Stephen Porges

THE CLAIM

Stephen Porges made safety a physiological state you can measure

Stephen Porges moved the word safety out of the language of feeling and gave it an address in the brainstem. Born in 1945 in New Brunswick, New Jersey, Porges trained as a psychophysiologist and served as president of the Society for Psychophysiological Research from 1993 to 1994. He introduced Polyvagal Theory in his 1994 presidential address and published it the following year as Orienting in a Defensive World: Mammalian Modifications of Our Evolutionary Heritage. A Polyvagal Theory, in Psychophysiology, volume 32, issue 4, pages 301 to 318 (Porges 1995). The founding claim is anatomical before it is psychological. The vagus, the tenth cranial nerve, is not one undifferentiated calming line running from brain to gut. Its motor fibers leave the medulla from two separate source nuclei, and those two nuclei do different work.

Ask what that commits you to. If the vagus carries two motor systems rather than one, then calm and collapse are not two readings on a single dial. They are two circuits, and a body can be profoundly still in either. From that one distinction Porges built a phylogenetic hierarchy of autonomic response, a subcortical risk detector he named neuroception, and a face to heart circuit he called the Social Engagement System. The theory reshaped trauma care, pediatric psychology and psychotherapy, and it is under sustained scientific challenge. Both of those sentences are true at once. This page holds them together, because a library that reports only the flattering half of a story has stopped being a library.

THE PARADOX

The theory began as a contradiction on a hospital monitor

Polyvagal Theory started as an attempt to explain how the same nerve could protect a newborn and endanger one. Porges spent years measuring heart rate patterns in preterm infants, and two facts sat badly together. Respiratory sinus arrhythmia, the rise and fall of heart rate across the breath cycle, is vagally mediated and reads as a marker of health and resilience. Clinically dangerous bradycardia, the sudden profound slowing that threatens a preterm newborn, is also attributed to the vagus. Porges called this the vagal paradox, and he was still working on it in 2023, when Comprehensive Psychoneuroendocrinology published his paper The Vagal Paradox: A Polyvagal Solution in volume 16 (Porges 2023).

One clinical detail cracked it open. Infants at high risk reliably showed low amplitude RSA, a nearly constant beat to beat rate, before a bradycardic event. Not a strong rhythm collapsing. A missing rhythm, and then the crash. Ask what that sequence means. If the protective oscillation has to be absent for the dangerous slowing to appear, the two cannot be the same pathway running at different intensities. One pathway is withdrawing while another takes over. That is the whole theory compressed into a single observation, and the order matters. The observation came first. The evolutionary framework was built afterwards to account for it.

ANATOMY

There is one vagus nerve, and it has two motor origins

The phrase two vagus nerves is a teaching shorthand that has hardened into an error. Anatomically there is one paired cranial nerve. What is doubled is the origin of its motor fibers inside the medulla. The dorsal motor nucleus of the vagus sends unmyelinated fibers, weighted towards organs below the diaphragm, and it is shared with most vertebrates. The nucleus ambiguus sits more ventrally and sends myelinated fibers, including the fibers to the sinoatrial node that generate respiratory sinus arrhythmia. Porges named that ventral grouping the ventral vagal complex. Sensory traffic, which accounts for the larger share of the nerve, arrives at the nucleus of the solitary tract.

The real anatomy is more crowded than the popular version admits. Winfried Neuhuber and Hans-Rudolf Berthoud, writing in Biological Psychology, volume 174, 2022, point out that vagal afferents also target the spinal trigeminal and paratrigeminal nuclei, and that the solitary nucleus integrates somatic input alongside visceral input (Neuhuber and Berthoud 2022). Porges reports that the number of myelinated vagal fibers increases in a roughly linear fashion from 24 to 28 weeks gestation through to full term birth, with myelination continuing across the first months of life. Read that number slowly. The calming circuit is not finished at delivery. It is built on a schedule, in contact with a caregiver, inside a particular acoustic and social environment.

THE HIERARCHY

Three circuits, and the newest one is used first

Polyvagal Theory ranks autonomic responses by evolutionary age and predicts the order in which they are recruited. Porges describes three subsystems. Social communication depends on the myelinated vagus and fosters calm behavioral states. Mobilization depends on the sympathetic nervous system and produces fight and flight. Immobilization depends on the unmyelinated vagus and produces feigned death, vasovagal syncope and behavioral shutdown. In the account he published in the Cleveland Clinic Journal of Medicine, volume 76, supplement 2, pages S86 to S90, in 2009 (Porges 2009), the sequence is explicit. The newest circuit is used first, and if it fails to provide safety the older circuits are recruited in turn.

The frame is borrowed openly. Porges cites John Hughlings Jackson, who argued that higher neural circuits inhibit lower ones, and that when the higher are suddenly rendered functionless the lower rise in activity. He cites Walter Hess, who proposed more than fifty years earlier that the autonomic nervous system is an integrated system rather than a merely vegetative one. Ask what the ordering commits you to. If the newest circuit runs first, then a person who drops straight into shutdown has not chosen badly. The top of the stack failed to deliver safety, and the system did what its architecture says it will do. The clinical target moves off the behavior and onto the availability of the circuit above it.

NEUROCEPTION

The body files a verdict on risk before you know a verdict was needed

Neuroception is the name Porges gave to risk detection that runs without conscious awareness and without asking permission. He introduced the term in 2004, in Zero to Three, volume 24, issue 5, pages 19 to 24, and carried it into the peer reviewed literature three years later (Porges 2007). The word was coined to mark a distinction. Perception implies awareness. Neuroception does not. The nervous system evaluates sensory information from the environment and from the viscera continuously, sorts it into safe, dangerous or life threatening, and shifts physiological state accordingly, all of it beneath the level at which a person could report on the process.

Ask what follows if that is right. You cannot argue a nervous system into safety, because the argument arrives at a structure that was never consulted. What does move the verdict is the class of input the detector actually samples. The prosody of a voice. The mobility of a face. The rate and depth of somebody else's breathing. The predictability of what happens next. The quality of contact. Porges also allows the detector to be wrong, reading safe contexts as dangerous or dangerous contexts as safe. Neuroception is a construct rather than a measured structure. It names a process and organizes evidence about it. It does not point to a single site on a scan.

FACE AND HEART

The nerves that run the face and the nerve that slows the heart share a circuit

In mammals the muscles of the face and head are yoked to cardiac vagal regulation inside one integrated system. Porges calls it the Social Engagement System. Its substrate is the special visceral efferent pathways of five cranial nerves, the trigeminal, the facial, the glossopharyngeal, the vagus and the accessory, working together with the myelinated vagal fibers from nucleus ambiguus. Those pathways control eye gaze, facial expression, chewing, the middle ear muscles, vocalization and head turning. As Porges puts it, a face to heart connection evolved as the source nuclei of vagal pathways shifted ventrally in the brainstem.

The detail that makes the claim testable is the middle ear. The same neural pathways that raise the eyelids also tense the stapedius muscle, which stiffens the ossicular chain and biases hearing towards the frequency band of the human voice. Follow that through. A person in a defensive state carries a flatter face, a narrower vocal range and more difficulty picking speech out of background noise, and none of those are character traits. They are the readable surface of a circuit that has stepped back. Prosody runs in both directions. A voice with contour is an input to another nervous system, not only an output of your own.

THE BRAKE

Vagal regulation is a brake that is applied and released, not a sedative

Porges describes the myelinated vagus as an active brake on the heart, one whose rapid inhibition and disinhibition can mobilize or calm a person within a beat or two. That framing carries more weight than it looks. A brake is not a tranquilliser. The sinoatrial node, left alone, fires faster than a resting heart rate. Vagal outflow holds it down, and releasing that hold produces immediate arousal with no sympathetic recruitment at all. Porges measured the outflow through the amplitude of respiratory sinus arrhythmia, and he built instrumentation to do it. United States patent 4,510,944, granted in 1985, covers his method for quantifying rhythmic oscillations in aperiodic physiological response systems (Porges 1985).

Now force the implication. If regulation is a brake, then health is not maximum vagal output. It is the capacity to release cleanly under demand and to reapply cleanly afterwards. Amplitude, timing and recovery all carry information. A person locked into high vagal restraint is not regulated, only rigid. A person who cannot reapply the brake after exertion is not relaxed, only depleted. This is a systems reading of physiology. Oscillation, feedback, gain, and the ability of a control loop to move without losing its set point. Read that way, a heartbeat is a running record of how one nervous system negotiates with the world around it.

STILLNESS

Immobilization comes in two forms and fear is what separates them

Defensive immobilization is the oldest circuit in the stack and the most costly one for a mammal to use. Porges describes it as phylogenetically ancient, associated with reduced metabolic demand and a raised pain threshold. For a reptile with a wide tolerance for low oxygen, going still is an efficient defense. For an oxygen hungry mammal the same maneuver is dangerous, which is why the human versions show up on the record as vasovagal syncope, dissociation and behavioral shutdown rather than as a dependable survival strategy. This is also where the theory earned its clinical audience, because it gives a physiological account of freezing that does not read as weakness.

The second form is the one most summaries drop. Porges describes immobilization without fear, produced by co-opting the same structures to serve reproduction, nursing and pair bonding, and he notes that the ventrolateral periaqueductal gray is rich in receptors for oxytocin. Stillness in the arms of somebody safe runs on ancient hardware repurposed for intimacy. So the dorsal vagal pathway is not the villain of the story, whatever the popular retellings suggest. It runs digestion, restoration and rest. It becomes a problem only when it is recruited as a defense, and it is recruited as a defense only when the circuits above it have failed to deliver.

THE ARGUMENT

The theory is contested, and saying so is part of teaching it

Polyvagal Theory is influential and disputed, and a serious page reports both. Paul Grossman published a direct challenge in Biological Psychology, volume 180, 2023, article 108589, titled Fundamental Challenges and Likely Refutations of the Five Basic Premises of the Polyvagal Theory (Grossman 2023). Two of his objections carry most of the load. First, respiratory sinus arrhythmia is not exclusive to mammals, so treating mammalian cardiorespiratory coupling as a unique evolutionary signature is contested. Second, RSA is an approximate index of cardiac vagal control rather than a synonym for it, and a great deal of the theory rests on that equivalence. Neuhuber and Berthoud raised parallel anatomical questions in the same journal in 2022 (Neuhuber and Berthoud 2022). The exchange escalated in 2026, when Clinical Neuropsychiatry, volume 23, carried a multi author evaluation arguing the theory is untenable (Grossman 2026), alongside a published reply from Porges in the same volume (Porges 2026).

Ask what a clinician should do with an open argument. Not pick a side by temperament. Separate the layers. The clinical observations that generated the theory are not the part in dispute. Preterm infants with flat heart rate patterns are at higher risk. Cardiac vagal control is real and measurable. Cues of voice and face change autonomic state. People in defensive states behave and heal differently from people who are not. The evolutionary story that binds those observations into a single hierarchy is a model, and a model is held loosely by anyone who intends to keep learning. Porges himself frames the theory as a generator of testable hypotheses. Take him at his word and treat it that way.

PRACTICE

If safety is a state, it can be offered, and offering it is a skill

The practical yield of Polyvagal Theory is a shift in what a practitioner counts as the intervention. If neuroception samples voice, face, pace and predictability, then the room is not a neutral container around the treatment. The room is part of the treatment. Volume, tempo, vocal contour, announcing what will happen before it happens, the speed of a hand approaching, the amount of choice a person keeps: every one of those is an input to a risk detector that never switches off. Porges pushed this into a product as well as a principle, publishing the Safe and Sound Protocol in 2017, an acoustic intervention built on filtered music and protected by United States patent 10,029,068, granted in 2018 (Porges 2018).

Be accurate about the evidence here. The theoretical account is stronger than the outcome literature for any specific protocol built on it, and the honest position is that the intervention research is still developing. What survives that caution is the ordinary craft. A practitioner who lowers vocal pitch and slows tempo, who narrates before touching, who leaves a person an exit, is moving measurable physiology whether or not the polyvagal framework holds in every detail. The mechanism is arguable. The requirement to earn safety before asking a body to change is not.

PORGES AND THE MODEL

What Porges gave the Unified Model of Tone, and where the model parts from him

Porges gave the Unified Model of Tone a vocabulary rather than a mechanism. The model uses vagal tone constantly and does not adopt Polyvagal Theory. Both of those hold at once. His durable contribution is that autonomic state became something a clinician could name, measure and discuss with a patient. A hidden variable became a public one. The model's own claims stand on their own footing, and none of them depend on the evolutionary hierarchy under argument in Biological Psychology and Clinical Neuropsychiatry.

The clearest example of tone is the autonomic nervous system, which runs on two opposing tones: a sympathetic tone that accelerates the body and a parasympathetic, or vagal, tone that restrains it. The body is always setting the balance between the two. When sympathetic tone outweighs vagal tone, heart rate, blood pressure and blood glucose climb. When vagal tone outweighs sympathetic, they fall. Health is not a fixed setting of that balance. Health is the ability to move it wherever the moment demands.

Vagal tone is one of the named tones the model unifies. It sits beside muscle tone, vascular tone, cortical tone and fascial tone as readings of one system rather than a family of properties that happen to share a word. Immunity shows what the reading buys. The counterweight to inflammation is the inflammatory reflex, carried mainly by the vagus, which releases acetylcholine onto immune cells and restrains cytokine production. When vagal tone is high, inflammation is held in check. When it is diminished, inflammation runs unchecked. That work belongs to Kevin Tracey, who described the inflammatory reflex in Nature in 2002 (Tracey 2002). It is not Porges's, and the model cites it on its own terms.

The model reads autonomic dysregulation as upstream rather than downstream. Autonomic dysfunction can be documented before the clinical onset of a disease like rheumatoid arthritis. In essential hypertension, elevated sympathetic firing to the heart, kidneys and vasculature is present early, before the pressure has settled into its raised state. The test is prospective. The model predicts that a cohort phenotyped for autonomic regulation while still well will show a measurable loss of autonomic flexibility before clinical onset, in people whose conventional markers are still normal. Prospective data in which autonomic dysregulation reliably appeared only after diagnosis would confine the finding to the conditions where it has already been shown.

Where the model differs from Porges is scale, and that difference is the model's to own. His account stops, deliberately and correctly, at the autonomic circuits and the striated muscles of the face and head. He wrote about the heart, the middle ear, the larynx and the facial nerve. He did not write about fascia, about resting muscular length, or about the tension a hand meets in a paraspinal region, and he made no claim about manual care. Guarding is the same autonomic state read at a different scale. So is breath held high in the chest, a jaw that will not release, and a segment that resists at a load the person carried easily last week. That is what this library means by tone, the living state of the nervous system expressed as tissue tension at every scale from a heartbeat to a posture. Porges measured one register of it with real precision. Julian Thayer extended the reading upward into cortical control. The extension outward into tissue is the model's own claim. The measurements are theirs.

WHAT THE RECORD SHOWS

The record on Porges, the vagus and autonomic tone

• 1994. Porges introduced Polyvagal Theory in his presidential address to the Society for Psychophysiological Research and published it the following year in Psychophysiology, volume 32, issue 4, pages 301 to 318 (Porges 1995).
• One nerve, two nuclei. The vagus is a single paired tenth cranial nerve. Its motor fibers leave the medulla from two source nuclei: the nucleus ambiguus, which sends myelinated fibers, and the dorsal motor nucleus, which sends unmyelinated ones.
• 24 to 28 weeks. Porges reports that myelinated vagal fibers increase in a roughly linear fashion from 24 to 28 weeks gestation through to full term birth, with myelination continuing across the first months of life.
• 2004. Porges coined neuroception in Zero to Three, volume 24, issue 5, pages 19 to 24, and carried it into the peer reviewed literature in 2007 in Biological Psychology, volume 74, issue 2 (Porges 2007).
• 1985 and 2018. United States patent 4,510,944 (Porges 1985) covers his method for quantifying rhythmic oscillations in aperiodic physiological response systems, and patent 10,029,068 (Porges 2018) covers the Safe and Sound Protocol he published in 2017.
• 2002. Kevin Tracey described the inflammatory reflex in Nature (Tracey 2002): the vagus releases acetylcholine onto immune cells and restrains cytokine production. Inflammation is held in check when vagal tone is high and runs unchecked when it is diminished.
• 2022 to 2026. Neuhuber and Berthoud raised anatomical objections in Biological Psychology in 2022 (Neuhuber and Berthoud 2022), and Grossman challenged the five basic premises in the same journal in 2023 (Grossman 2023). Clinical Neuropsychiatry volume 23 carried a critique with a reply from Porges in 2026.