Most of what is sold as a somatic exercise is three things in a trench coat: slow movement, attention to physical sensation, and a longer exhale. Those three have reasonable evidence behind them. The theory usually bolted on top — polyvagal theory — does not, and its core premises have been formally challenged in the same journal that helped popularise them. The useful move is to keep the exercises and drop the explanation.
What does “somatic” actually mean here?
Soma is Greek for body. A somatic approach is one that works through bodily sensation rather than through discussing events — the target is what you notice happening in you, and movement or breath is the route there. That is a coherent category with real lineages behind it: Feldenkrais, Hanna Somatics, Rolfing, Somatic Experiencing, sensorimotor psychotherapy.
The word has since stretched to cover almost anything performed by a body. A sixty-second clip of someone shaking their arms is now “somatic release.” So is a hip stretch. So is lying down. The label tells you very little about what you are being given.
A workable test: does the instruction tell you what to notice, or only what to do? If it is only movement, it is exercise. If the instruction directs attention to sensation and asks you to stay with it, it is somatic in the sense that has any research attached to it. Nearly all of the viral versions are the first kind wearing the name of the second.
The polyvagal problem
Almost every somatic-exercise page opens with the same map: ventral vagal for safe and social, sympathetic for fight or flight, dorsal vagal for shutdown. It comes from Stephen Porges’s polyvagal theory, and it is presented as settled neuroscience. It is not.
Two specific problems, both from the physiology literature rather than from sceptics with no stake in it.
First, the measurement. Paul Grossman and Edwin Taylor, writing in Biological Psychology in 2007, showed that respiratory sinus arrhythmia is not a clean index of cardiac vagal tone. It varies with breathing rate and depth, so a person who slows their breathing produces a bigger number without necessarily changing anything underlying. Since “vagal tone” in popular somatic content usually means exactly this measure, the headline claim is shakier than it sounds.
Second, the theory itself. Grossman returned to it in Biological Psychology in 2023 with a paper titled “Fundamental challenges and likely refutations of the five basic premises of the polyvagal theory”, arguing that each of the five load-bearing claims is contradicted by comparative physiology — including the evolutionary story that the myelinated vagus is a mammalian innovation supporting social engagement.
What this changes in practice is smaller than it sounds, and worth being precise about. It does not mean the exercises are useless. It means that when a video tells you an exercise is “toning your ventral vagal complex,” you are being given a mechanism that has not been established. The same thing happens with sound frequencies, which I have written about in the binaural beats evidence and in what vagus nerve exercises can and cannot do.
Eight exercises, and what each one is actually doing
These are the ones that recur across the somatic-exercise genre. The third column is the honest version of the mechanism.
| Exercise | How | What is likely doing the work |
|---|---|---|
| Extended exhale | In for 4, out for 8, twelve rounds | Slow-paced breathing. Best-evidenced item on the list. |
| Physiological sigh | Two inhales through the nose, one long exhale through the mouth | Rapid carbon dioxide offload plus a long exhale. Fast-acting, short-lived. |
| Orienting | Turn the head slowly, name five things you can see | Moves attention outward and interrupts rumination. Simple and underrated. |
| Grounding through pressure | Press both feet into the floor, or push palms against a wall | Strong proprioceptive input gives attention something unambiguous to hold. |
| Self-hug or butterfly tap | Cross arms, alternate slow taps on each shoulder | Rhythmic bilateral touch. Reliably soothing for many people; the mechanism is not established. |
| Shaking or tremoring | Let knees bounce, allow the shake to spread, two to five minutes | Movement plus attention plus, usually, a longer exhale. The “releasing stored trauma” framing has no support. |
| Progressive muscle relaxation | Tense a muscle group hard for five seconds, release, notice the difference | The oldest and best-documented of the set. The contrast is the point, not the tensing. |
| Humming or a long “voo” | One low sustained note on each exhale | A long exhale you cannot rush. The vocal-fold vagal-stimulation claim is speculative. |
Count the third column and something becomes obvious: five of the eight contain an extended exhale, and two more contain sustained attention to sensation. The genre is less varied than its branding suggests. If you only ever did the first and third rows, you would have most of the available benefit.
On sequencing — if you are very activated, start with orienting or pressure rather than breath. Asking someone in a spike to control their breathing often makes them more aware of how hard it is, which adds to the spike. That is covered in more detail in getting out of fight or flight.
What the research actually supports
Slow-paced breathing. Sylvain Laborde and colleagues published a systematic review and meta-analysis in Neuroscience & Biobehavioral Reviews in 2022 on voluntary slow breathing and heart rate variability. Slow breathing increases HRV fairly consistently; the effect on heart rate itself is smaller and less reliable. This is the firmest ground in the whole field, and it is also the least interesting-sounding, which is presumably why it is not what gets marketed.
Somatic Experiencing. Danny Brom and colleagues ran a randomised controlled trial published in the Journal of Traumatic Stress in 2017 — 63 participants with PTSD, with symptom reductions against a waitlist control. A scoping review by Marie Kuhfuß and colleagues in the European Journal of Psychotraumatology in 2021 found a positive signal across a small literature, with small samples and few controlled trials.
That is genuinely promising, and it comes with a distinction that gets dropped constantly: both studied Somatic Experiencing delivered by a trained practitioner over multiple sessions. Neither studied following exercises from a video. Citing Brom to justify a shaking clip is a category error, and it happens on almost every page that cites him.
Progressive muscle relaxation. Dating to Edmund Jacobson in the 1930s and studied ever since, with a decent record for anxiety and sleep onset. Unfashionable, free, and better supported than most of what has replaced it.
Interoception. The most plausible shared mechanism is not vagal anything — it is learning to notice bodily sensation without being alarmed by it. That capacity is measurable, and it appears to be trainable. Which would explain why the specific exercise seems to matter less than whether you keep doing one.
What I actually kept
I tried most of the list. Two survived.
Orienting, when I catch myself having been at a screen long enough that my jaw is set and my shoulders are somewhere near my ears. Head turns slowly left, then right, and I name things — window, kettle, the chipped bit of the doorframe. It takes about twenty seconds and it works on the first try, which is unusual for anything in this category.
And a 4-in, 8-out count, twelve rounds, usually sitting in the car before going into something I am dreading. It took roughly three weeks before I stopped having to count deliberately.
The shaking one I dropped. It did produce a change in how I felt, but I could not tell it apart from the effect of having stood up and moved for four minutes, and the framing around it — that I was discharging something stored — made me pay attention to my body in a searching, slightly anxious way that was worse than not doing it. Your result may differ. That was mine, and I would rather say so than pad the list.
When this is the wrong tool
Turning attention toward the body is not neutral for everyone. For some people with a trauma history, interoceptive focus increases distress rather than reducing it — this is documented, it is not a sign you are doing it wrong, and it is a reason to work with someone rather than alone with a video.
Stop and speak to a professional rather than continuing on your own if: an exercise reliably brings on flashbacks, dissociation or panic; you feel further from your body afterwards rather than closer; symptoms are severe enough to interfere with work, sleep or relationships; or you are using these to manage something that has never been assessed. Persistent physical symptoms — racing heart, breathlessness, tremor, gut problems — deserve a medical opinion before they are filed under dysregulation. Thyroid disease, anaemia, arrhythmias and medication effects all produce the same feelings.
What the evidence doesn’t support
- That trauma is stored in tissue and can be shaken out. Trauma involves memory and physiology; there is no evidence of a discrete deposit being physically released.
- That polyvagal theory is established science. Its five basic premises were formally challenged in Biological Psychology in 2023.
- That a short video sequence delivers what Somatic Experiencing trials tested. Those trials used trained practitioners across multiple sessions.
- That HRV readings from a wearable tell you your vagal tone. HRV moves with breathing rate, posture, caffeine, alcohol and illness.
- That a nervous system can be reset in one session. Nothing in the literature describes a reset; what it describes is gradual change with repetition.
- That these substitute for treatment of PTSD or an anxiety disorder. They can sit alongside it.
Common questions
What are somatic exercises for nervous system regulation?
Movement, breath or touch practices where the instruction directs your attention to bodily sensation. In practice the effective ones come down to slowing the exhale, giving attention a physical anchor, and orienting outward. Everything else in the genre is a variation on those.
Do somatic exercises actually work?
The components do, modestly. Slow-paced breathing has meta-analytic support for raising HRV, and progressive muscle relaxation has a long record for anxiety and sleep. Practitioner-delivered Somatic Experiencing has early controlled evidence for PTSD. What does not hold up is the mechanism most often quoted at you, and the implied size of the effect.
What are the best nervous system reset exercises?
“Reset” is the wrong word — nothing here switches a state off and on. If you want the two with the best ratio of evidence to effort: a 4-in, 8-out breath for twelve rounds, and orienting by turning your head slowly and naming what you can see. Doing one of them daily beats rotating through eight.
How long before I notice a difference?
In the moment, orienting and the physiological sigh work within a minute or they do not work for you. The durable part — reacting slightly less, recovering slightly faster — took me about three weeks of near-daily practice to notice, and it showed up as other people commenting before I spotted it myself. What that change looks like is its own subject.
Can somatic exercises make you feel worse?
Yes, and this is under-said. Directing attention inward can amplify distress for some people, particularly with a trauma history. If an exercise consistently produces panic, flashbacks or a numb, far-away feeling, that is a reason to stop and get support, not to push through.
Is this the same as yoga or breathwork?
Overlapping, not identical. Both are covered separately — yoga for nervous system regulation and breathwork for nervous system regulation — and the broader picture is in nervous system regulation. If you are trying to work out whether any of this applies to you, start with the signs of dysregulation.
This article is general information, not medical advice. Nothing here is a substitute for care from a qualified professional, and no practice described should replace prescribed treatment. Sourcing and evidence standards are set out in the editorial policy.
