype or untested

• Humming and singing. The study everyone cites found humming raised nitric oxide in the nose 15-fold in ten healthy subjects (Weitzberg and Lundberg, American Journal of Respiratory and Critical Care Medicine, 2002). That is a sinus finding; it says nothing about anxiety or the vagus. The choir study that showed singers’ hearts synchronizing attributes the effect to the song pacing everyone’s breathing to about six breaths a minute, and did not test wellbeing (Vickhoff and colleagues, Frontiers in Psychology, 2013). Humming and singing are plausible because they force a long, slow exhale. They are slow breathing with a soundtrack, not a separate mechanism.
• Gargling. Appears in swallowing rehabilitation after stroke. We found no trial of gargling for anxiety or HRV.
• Ear massage, foot massage, a book on your belly, the “half smile.” No controlled trials found. Harmless, free, and not worth paying anyone to teach you.
• Clip-on and handheld “vagus” devices. The ear-clip approach (transcutaneous auricular VNS) is the basis of most consumer gadgets. A living meta-analysis of 16 sham-controlled studies found no effect on vagally mediated HRV: g = 0.014, with Bayesian “strong evidence for the null hypothesis” (Wolf and colleagues, Psychophysiology, 2021). The best-known non-invasive vagus device with FDA clearance is a handheld neck stimulator cleared for cluster headache and for migraine, acute and preventive (electroCore, 2018; 2020). None of it is cleared for anxiety.
• “VNS raises GABA.” True, in a specific sense: the GABA findings come from surgically implanted stimulators in people with drug-resistant epilepsy, such as a study of GABA-A receptor density in 10 implanted patients (Marrosu and colleagues, Epilepsy Research, 2003). Every GABA finding in the vagus literature comes from a wire on the nerve in epilepsy patients. None of it comes from humming or an ear clip.
• “Vagus nerve reset” programs and supplements. No controlled trials found for any commercial program or supplement marketed this way.

Safety: the one thing not to do

Several posts suggest massaging the sides of the neck “near the carotid sinus” to stimulate the vagus. Carotid sinus massage is a cardiology test. The European syncope guideline says it should be done with continuous ECG and beat-to-beat blood pressure monitoring, because in susceptible people it causes the heart to pause for three seconds or more or blood pressure to drop by more than 50 mm Hg (ESC 2018 syncope guideline, ACC summary). Leave the sides of your neck alone.

What real vagus nerve dysfunction looks like

Feeling stressed is not a sign of a damaged vagus nerve. Actual damage, from surgery on the esophagus or stomach, from diabetes, or from a lesion, shows up as hoarseness, difficulty swallowing, slurred speech and the uvula pulling to one side (StatPearls), or as gastroparesis: feeling full soon after starting a meal, nausea, vomiting, bloating and upper abdominal pain, because “if the vagus nerve is damaged or stops working, the muscles of the stomach and small intestine do not work normally” (NIDDK, gastroparesis). Those are doctor problems, not breathing-app problems.

When anxiety needs a clinician

Breathing, training and sleep are a reasonable first move for everyday stress. They are not a treatment plan for an anxiety disorder. The National Institute of Mental Health describes generalized anxiety as worry that is hard to control on most days for at least six months and does not go away, and its advice is to talk to a health care provider (NIMH, Generalized Anxiety Disorder; NIMH, Anxiety Disorders). A concrete line: the GAD-7, a seven-question screen your doctor can give you in two minutes, uses a score of 10 or higher as its cut point, with 89 percent sensitivity and 82 percent specificity in the validation study (Spitzer and colleagues, Archives of Internal Medicine, 2006). Ten or more, talk to someone. If you are having thoughts of harming yourself, call or text 988, the Suicide and Crisis Lifeline, now.

How we use this

We are trainers, not therapists, and the evidence above is why our 90-day lifestyle coaching program is built on the things with the best data rather than on hacks: private training sessions two or three times a week at an intensity that does something, and a sleep plan that protects the nights; the slow-breathing practice above is a five-minute daily habit we can add alongside it. The 75-minute kickoff assessment comes first, before anyone prescribes anything. If you are already training with us, ask your trainer to add a five-minute breathing cool-down at the end of a session; it costs nothing and it is the one vagus technique that has earned its place.

Frequently asked questions

How long until slow breathing helps?

Within a session for how you feel, and within minutes for HRV: the 2022 meta-analysis found vagally mediated HRV rose during breathing and immediately after one session. For a change in your baseline mood, the Cell Reports Medicine trial used five minutes a day for one month. Give it a month of daily practice before judging.

Can my watch or ring measure my vagal tone?

It can estimate overnight HRV. A validation study of the Oura ring against a chest ECG in 35 healthy adults found nightly RMSSD correlated at r = 0.96 but read about 15 ms low, and the study covered sleep only, not daytime or stress (Cao and colleagues, Journal of Medical Internet Research, 2022). Use your own trend on one device. Never compare your number with someone else’s or across devices, and remember it is an index of the heart’s rhythm, not a reading of the nerve.

Is higher HRV always better?

No. HRV falls with age, the healthy range is wide (short-term RMSSD averaged 42 ms with a range of 19 to 75 across 21,438 healthy adults), and abnormally high HRV can come from heart conduction problems; the authors of the standard reference write that “higher HRV is not always better since pathological conditions can produce HRV” (Shaffer and Ginsberg, Frontiers in Public Health, 2017). A sudden spike is a reason to ask a doctor, not to celebrate.

Does the vagus nerve control inflammation?

There is a real “inflammatory reflex,” and the headline study is real science: an implanted vagus stimulator reduced TNF production and disease activity in a small open-label study of 17 people with rheumatoid arthritis (Koopman and colleagues, PNAS, 2016). That involved a surgically implanted electrode. It does not mean humming lowers your inflammation.

Is box breathing or 4-7-8 better than just breathing slowly?

We found no trial showing any pattern beats plain slow breathing for anxiety. In the one trial that compared patterns head to head, cyclic sighing (a double inhale, then one long exhale) was the pattern that separated from meditation on mood; box breathing was not, and all four groups improved. The common ingredient is a slower rate with a longer exhale. Pick the one you will do.

Should I try a vagus nerve stimulation device?

Implanted VNS is approved for drug-resistant epilepsy (1997) and treatment-resistant depression (2005), after multiple failed medications, and it is surgery (CMS coverage document). No vagus nerve device is cleared for anxiety, and the ear-clip approach did nothing measurable to vagal HRV in sham-controlled studies. Spend the money on a gym membership, a trainer or a therapist.