Dorsal vagal complex

In brief

Three nuclei that are read as one organ

The medulla oblongata is the lowest part of the brainstem, the stalk where the brain becomes the spinal cord. On its back surface, under the floor of the fourth ventricle - the fluid-filled cavity that sits between the brainstem and the cerebellum - three cell groups lie within a millimeter or two of one another. From the ventricular surface downward they are the area postrema, a small paired swelling of tissue; the nucleus of the solitary tract, a long column of cells wrapped around a bundle of incoming sensory fibers; and the dorsal motor nucleus of the vagus, a column of motor neurons beneath it. Written together, they are the dorsal vagal complex ^1,3.

The grouping is not merely a matter of proximity. It is a claim that these three do one job. Vagal sensory neurons, whose cell bodies sit in the nodose ganglia in the neck, carry information from the viscera into the brainstem and terminate in the nucleus of the solitary tract ^4,6. That nucleus projects, over short distances, onto the dorsal motor nucleus. The dorsal motor nucleus holds the preganglionic neurons of the parasympathetic supply to the gut - the first neuron of a two-neuron chain - and its axons leave the brainstem in the vagus nerve and travel back down to the abdomen ⁴. Organ to brainstem to organ, in two or three synapses - and possibly in a single central synapse, because dorsal motor nucleus neurons send dendrites up into the very parts of the solitary nucleus where gastric afferents terminate ³⁶. Rogers, McTigue and Hermann named the arrangement for what it does in 1995: vagovagal reflex control of digestion, sited in the dorsal vagal complex, and exposed to blood-borne as well as neural influence because the region lacks a complete blood-brain barrier ¹.

The area postrema is the odd member. It is a circumventricular organ, one of a small set of brain structures whose capillaries are fenestrated and which therefore lack the barrier that keeps blood-borne molecules out of nervous tissue ². It has no efferent axons running to the gut. Its job is chemical input: its neurons sense substances in the blood and project into the adjacent nucleus of the solitary tract, feeding the reflex arc from the side ³⁷. Reading it as part of the complex means reading the complex as a device with two kinds of input - nerve traffic from the organs, and chemistry from the bloodstream - converging on one output.

A 2014 review by Browning and Travagli set out the division of labor that makes this matter. The intestines have intrinsic nerve plexuses and can run themselves to a considerable degree; the stomach and esophagus are much more dependent on extrinsic input from parasympathetic and sympathetic pathways, and the parasympathetic part of that input, which both excites and inhibits gastric tone and motility, travels in the vagus ³. So the practical reach of the dorsal vagal complex is largest at the top of the gut and diminishes as you go down it.

One caution about the boundaries. Neuhuber and Berthoud pointed out in 2022 that the classical list of vagal brainstem nuclei - the dorsal motor nucleus, the nucleus ambiguus and the nucleus of the solitary tract - is incomplete in both directions: vagal afferents also target the spinal trigeminal and paratrigeminal nuclei, brainstem relays usually associated with sensation from the face, head and airways, and the nucleus of the solitary tract receives somatic afferents, from skin and muscle, as well as visceral ones ³⁸. The complex is a useful unit, not a sealed compartment. That paper carries a published corrigendum, issued in 2023, which readers should consult alongside it ³⁹.

The three members, briefly, and where to read about each

Each of the three nuclei has its own entry in this Atlas, and the detail belongs there. What follows is only enough to see how the division of labor works.

The area postrema is the chemical sensor. It is a circumventricular organ with fenestrated capillaries, it carries the receptors for a long list of circulating signals including GFRAL, the receptor for the stress hormone GDF15 (growth differentiation factor 15), and the receptor for GLP-1 (glucagon-like peptide-1, a gut hormone released after a meal), and it has no output to the gut: it projects into the nucleus of the solitary tract ^2,13,37. A single-nucleus atlas - a catalog built by reading which genes are active in thousands of individual cell nuclei - published online in December 2020 and in print in 2021 resolved a small number of neuron types here and found that activating particular excitatory populations induced nausea-related behaviors in mice, with one type mediating aversion imposed by several different poisons ³⁷. Read the area postrema page for the chemoreceptor trigger zone, the antiemetic drug classes and area postrema syndrome.

The nucleus of the solitary tract is the sensory relay and the integrator. Most vagal afferents from the viscera end here, although, as Neuhuber and Berthoud emphasize, some vagal afferents also reach the spinal trigeminal and paratrigeminal nuclei, and the nucleus receives somatic afferents as well, which is one reason it is not a simple visceral switchboard ^6,38. It also contains the swallowing pattern generator ⁴⁴ and the inhibitory interneuron population that sets the resting state of the gastric reflexes ⁵. It is the member of the complex that does the computing. Read its page for the subnuclei, taste, baroreflex and the rest.

The dorsal motor nucleus of the vagus is the output. Its cholinergic preganglionic neurons project through the vagus to enteric neurons of the gut wall rather than to muscle, and its activity determines gastric tone, motility and secretion ^4,5. It is, together with the anterior olfactory nucleus, one of the two sites at which the pathology of sporadic Parkinson disease first appears ²⁶. In mammals it contributes only a small part of the vagal supply to the heart, most of which comes from the nucleus ambiguus ^45,46, though in rats its neurons do regulate the excitability and the contractility of the ventricles ⁴⁷. Read its page for the cell columns, the Parkinson story and the cardiac question in detail.

One more nucleus belongs in the reader's mind even though it is not part of the complex: the nucleus ambiguus, which lies ventrolaterally, supplies the muscles of the larynx and pharynx, and in mammals provides the majority of cardiac vagal outflow ^45,46. Most of the confusion in the polyvagal dispute comes from claims that attach cardiac phenomena to the dorsal nucleus when the anatomy assigns them to the ventral one.

How the complex was assembled, and where the field reversed itself

The dorsal vagal complex is a twentieth-century synthesis rather than a discovery. Each of its three parts has its own history, told on its own page; what changed is that physiologists began to treat them as one circuit, and then as one chemosensory interface, and then as one target for drugs. The timeline below marks the steps for which a primary source is available, and marks explicitly the places where the field reversed a previous position.

Six of the entries are marked as reversals, in four strands, and they are worth stating plainly because they are the kind of correction this page exists to record. In 2007 Grossman and Taylor argued that respiratory sinus arrhythmia - the normal speeding of the heart with each breath in and slowing with each breath out, the measure on which a generation of autonomic psychology had been built - is not equivalent to cardiac vagal tone, the continuous background slowing of the heart by the vagus, and that the evolutionary story attached to it was wrong ⁴⁸. In 2023 Korsak and colleagues recorded directly from vagal preganglionic neurons during exercise in rats and found that their activity rises rather than falls, arguing against the standard notion of central vagal withdrawal during exercise ⁴⁹. Across 2022, 2023 and 2026 a series of papers first showed the anatomy the theory relies on to be incomplete ³⁸ and then argued that its anatomical and evolutionary premises are untenable ^27,28. And in 2026 a mouse study contradicted the 2024 finding that the weight loss and the nausea produced by the GLP-1 drugs run through separable populations of neurons ³⁴.

The term is younger than the anatomy. Among the earliest records in PubMed that use it are two 1980 papers from Swanson's laboratory on the projections of the hypothalamic paraventricular nucleus, one of which carries the phrase dorsal vagal complex in its title ^50,51. By 1985 Leslie could title a review of the region's neurochemistry with all three members named - nucleus of the tractus solitarius, area postrema and dorsal motor nucleus of the vagus ⁵². Rogers, McTigue and Hermann used the term in the first sentence of the abstract of their 1995 review without defining it, which is usually the sign that a term has settled ¹. PubMed's abstract coverage before the mid-1970s is thin, so older uses may exist.

The leakiness gradient, and the glia that stand in for the barrier

Almost everywhere in the brain, capillaries are sealed. The endothelial cells lining them are stitched together by tight junctions, so that what circulates in the blood does not simply diffuse into nervous tissue. That seal is the blood-brain barrier. The area postrema is one of the handful of places where it is deliberately incomplete, and the practical meaning of the dorsal vagal complex as a unit is that a hormone in the blood can reach a reflex circuit that controls the stomach without ever crossing an intact barrier ^1,2.

The opening is not a hole. Morita and colleagues infused tracers of different sizes into adult mice and watched where they stopped. Fluorescein isothiocyanate, a molecule of 389 daltons, and a dextran - a sugar polymer used as a size-graded tracer - of 3,000 daltons diffused out of the vessels into the tissue; a dextran of 10,000 daltons stayed in the perivascular space, the thin space around the vessel wall, between its inner and outer basement membranes ². Permeability to the smallest tracer was higher at the central subdivisions of these organs than at their distal subdivisions, and the tight-junction proteins occludin, claudin-5 and zonula occludens-1 were undetectable centrally but detectable distally ². So the barrier is graded across a few hundred micrometers rather than present or absent.

The second finding of that study matters more for the complex as a whole. Neither of the tracers that escaped the vessels diffused beyond a dense network of glial fibrillary acidic protein-positive astrocytes and tanycytes at the organ's edge, and electron microscopy showed capillaries surrounded by numerous layers of astrocyte processes and dendrites ². Occludin and zonula occludens-1 were found as puncta on these glial cells themselves. The authors proposed that the outer basement membrane and the dense astrocyte and tanycyte connection are possible alternative mechanisms for a diffusion barrier, in place of the classical endothelial barrier ². In plain terms: the gate may lie not in the vessel wall but in the glia a few cell-widths away.

Tanycytes are elongated glial cells with a cell body on a ventricular surface and a long unbranched process reaching into the tissue; they are best known from the hypothalamus, where they gate access between the third ventricle and the brain. Tanycyte-like cells of the same general shape line the interface between the area postrema and the nucleus of the solitary tract. Dallaporta and colleagues found leptin receptors on a subpopulation of glial fibrillary acidic protein-positive cells at exactly that border in rats, with unbranched processes radiating rostro-caudally from the fourth ventricle wall, and used a fluorescent tracer injected intravenously and into the ventricle to argue that these cells may constitute a diffusion barrier regulating what enters the nucleus of the solitary tract ¹⁰. They also found that acute or chronic food restriction changed the expression of the short leptin receptor isoforms in the complex, and hypothesized that these glial cells take part in transporting leptin into the brainstem ¹⁰.

Guillebaud and colleagues then showed that these cells are not only a barrier but a source of signal. Diazepam-binding inhibitor, the precursor of a family of peptides called endozepines, is strongly expressed by the ependymal cells lining the fourth ventricle, by tanycyte-like cells and by astrocytes near the area postrema and nucleus of the solitary tract interface in rats ⁴⁴. Electron microscopy placed these peptide-containing processes in close apposition to dendrites and axon terminals. Injecting either of two fragments of it - an eighteen-amino-acid peptide called octadecaneuropeptide, or an eight-amino-acid piece of its tail - into the fourth ventricle triggered c-Fos expression - a marker of recent neuronal activation - throughout the dorsal vagal complex and strongly reduced food intake, and microinjecting it into the swallowing pattern generator within the nucleus of the solitary tract inhibited the swallowing reflex ⁴⁴.

The gradient also explains a structural asymmetry that is easy to miss. The area postrema is outside the barrier; the nucleus of the solitary tract lies next to it, behind the glial border described above ^2,10; and the dorsal motor nucleus lies deeper still, although in rats a few of its dendrites reach into the area postrema and the lining of the fourth ventricle ³⁶. A plausible consequence, not directly tested in the studies cited here, is that a drug or hormone in the blood has graded access to the complex, strongest at the ventricular surface and falling off with depth.

What the complex actually runs: accommodation, acid, motility, emptying

The reason physiologists care about the dorsal vagal complex is that it runs the stomach. The stomach has an unusual mechanical problem: it must expand to receive a meal without the pressure inside it rising, because a rise in pressure would push the contents back up the esophagus. That expansion is called gastric accommodation, and it is not passive stretch. It is an active relaxation, driven from the brainstem, triggered by the arrival of food and by distension itself ^3,5.

The output arm does not work the way a motor nerve to a muscle works. Dorsal motor nucleus neurons are cholinergic, meaning they release acetylcholine, but they do not contact gastric muscle cells. They end on enteric neurons in the myenteric plexus, the nerve net between the two muscle layers of the gut wall, and those enteric neurons act both as muscle motor neurons and as function-specific pattern generators that can run particular motility and secretion programs ⁴. The brainstem therefore selects a program rather than pulling a lever. Some of those enteric neurons are excitatory and cholinergic; others are inhibitory and nitrergic, signaling with nitric oxide, and it is the latter that produce relaxation ^4,5.

Gillis and colleagues spent four decades mapping which brainstem connection serves which part of the stomach, and their 2022 review sets out the result. Reflex control of the fundus, the dome at the top of the stomach that does the accommodating, requires a noradrenergic connection from the nucleus of the solitary tract to the dorsal motor nucleus and also requires second-order nitrergic neurons within the nucleus of the solitary tract. Reflex control of the antrum, the muscular lower part that grinds and pumps, requires a GABAergic connection instead ⁵. Under most conditions a single efferent pathway out of the dorsal motor nucleus suffices; the accommodation reflex is the exception, where dual effectors controlling cholinergic and non-adrenergic, non-cholinergic input to the stomach appear to be involved ⁵.

The resting state is also actively set. Somatostatin-expressing GABA interneurons within the nucleus of the solitary tract determine basal gastric tone and phasic contractions, holding the system down; for a vagovagal reflex to become operational, an endogenous opioid is released in the nucleus of the solitary tract and the activity of those interneurons is suppressed ⁵. This is the sense in which the gastric reflexes have a brake: inhibitory interneurons sitting between the afferent input and the motor nucleus, released when a reflex is needed ⁵. It is a different thing from the vagal brake of the psychological literature, a term Porges and colleagues introduced in 1996 for the moment-to-moment withdrawal and reinstatement of cardiac vagal tone, measured as respiratory sinus arrhythmia, in a study of 24 infants ⁵⁵. Tonic vagal restraint of the heart is real - vagal tone is the major determinant of heart rate across vertebrates ⁴⁵ - but in mammals it is exerted mainly from the nucleus ambiguus rather than from the dorsal motor nucleus ^45,46, so the cardiac vagal brake is not a reflex of this complex.

Acid is the oldest demonstration. Early in the twentieth century Pavlov showed, in dogs with a chronic esophagostomy - a surgically diverted esophagus - that the sight, smell or taste of food produces vagally dependent gastric acid secretion, which established the cephalic phase of digestion ³⁰. In humans, sham feeding - chewing and spitting out food so that nothing reaches the stomach - increases bicarbonate secretion in both the stomach and the duodenum, and the gastric component is mediated by cholinergic vagal fibers while the duodenal component appears not to be ³¹. The brainstem mechanism was worked out in rats: thyrotropin-releasing hormone, a three-amino-acid peptide, reaches the dorsal motor nucleus from raphe neurons, the serotonin-rich cells of the brainstem midline, excites its firing through the receptor subtype TRH-R1, and drives vagal output to acid-secreting, pepsin-secreting and ghrelin-secreting cells and to enteric neurons; blocking the peptide or its receptor in the brainstem abolishes the acid response to sham feeding ³⁰. A 2026 review extends the same logic to glucose, arguing that taste-driven cephalic phase responses are assembled across the rostral nucleus of the solitary tract, the dorsal vagal complex and vagal parasympathetic output, and can blunt the post-meal rise in blood sugar before any nutrient is absorbed ⁶⁴.

The complex also controls the valve above the stomach. In decerebrate, unanesthetized ferrets - the forebrain surgically disconnected, so that the animal feels nothing while the brainstem still works - inflating the stomach with 30 mL relaxed the lower esophageal sphincter by 70 percent from baseline, plus or minus 8 percent, a vagovagal reflex that can result in gastroesophageal reflux ⁵⁷. Delta-9-tetrahydrocannabinol, given intravenously at 0.2 mg per kg or applied directly to the dorsal hindbrain surface at 0.002 mg, blunted that relaxation, and the effect of the intravenous dose was reversed by a selective cannabinoid-1 receptor antagonist ⁵⁷. Cannabinoid-1 receptor staining was present in cell bodies in the area postrema, nucleus of the solitary tract and nodose ganglion, with dense terminal-like staining in the nucleus of the solitary tract and dorsal motor nucleus that survived both nodose ganglionectomy and vagotomy, so the receptors are on intrinsic brainstem elements rather than on vagal fibers ⁵⁷. This is ferret physiology, not human, but it is a clean demonstration that a drug acting inside the complex can retune a reflex rather than block it.

The complex also sets the pace of gastric emptying, and a 2026 mouse study shows how stress reaches it. Acute stressors rapidly activated neurons of the hypothalamic paraventricular nucleus that carry the neuropeptide Y1 receptor; those neurons synapse directly on cholinergic neurons of the dorsal vagal complex; activating the pathway suppressed feeding and disordered gut motility, and inhibiting it reversed stress-induced loss of appetite and rescued stress-induced slowing of gastric emptying ⁶³. It is a mouse circuit, and it is the closest thing in the physiological literature to a dorsal vagal stress response - one that acts on the gut, not on the heart.

Finally, the afferent side is wider than the stomach. Tracing studies place vagal sensory endings in the esophagus, lower airways, heart, aorta, the whole gastrointestinal tract, liver, portal vein, biliary system and pancreas - but not the spleen ⁶. A 2026 anterograde tracing study from the nodose ganglion found a dense vagal sensory innervation of the mouse gall bladder, with axons entering by the cystic duct and dividing copiously in the wall, including laminar endings within gall bladder ganglia ⁷. The complex is the destination for all of it.

What it reads out of the blood: leptin, insulin, GLP-1, GDF15, cytokines

The second input to the complex is chemical. Because part of it lies outside the blood-brain barrier and the rest lies just behind a glial gate, circulating hormones can act on it directly. The list below is not exhaustive and the quality of evidence differs sharply from one entry to the next.

Leptin is the hormone released by fat tissue in proportion to fat mass. Leptin receptors are present on neurons of the dorsal vagal complex and, unusually, on the glial cells at the area postrema and nucleus of the solitary tract border, where they may serve transport rather than signaling ¹⁰. Functionally, leptin delivered into the hindbrain of rats raised brain-derived neurotrophic factor content in the dorsal vagal complex, and the intake-suppressing effect of leptin injected into the medial nucleus of the solitary tract was attenuated by a selective TrkB receptor antagonist - a drug blocking TrkB, the receptor for brain-derived neurotrophic factor - given at the same site - so part of leptin's action here runs through a growth-factor pathway rather than directly ¹¹. A harmonized rodent atlas published in 2025 described a leptin receptor-expressing neuron population in the rat area postrema marked by a progenitor cell marker, a population that had not previously been separated out ⁸.

Insulin is sensed here too, and the 2025 finding is that the sensing cell is not a neuron. In male rats, insulin receptors on astrocytes of the nucleus of the solitary tract were required for the brainstem's control of glucose production by the liver. Insulin evoked release of endozepines from primary astrocytes; infusing endozepines directly into the nucleus of the solitary tract reproduced the effect of insulin; blocking the benzodiazepine binding site of the GABA-A receptor prevented both; and endozepine administration bypassed the insulin resistance produced in this nucleus by a high-fat diet ¹². That is a complete mechanistic chain, but in one sex of one species, and it has not been tested in people.

GLP-1, glucagon-like peptide-1, is an incretin: a gut hormone released after eating that increases insulin secretion and reduces appetite. Its receptor is expressed on neurons in both the area postrema and the nucleus of the solitary tract, and those neurons are the ones that matter for the obesity drugs built on this hormone. Systematic investigation across the drug-accessible GLP-1 receptor populations found that only hindbrain neurons were required for the efficacy of GLP-1-based obesity drugs in mice ¹³.

GDF15, growth differentiation factor 15, is a stress-induced hormone that acts on a receptor called GFRAL, found almost nowhere in the body except the area postrema. In pregnancy, mass spectrometry - the precise weighing of molecules - of a naturally occurring variant of the hormone that tags its source showed that the vast majority of GDF15 in maternal plasma comes from the feto-placental unit; higher maternal levels were associated with vomiting and with hyperemesis gravidarum, the severe and persistent vomiting of pregnancy; and women with beta-thalassemia, an inherited anemia in whom GDF15 is chronically high, reported very low levels of nausea and vomiting in pregnancy ¹⁷. A 2026 multi-ancestry genome-wide association study of 10,974 cases and 461,461 controls identified ten loci, that is ten regions of the genome: four already known - GDF15, IGFBP7 (insulin-like growth factor binding protein 7), PGR (the progesterone receptor gene) and GFRAL - and six new, with opposing effects for GDF15 between maternal and fetal genotype ³⁵. Hyperemesis gravidarum has its own page; what belongs here is that the receptor sits in this complex.

The sickness signals of infection and chemotherapy also arrive here, and the 2026 evidence complicates the textbook account. In mice, deleting the 5-HT3 receptor - a serotonin receptor, the target of ondansetron, and the standard explanation for chemotherapy-induced vomiting - did not prevent cisplatin, a platinum chemotherapy drug, from producing the full malaise-like state: reduced food intake, weight loss, reduced spontaneous activity, pica and conditioned taste aversion all appeared to a similar extent in knockout and wild-type animals ¹⁸. In both groups, c-Fos-positive and c-Fos and GFRAL double-positive cells increased in the area postrema and nucleus of the solitary tract, and plasma GDF15 stayed elevated for at least a day ¹⁸. The authors conclude that GDF15-GFRAL signaling may drive cisplatin malaise independently of the 5-HT3 receptor. This is a mouse study, and malaise in a mouse is inferred from behavior, not reported.

Cytokines complete the list. The inflammatory reflex describes a circuit in which the vagus carries information about peripheral immune activity to the brainstem and, through efferent vagal and splenic nerve traffic, acetylcholine from choline acetyltransferase-expressing T cells downregulates macrophages bearing alpha-7 nicotinic receptors ⁶⁵. A 2026 study using cecal ligation and puncture, a rodent sepsis model, dissected the brainstem end: chemogenetically activating dorsal vagal complex neurons - switching them on with an engineered receptor and a designer drug - worsened sickness behavior and produced lasting anxiety-related behavior, while activating the upstream nodose neurons did not, and transient inhibition at either site had no effect; cutting the vagus below the diaphragm reduced acute brain activation and partly attenuated anxiety in survivors ¹⁹. A separate guinea-pig study gave fine particulate matter into the nose twice daily for 28 days and measured cytokines inside the complex: interleukin-1 beta was 132.7 plus or minus 17.6 picograms per milliliter in exposed animals and 105.3 plus or minus 14.3 in exposed animals also given melatonin, tumor necrosis factor alpha was 143.8 plus or minus 30.4 against 113.0 plus or minus 23.5, activated microglia numbered 54.6 against 25.1, and coughs to a citric acid challenge 29.1 against 18.8; the abstract does not report a brainstem cytokine level for the unexposed animals, so the size of the exposure effect itself is not stated ²⁰. These are animal models of different diseases; the common thread is that this piece of brainstem reads inflammation.

The glial turn, 2020 to 2026

For most of the twentieth century the dorsal vagal complex was described as a wiring diagram of neurons, with glia as support. That has changed, and the change is recent enough that most textbooks have not caught up.

The first clear demonstration came in a paper published online at the end of 2019 and in print in 2020. Mice given high-fat chow for twelve hours during their active phase showed a 65 percent increase in the number of glial fibrillary acidic protein-immunoreactive cells adjacent to the area postrema, along with greater morphological complexity, compared with mice on control chow ⁴¹. The authors then used a chemogenetic approach - delivering an engineered receptor to astrocytes by viral vector and activating it with a designer drug - and found that switching dorsal vagal complex astrocytes on reduced dark-phase feeding by 84 percent at four hours and cut refeeding after an overnight fast by 71 percent, without affecting locomotion, and with c-Fos appearing in downstream neuronal feeding circuits including the parabrachial nucleus, a relay in the pons that carries visceral and aversive signals forward ⁴¹. Astrocytes here respond to nutritional excess and, when driven, stop the animal eating.

A 2026 study extended this to a physiological rhythm. Food intake in female rats falls during the high-estrogen phase of the estrous cycle. Astrocyte staining and morphological complexity in the dorsal vagal complex rise in the same phase; chemogenetic inhibition of these astrocytes eliminated the cycle-linked oscillation in food intake altogether, while chemogenetic activation reduced food intake regardless of cycle stage ⁴². Recordings from dorsal motor nucleus neurons tied the effect to NMDA receptor activation, and blocking either estrogen receptors or the astrocytes prevented it ⁴². This is a tidy result, but it is rat physiology and the manipulations are artificial: chemogenetic activation is not a dose of anything a body makes.

Not every glial intervention changes the animal, and this is worth stating because selective reporting in this field would be easy. Connexin 43 is a gap-junction protein expressed at remarkably high levels by astrocytes in this region. Knocking it down with RNA interference changed astrocyte and microglial morphology and phenotype, producing denser astrocytic tiling and hyper-ramified microglia, but produced no significant change in metabolic phenotype under either standard feeding or a high-fat, high-sugar diet ⁴³. The authors read the negative result as evidence of a buffering system that preserves autonomic and metabolic output when one pathway fails - a reasonable interpretation, and also the interpretation that a null result invites, which is a reason to hold it loosely.

Diet changes the cellular environment more broadly. In eight-week-old male mice, one week of a high-fat diet produced a stronger response in neural stem and progenitor cells of the area postrema and the central canal than four weeks did; the diet increased astrocyte density in the nucleus of the solitary tract and the dorsal motor nucleus and increased microglial and macrophage density in the area postrema and solitary nucleus, and the longer exposure induced mild inflammation in the medulla ⁶⁶. The area postrema is one of the few sites in the adult brain where neural stem cells persist, so the complex can in principle remodel itself.

Astrocytes in this region also respond to the same gut hormones the neurons do. A 2026 study built a slice culture in which the jugular and nodose ganglia remained attached to a brainstem slice through intact vagal rootlets, and imaged calcium in identified cell types. Cholecystokinin, a gut hormone released in response to fat and protein, directly activated a subpopulation of neurons in the dorsal and ventrolateral medulla without requiring action potentials, and both cholecystokinin and substance P, a neuropeptide involved in nausea and pain, triggered robust calcium waves in brainstem astrocytes, while muscarine, a drug that mimics acetylcholine at one class of its receptors, and thyrotropin-releasing hormone activated neurons without recruiting the astrocytic network ⁶⁷. Selectivity of that kind is what separates a signaling system from a general stress response.

The most translational glial result came in 2025. Glia of the dorsal vagal complex make octadecaneuropeptide, an appetite-suppressing fragment of a protein called diazepam-binding inhibitor. In rats, octadecaneuropeptide targeted to the hindbrain caused weight loss, counteracted the feeding response to glucose deprivation and improved glucose clearance; blocking central octadecaneuropeptide signaling blunted the appetite-suppressing response to GLP-1 receptor agonists; and a derivative given by injection improved insulin sensitivity in obese mice and caused weight loss without pica in rats or vomiting in the musk shrew, a mammal that can vomit, and without any change in heart rate, core temperature or physical activity ⁶². A companion study found that female rats carry more diazepam-binding inhibitor protein throughout the complex and are more sensitive to the peptide than males ⁶⁸. Microglia join in too: in mice, a single injection of the cytokine tumor necrosis factor alpha into the abdomen activated neurons lateral to the area postrema, and suppressing microglia with minocycline, an antibiotic that also damps microglial activity, given into the fourth ventricle reduced that activation ⁶⁹. All of this is animal work; the gliopeptide has not been given to a person.

The modeling literature has begun to follow. A 2025 computational framework coupled a quadripartite synapse - excitatory neurons, inhibitory neurons, astrocytes and vascular smooth muscle cells - to a three-dimensional microcirculatory network in the dorsal vagal complex, and reproduced the feedback in which transmitter release triggers astrocytic responses that alter vessel radius and so local oxygen delivery ⁷⁰. It is a simulation, generating no new measurements, and should be read as a hypothesis-organizing tool rather than evidence.

The complex as a drug target, 2024 to 2026

The commercial interest in this piece of brainstem is new and intense. The GLP-1 receptor agonists reduce body weight substantially and cause nausea in a large minority of patients, and in mice both effects depend largely on hindbrain neurons of the dorsal vagal complex ^13,14,34; in people that has not been shown directly. In the STEP 1 trial (Semaglutide Treatment Effect in People with obesity), 1,961 adults with overweight or obesity were randomized two to one to once-weekly semaglutide 2.4 mg or placebo for 68 weeks; mean weight change was minus 14.9 percent against minus 2.4 percent, a treatment difference of 12.4 percentage points with a 95 percent confidence interval - the range within which the true difference very probably lies - of 11.5 to 13.4; nausea and diarrhea were the most common adverse events, described as typically transient and mild to moderate, and 4.5 percent of the semaglutide group discontinued for gastrointestinal events against 0.8 percent on placebo ³³. The trial was funded by Novo Nordisk and three of the listed authors were Novo Nordisk employees ³³.

The central question the preclinical work has been asking is whether weight loss and nausea can be separated. In 2024 Huang and colleagues imaged hindbrain GLP-1 receptor neurons in behaving mice with two-photon microscopy and found that most individual neurons were tuned either to nutritive or to aversive stimuli but not to both. Area postrema GLP-1 receptor neurons were broadly responsive; nucleus of the solitary tract GLP-1 receptor neurons were biased toward nutritive stimuli. Activating the nucleus of the solitary tract population produced satiety without aversion; activating the area postrema population produced strong aversion together with reduced food intake; and the two projected to different downstream regions ¹³. Critically, GLP-1 receptor agonists still reduced food intake when the aversion pathway was inhibited ¹³. That separation has since been contested. In 2026 Yacawych and colleagues silenced GLP-1 receptor neurons in either the nucleus of the solitary tract or the area postrema of mice, or restored the receptor in only one of those two places on an otherwise receptor-deficient background. The nucleus of the solitary tract neurons contributed to the normal restraint of food intake and body weight but were not needed for weight loss on a GLP-1 receptor agonist; the area postrema neurons played no detectable role in ordinary feeding but mediated both the weight-lowering and the aversive effects of the drugs, which the authors say prevents the separation of nausea from weight loss at the level of the circuit ³⁴. Several co-authors are Novo Nordisk employees or work at the Novo Nordisk Foundation Center for Basic Metabolic Research ³⁴. Both studies are in mice and use different manipulations - activation and inhibition in one, silencing and receptor restoration in the other - and the disagreement is unresolved.

In 2025 a different group identified the cells that carry semaglutide's effect. Many semaglutide-activated neurons in the area postrema and nucleus of the solitary tract express Adcyap1, the gene for pituitary adenylate cyclase-activating polypeptide. Reactivating semaglutide-responsive dorsal vagal complex neurons reproduced the drug's reduction of food intake and body weight, its shift toward fat utilization, and conditioned taste aversion. Ablating the Adcyap1 neurons largely reversed semaglutide's effects on energy balance, acutely in lean mice and in subchronically treated obese mice, and the loss was mostly of fat rather than lean mass, with only a modest effect on conditioned taste aversion ¹⁴.

Two further incretin systems converge on the same region. GIP, glucose-dependent insulinotropic polypeptide, the second incretin, is pharmacologically strange because both agonists and antagonists of its receptor reduce body weight. Single-nucleus RNA sequencing of the dorsal vagal complex showed opposing transcriptional effects of GIP receptor agonism and antagonism, with antagonism, not agonism, closely resembling GLP-1 receptor signaling, and both GIP receptor antagonism and GLP-1 receptor agonism regulated genes implicated in synaptic plasticity ¹⁵. That study includes Eli Lilly co-authors ¹⁵. A 2026 study then showed that GIP receptors in the dorsal vagal complex were required for the acute appetite-suppressing effect of a GIP receptor agonist but not for its anti-aversive effect, which depended instead on neurons of the parabrachial nucleus that make calcitonin gene-related peptide, a marker of aversion-signaling cells: GIP receptor agonism blunted the aversive effect of interleukin-1 beta while enhancing its appetite-suppressing effect ¹⁶.

Amylin is the third. A 2026 cross-species atlas profiled more than 530,000 cells of the caudal brainstem in rat, mouse and macaque, resolved 80 neuronal populations, and mapped their spatial distribution in the rat dorsal vagal complex. The amylin analog cagrilintide acted on two conserved calcitonin receptor-expressing populations. Acute treatment altered gene expression in area postrema neurons expressing the calcitonin receptor and Ramp3, the gene for receptor activity-modifying protein 3, which pairs with the calcitonin receptor to make an amylin receptor, but chemogenetically activating those neurons in rats did not change long-term food intake or body weight. Long-term cagrilintide instead upregulated prolactin-releasing hormone in nucleus of the solitary tract cells expressing the calcitonin receptor and Prlh, the prolactin-releasing hormone gene, a population conserved in rodents, macaques and humans, and knocking down that peptide in the dorsal vagal complex abolished the effects of cagrilintide but not those of semaglutide in rats ⁹. The authors include Novo Nordisk employees and the study concerns a Novo Nordisk compound ⁹.

Two 2025 mouse studies sharpen the anatomy. Fluorescently labeled tirzepatide, which activates both the GLP-1 and the GIP receptor, entered the brain mainly at the circumventricular organs, and both receptors in the dorsal vagal complex were needed for the effects of systemically given agonists; GIP receptor agonism alone directly activated a distinct population of GABA-releasing neurons in the area postrema, while the full synergy of the two drugs on body weight required neurons of the central amygdala that make the peptide neurotensin downstream ²². Several authors were Eli Lilly employees, and Lilly makes tirzepatide ²². For cagrilintide, weight loss in obese mice depended on amylin receptors built from the calcitonin receptor together with the receptor activity-modifying proteins RAMP1 and RAMP3; without those proteins, activation of the area postrema by the drug fell by 57 percent. That work was funded by a Novo Nordisk consortium grant and included Novo Nordisk employees ²³.

Two more 2026 papers widen the frame. One reports that in a mouse model of lipopolysaccharide-induced neuroinflammation, semaglutide prevented brain infiltration of neutrophils and suppressed inflammatory transcriptional signatures in microglia, endothelial cells and a subset of pericytes, and identifies a subset of GLP-1 receptor-expressing dorsal vagal complex neurons that regulate anti-inflammatory genes on treatment; the work was done largely by Novo Nordisk researchers ²¹. The other sequenced 75,787 nuclei from the arcuate nucleus and the ventromedial hypothalamus, two appetite-regulating regions of the hypothalamus, and from the dorsal vagal complex of control-diet and high-fat-diet mice after intravenous glucose or saline, and found that the transcriptional response to high blood sugar, conserved across regions in astrocytes and oligodendrocytes, the glia that make myelin, was largely absent or reversed after a high-fat diet ⁷¹. Both are mouse work by groups with industry ties, and both should be read as hypothesis-generating.

A preprint posted in 2026, not yet peer reviewed, reports optical recordings in behaving mice showing that GFRAL neurons - the GDF15 sensors, usually described as nausea neurons - are activated by eating food rich in fat, that this activation is required for fat satiation and does not involve GDF15 or the usual gut-brain pathways, and that GIP receptor neurons which inhibit them are selectively activated by sugar ⁷². If it survives review, it would mean that the GFRAL and GIP receptor neurons of the area postrema, which the authors describe as key targets of tirzepatide, have ordinary nutritional jobs - sensing dietary fat and sugar respectively - quite separate from sickness. It should be treated as provisional until it does.

Two things follow for a reader trying to judge this literature. First, essentially all of the circuit evidence is mouse and rat, with a single macaque dataset; the human contribution is a set of randomized trials that measure weight and side effects without being able to see the brainstem. Second, a high proportion of the key papers were produced by, funded by, or include employees of the companies that sell these drugs, which does not make the results wrong but does mean that negative or inconvenient findings are less likely to be pursued with the same intensity.

Dorsal vagal complex, dorsal vagus complex, and what the index knows

Both spellings appear in print, and they mean the same thing. As of 19 October 2026 a PubMed search for the exact phrase dorsal vagal complex returns about 740 records; the exact phrase dorsal vagus complex returns 4, only two of which carry it in the title ²⁵. One of those two is a 2014 rat tracing study of whether the substantia nigra, the midbrain nucleus whose dopamine neurons are lost in Parkinson disease, projects directly to the complex - it does not, and the authors propose an indirect route through the hypothalamus or locus coeruleus - which uses dorsal vagus complex in its title and defines it as the dorsal motor nucleus of the vagus together with the nucleus tractus solitarius, a narrower definition that leaves out the area postrema included on this page ²⁴. The minority form is not an error; it is simply rare.

Vagal is the adjective formed from vagus. Gut physiologists usually write dorsal motor nucleus of the vagus; cardiovascular and comparative physiologists often write dorsal vagal motor nucleus ^45,49. The complex, by contrast, has settled on the adjective. Writers should use dorsal vagal complex and should expect to meet the other form occasionally.

A more consequential point is that neither phrase is a controlled indexing term. PubMed expands area postrema and nucleus tractus solitarius into the Medical Subject Headings Area Postrema and Solitary Nucleus when they are searched, but treats dorsal vagal complex as a plain string matched in all fields, with no heading behind it ²⁵. There is also no heading for the dorsal motor nucleus of the vagus as such; papers about it are indexed under Vagus Nerve or Medulla Oblongata ²⁵. The practical consequence for anyone searching this literature is that a subject-heading search will not retrieve the complex as a unit, and a phrase search will miss papers that discuss all three nuclei without using the collective name.

The three abbreviations in common use are AP for area postrema, NTS or NST for the nucleus of the solitary tract, and DMV or DMN or 10N for the dorsal motor nucleus of the vagus. DMN is ambiguous, because in imaging literature it almost always means default mode network, and 10N is an atlas convention derived from the vagus being the tenth cranial nerve. The nucleus of the solitary tract also appears as nucleus tractus solitarii, in which tractus and solitarii are both Latin genitive singular - nucleus of the solitary tract - and that is the form Neuhuber and Berthoud use ³⁸; the commoner nucleus tractus solitarius makes the adjective agree with nucleus instead, which is why the same structure is written with and without the final i.

A fourth structure sometimes appears in this neighborhood in the psychological literature: the ventral vagal complex, a term used in the polyvagal theory for the nucleus ambiguus together with associated nuclei of the cranial nerves serving the face and head ^60,61. It is not a standard term in neuroanatomy in the way the dorsal vagal complex is, and the reader should be aware when the two are set against each other that only one of them has a settled anatomical referent.

The correction this page exists for: what the polyvagal theory claims

If you have encountered the dorsal vagal complex outside a physiology course, it was probably in a therapy room. The polyvagal theory, proposed by Stephen Porges, is taught widely in trauma psychotherapy, somatic therapies, bodywork and coaching, and it has made the phrase dorsal vagal into everyday language for a state of collapse, shutdown, freeze or dissociation. Because that use attaches a specific psychological meaning to a specific piece of brainstem, it is worth setting out exactly what is claimed and exactly what the physiological literature says about it.

Porges introduced the theory in 1995. He argued that although vagal fibers from both the nucleus ambiguus and the dorsal motor nucleus end on the heart's pacemaker, only those from the nucleus ambiguus produce respiratory sinus arrhythmia; that divergent shifts in respiratory sinus arrhythmia and heart rate reflect independent actions of the two nuclei; that a ventral vagal complex distinguishes mammalian from reptilian brainstem organization; and that disorders such as sudden infant death syndrome and asthma may reflect competition between the two nuclei ⁵⁴. In 2001 he described three evolutionary stages, the first of them a primitive unmyelinated visceral vagus that fosters digestion and responds to threat by depressing metabolic activity, associated with immobilization behaviors ⁵⁶, and in 2007 he published an explicit statement of the theory in the same volume of Biological Psychology that carried Grossman and Taylor's measurement critique ^48,58. Porges's own current statement of the theory, published in 2025, describes an evolutionarily informed neurophysiological framework in which the autonomic nervous system is hierarchically organized, with the ventral vagal complex playing a distinctive role in supporting social behavior and physiological flexibility, through mechanisms the theory names neuroception; co-regulation, the idea that one person's physiological state calms or unsettles another's; and dissolution, a term borrowed from the neurologist John Hughlings Jackson for the idea that under threat the nervous system falls back on evolutionarily older circuits ⁶⁰. A companion piece the same year adds that the theory is to be read as a systems-level account incorporating transcriptomics, neurophysiology and clinical application, and that neuroception is the neural process by which safety and threat are detected without conscious awareness ⁶¹. The theory's clinical reach is wide: the 2025 status paper discusses application across trauma, chronic pain, autism, developmental disorders and mood disorders, and advocates a science of safety in clinical practice, education and public health ⁶⁰. The framework has also been applied by other authors to areas such as neonatal sleep and sudden infant death syndrome, in a 2025 narrative review that summarizes the association between vagal activity and neonatal sleep states and proposes the polyvagal frame as a useful lens ⁷³ - a review, not new evidence.

Grossman's 2023 paper is the clearest statement of what the theory asserts about the dorsal vagal complex in particular, and because it is written by a critic it should be checked against the theory's own statements, which the 2025 papers supply ^60,61. As Grossman sets it out, the polyvagal hypotheses rest on a primary assumption that in mammals the ventral and dorsal vagal regions of the brainstem each have their own distinct mediating effect on the control of heart rate; that these putative differences map onto socioemotional behavior, the dorsal to defensive immobilization and the ventral to social affiliative behavior; and that they track trends in the evolution of the vagus nerve ²⁷. He adds that a single measurable phenomenon, respiratory sinus arrhythmia, serves as the linchpin for virtually every premise, and that the theory assumes respiratory sinus arrhythmia to be a mammalian phenomenon, quoting the 2011 statement that it has not been observed in reptiles ²⁷. The theory's most specific physiological case for a dorsal vagal effect on the heart concerns the preterm infant. Porges argued in 2023 that in the fetus and the preterm infant the ventral vagus is not fully functional or myelinated, so cardioinhibitory pathways from the dorsal motor nucleus may dominate, and that the apnea and bradycardia of preterm newborns are, in his words, hypothetically mediated through chronotropic dorsal vagal pathways ⁵⁹. That is a testable hypothesis about immature humans, stated as a hypothesis, and it is narrower than the adult shutdown state taught clinically.

Translated into clinical language, the claim is that a mammal under inescapable threat falls back on an evolutionarily older, dorsal vagal circuit that produces immobilization, bradycardia and shutdown, and that this is what a dissociating patient is doing ⁵⁶. The intuition is powerful and the clinical observation it rests on - that some frightened people go still, pale and faint rather than fight or flee - is real and well described: reviews of human and animal defense responses describe freezing as a form of behavioral inhibition accompanied by a parasympathetically dominated slowing of the heart, which in rodents depends on projections from the amygdala to the periaqueductal gray, the midbrain region that organizes defensive responses ⁷⁴, and clinical reviews describe a defense cascade running from arousal through fight, flight and freezing to tonic and collapsed immobility as responses of last resort, mediated through the amygdala, hypothalamus, periaqueductal gray and both sympathetic and vagal nuclei ⁷⁵. The question is whether the brainstem account attached to it is correct.

One structural note before the evidence. The theory's author published both the 2025 status paper and, in 2026, the reply to the critique ^29,60. That is the ordinary situation of an originator defending a framework and should be weighted as such; whether the author has financial interests in polyvagal-informed programs should be checked against each paper's disclosure statement, which the PubMed records do not reproduce. None of the papers cited in this section carries a retraction or an expression of concern, and the one paper in this group with a formal correction is Neuhuber and Berthoud's 2022 review, which has a published corrigendum ^38,39.

What the published refutations show, and how strong each is

The challenge to the theory has come in three waves, and they differ in kind. The first is measurement: whether respiratory sinus arrhythmia can carry the weight placed on it. The second is anatomy: whether the two vagal pathways do what the theory says. The third is comparative and evolutionary: whether the story about mammals and reptiles holds. The decisive documents are reviews and an expert appraisal, which is the appropriate form for a dispute about the interpretation of existing data, but they rest on primary work: comparative tracing across the vertebrates ⁴⁵, direct recordings from vagal preganglionic neurons in rats ⁴⁹, and simultaneous cardiac and gastric recordings ⁷⁶. None of that primary work was done in people.

On measurement, Grossman and Taylor set out six caveats in 2007: respiratory parameters confound the relation between respiratory sinus arrhythmia and cardiac vagal tone; within-person relations are often strong while between-person associations may be modest; the measure is profoundly influenced by momentary physical activity, which biases estimation of individual differences; its magnitude is affected by beta-adrenergic tone; and the two can dissociate outright ⁴⁸. The sixth caveat was the theory itself: the paper argued that the polyvagal account does not accurately depict the evolution of vagal control of heart-rate variability - the beat-to-beat fluctuation in heart rate, of which respiratory sinus arrhythmia is the part linked to breathing - and that it ignores cardiac aliasing, the artifact that arises when heart rate samples a respiratory rhythm too slowly ⁴⁸. Notably, the authors did not conclude that the measure is useless: they wrote that it often provides a reasonable reflection of cardiac vagal tone when these complexities are taken into account ⁴⁸. In 2023 Grossman added the argument that treating an approximate index of a phenomenon as the phenomenon itself is a category mistake ²⁷. The critique has since come from outside his group. In 2025 McIntyre and Sloan reviewed work treating vagal activity as a readout of self-regulation, under both the polyvagal theory and the rival neurovisceral integration model, and described confusion around constructs, misleading claims about the relation between vagal activity and either respiratory sinus arrhythmia or the high-frequency band of heart rate variability, errors in neuroanatomy and contradictory findings ⁷⁷. In 2026 Bambha and colleagues surveyed 316 developmental studies using respiratory sinus arrhythmia, published over six and a half years, and found that their psychological interpretations rely heavily on the century-old binary model of the autonomic nervous system, which contemporary anatomy and molecular profiling show to be incomplete ⁷⁸. Both are reviews, not new measurements.

On anatomy, Neuhuber and Berthoud reviewed vagal functional anatomy in 2022 with the explicit question of how the polyvagal theory complies with it. Their answer was that the classical list of vagal brainstem nuclei, which includes the nucleus ambiguus, is itself incomplete: vagal afferents also target the spinal trigeminal and paratrigeminal nuclei, and the nucleus of the solitary tract integrates somatic as well as visceral afferents ³⁸. This page reads that finding as weakening the clean two-pathway architecture a hierarchical theory needs; the paper itself sets out the anatomy rather than drawing that conclusion ³⁸. The paper carries a corrigendum published in 2023, which should be read with it ³⁹.

On comparative physiology, the relevant data are old and not in dispute. Taylor reported in 1994 that in elasmobranch fish 8 percent of vagal preganglionic neurons lie outside the dorsal vagal motor nucleus, and that all of these are cardiac motoneurons; that the proportion outside the dorsal nucleus rises from fish through amphibians to more than 30 percent in mammals; and that in the cat 80 percent of cardiac vagal motoneurons are in the nucleus ambiguus ⁴⁵. In the neotenous axolotl all vagal preganglionic neurons sit in the dorsal motor nucleus, and 15 percent move lateral to it after metamorphosis is induced with thyroid hormone ⁴⁵. Taylor, Jordan and Coote set the same relocation out at length in 1999 ⁴⁶. The implication for the theory is direct: in mammals most cardiac vagal motoneurons lie in the nucleus ambiguus, 80 percent of them in the cat ⁴⁵, so a brainstem state described as dorsal vagal cannot be straightforwardly read from a cardiac measure.

Two further strands weaken the picture from different directions. Direct single-unit recording from both populations in rats found that vagal preganglionic neurons of the nucleus ambiguus and of the dorsal vagal motor nucleus are both strongly activated during bouts of exercise, and that exercise training raises resting activity in both while augmenting the excitatory responses of nucleus ambiguus neurons - data the authors describe as arguing against the notion of exercise-induced central vagal withdrawal, and as the first direct neurophysiological evidence that training increases vagal tone ⁴⁹. Separately, a rat study that recorded gastric antral motility and the electrocardiogram simultaneously found that the high-frequency band of heart rate variability correlated positively with proximal antral motility at baseline and after cholecystokinin, but not after ghrelin or food, and framed the comparison with the explicit statement that cardiac and gastrointestinal vagal outputs arise from distinct brainstem nuclei - the nucleus ambiguus and the dorsal motor nucleus respectively ⁷⁶. A cardiac measure is at best an unreliable proxy for what the dorsal vagal complex is doing.

The 2026 appraisal is the strongest form the challenge has taken. Grossman invited thirty-nine acknowledged experts in the physiology and evolution of the vagus and in vertebrate social behavior, many of whose work had been cited in the polyvagal literature as supporting the theory; all but one accepted, the dissenting scholar being unfamiliar with it ²⁸. The resulting paper appraises five major elements: whether respiratory sinus arrhythmia is a direct measure of central vagal drive to the heart; the theory's characterizations of the neuroanatomy and function of the nucleus ambiguus and the dorsal motor nucleus; its assertions about vagal evolution; its claims about the specificity of mammalian social behavior relative to other vertebrates; and its interpretations of earlier seminal physiological work ²⁸. The co-authors agreed that major tenets are unsupported by past or current knowledge and in several instances inconsistent with the broader evidence base, and concluded that the theory is untenable because it is not defensible on existing neurophysiological and evolutionary evidence ²⁸. The appraisal is less independent of the evidence reviewed above than its length suggests: its co-authors include Edwin W. Taylor, author of the comparative studies ⁴⁵; Winfried Neuhuber, co-author of the anatomical review ³⁸; and Alexander Gourine and Gareth Ackland, authors of the exercise recordings ⁴⁹, as its own author list shows ²⁸. It is a consensus of many of the people who produced the primary evidence - strong in expertise, weaker as independent confirmation.

Porges replied in the same issue. His argument is not that each objection is individually wrong but that the critique fails as a refutation: it evaluates, he writes, a reconstructed proxy of the theory shaped by category errors including the conflation of neuroanatomy with neurophysiology, the reduction of a theory to a measurement, and the substitution of phylogenetic continuity for functional organization; where disagreement exists, he argues, it reflects differences in measurement preference, level of analysis or framing rather than evidence against the theory's organizing principles, and an appendix presents a historical audit arguing that several of the claims in the critique were identified nearly two decades earlier as mischaracterizations ²⁹. A reader should hold both: the critics have the comparative anatomy and the measurement literature on their side, and the reply is correct that a systems-level psychological framework is not refuted simply by showing that one of its neuroanatomical premises is wrong - though it is substantially weakened, because the neuroanatomy is what the framework offered as its warrant.

What survives, and what does not. The clinical observation that some people respond to overwhelming threat with immobility, pallor, slowed heart rate and dissociation is not in question ^74,75, and neither is the usefulness of attending to a patient's physiological state. What the evidence does not support is the specific mechanistic story: that this state is produced by the dorsal motor nucleus of the vagus, that it is the activation of an older vagal branch, that respiratory sinus arrhythmia measures how much of it is present, or that mammals differ from reptiles in the way the theory describes ^27,28,45,48. For the purposes of this page, the statement the comparative data force is narrower: in mammals most of the vagal motoneurons that slow the heart lie in the nucleus ambiguus, 80 percent of them in the cat ^45,46, so a heart-rate measure such as respiratory sinus arrhythmia cannot be read as the activity of the dorsal motor nucleus. That nucleus is not irrelevant to the heart. In anesthetized rats, silencing its neurons shortened the ventricular effective refractory period - the shortest interval at which the ventricle will answer a second stimulus - lowered the threshold for triggering ventricular tachycardia and prolonged the corrected QT interval, a measure of ventricular electrical recovery read off the electrocardiogram ⁷⁹, and increased left ventricular contractility ⁸⁰; genetic silencing of the same population dramatically impaired exercise capacity, while optogenetic recruitment, switching the same cells on with light, enhanced cardiac contractility and prolonged exercise endurance ⁸¹. Those are rodent findings about the ventricles rather than about heart rate, and they point away from a shutdown circuit: in these experiments the nucleus supports cardiac performance ⁴⁷.

What goes wrong here in people

Human evidence about this structure is thinner than the animal literature, for an obvious reason: it is a few cubic millimeters of brainstem that cannot be sampled, stimulated or recorded from in a living person. What exists falls into three groups - neuropathology, drug side effects, and inference from disease patterns.

The neuropathology is the strongest human finding. Braak and colleagues examined incidental and symptomatic Parkinson cases and found that the alpha-synuclein lesions that define the disease appear first in the dorsal motor nucleus of the glossopharyngeal and vagal nerves and in the anterior olfactory nucleus, and that the brainstem process pursues an ascending course with little variation between individuals, reaching the substantia nigra only at the third stage ²⁶. In other words, the clinical disease that is named for a movement disorder begins, pathologically, in the output nucleus of the dorsal vagal complex. The staging describes a sequence; it does not by itself prove that the sequence is causal, and the initiating event remains unknown.

The clinical correlate is gastrointestinal, and it precedes the movement disorder by many years. In a Japanese single-center study of 94 patients with Parkinson disease, 78.7 percent complained of constipation and 71.1 percent met the study's definition of fewer than one bowel movement in three days; in 33 of the 74 constipated patients, that is 44.6 percent, the constipation had begun an average of 18.1 years before the motor disturbance, with a standard deviation of 18.8 years ³². The same study found lower total water intake in patients than controls, 604.0 milliliters per day with a standard deviation of 377.2 against 909.5 with a standard deviation of 531.6 ³². These are associations in a small case-control study with self-reported histories recalled across decades, and the authors' own framing is cautious; the numbers are reported here because they are among the few human numbers that exist, not because they settle anything.

A related rat finding limits one tempting explanation. Destroying the substantia nigra in rats decreased choline acetyltransferase staining and increased tyrosine hydroxylase staining in the dorsal vagal complex, which looks like a direct nigral influence on the complex - but retrograde tracing from the complex labeled no neurons in the substantia nigra, and anterograde tracing from the substantia nigra labeled no fibers in the complex, the retrograde tracer instead labeling the hypothalamic paraventricular nucleus, lateral hypothalamus, inferior olive and locus coeruleus ²⁴. Both tracers tested the same direction, so a projection running the other way, from the complex to the nigra, was not examined. In the rat, whatever links the two is indirect, and the authors propose a route through the hypothalamus and/or the locus coeruleus ²⁴.

The second group is drug side effects, and here the human data are abundant and the mechanism inferred. The GLP-1 receptor agonists reliably cause nausea in a substantial minority, with 4.5 percent of patients on semaglutide 2.4 mg discontinuing for gastrointestinal events in STEP 1 against 0.8 percent on placebo ³³. The mouse work places the aversive effect in area postrema GLP-1 receptor neurons and the satiety effect in nucleus of the solitary tract neurons, and shows that intake reduction survives inhibition of the aversion pathway ¹³. Whether that is a therapeutic lead is now disputed: a 2026 mouse study found that the area postrema neurons carry both the weight loss and the aversion, and that the satiety neurons of the solitary nucleus are not needed for the weight loss the drugs produce ³⁴.

Pregnancy sickness is the third. The hormone is fetal, the receptor is in the area postrema, and the human genetics is now substantial: a multi-ancestry genome-wide association study of 10,974 cases of severe pregnancy nausea and vomiting against 461,461 controls found ten loci, including GDF15 and its receptor GFRAL, with opposing effects for GDF15 between maternal and fetal genotype ³⁵. The mechanism proposed by Fejzo and colleagues is that maternal sensitivity, set partly by pre-pregnancy exposure to the hormone, determines how severely a given fetal output is felt - which, if correct, means the complex can be desensitized ¹⁷. The condition has its own page.

The GDF15 system is also where a drug acting on this complex's chemistry has been tested against placebo in people. In a phase 2 randomized, double-blind trial of 187 adults with cancer cachexia - the wasting of muscle and fat that accompanies advanced cancer - and a serum GDF15 of at least 1,500 picograms per milliliter, ponsegromab, an antibody that neutralizes GDF15, given under the skin every four weeks for three doses, increased weight at 12 weeks by a median of 1.22, 1.92 and 2.81 kilograms more than placebo at the 100, 200 and 400 milligram doses, with better appetite and physical activity at the highest dose ⁴⁰. The trial was funded by Pfizer, which makes the drug ⁴⁰. It shows that circulating GDF15 drives appetite loss in people; that the effect runs through the area postrema is inferred from animal work.

Sickness behavior in critical illness is the least developed and possibly the most important. In a rodent sepsis model, chemogenetically activating dorsal vagal complex neurons exacerbated sickness behavior and produced lasting anxiety-related behavior, whereas activating the upstream nodose neurons did not, and subdiaphragmatic vagotomy reduced acute brain activation and partly attenuated anxiety in survivors ¹⁹. Survivors of severe sepsis do show substantial and persistent new impairment: in a US cohort of 516 survivors drawn from a nationally representative survey, the prevalence of moderate to severe cognitive impairment rose by 10.6 percentage points after severe sepsis, a more than threefold increase in the odds (odds ratio 3.34, 95 percent confidence interval 1.53 to 7.25), and those with no previous limitations acquired a mean of 1.57 new limitations in the activities of daily living ⁸². The rodent result suggests a brainstem contribution and a possible target; it remains a mouse experiment using chemogenetic manipulations that have no human equivalent.

What is conspicuously absent from the human literature is any routine clinical test of the complex. There is no blood test, no imaging protocol in standard use, and no electrophysiological measure that isolates it. Heart rate variability is the measurement most often pressed into that role, and for the reasons set out above it is the wrong instrument for this structure ^45,48,76.

What changed, and what is still unknown

Four things about this structure have changed since about 2009 and have not yet reached most textbooks. The first is that the barrier appears to be glial as well as vascular: a dense network of astrocytes and tanycytes at the edge of the area postrema stops tracers that have left the leaky capillaries, and has been proposed, with the outer basement membrane, as an alternative diffusion barrier ^2,10; that the sheet is regulated by nutritional state is suggested only indirectly, by the change in leptin receptor isoforms with food restriction ¹⁰. The second is that the glia are participants in the circuit, not scaffolding: activating astrocytes here stops a mouse eating, inhibiting them abolishes a normal hormonal rhythm in food intake, and a peptide made by these glia produces weight loss in rats without the signs of nausea ^41,42,62. The third is that the region has been resolved into more than a hundred molecularly distinct cell types, with cross-species conservation now mapped from rat to macaque to, for one population, human ^8,9.

The fourth change is the one that will matter most to patients. The obesity drugs act here, and the field has spent three years trying to pry apart the useful effect from the nauseating one. The 2024 demonstration that nucleus of the solitary tract GLP-1 receptor neurons produce satiety without aversion while area postrema GLP-1 receptor neurons produce aversion is the clearest statement of the goal ¹³; the 2025 identification of Adcyap1 neurons as the carriers of semaglutide's effect on energy balance, with only a modest effect on conditioned taste aversion, is the clearest statement of a candidate target ¹⁴; and the 2026 finding that an amylin analog works through a different population again, conserved as far as humans, shows how much room there is in the circuit ⁹.

What is unknown is more interesting. Nobody knows what the human dorsal vagal complex looks like at the level of cell types, because the single-nucleus atlases are overwhelmingly rodent with one macaque dataset; a single human-conserved population has been identified, not a human map ^8,9. Nobody knows whether the glial mechanisms demonstrated in rats and mice operate in people, or at what scale - and the insulin-astrocyte-endozepine pathway, the most complete mechanistic chain in this literature, was demonstrated in male rats only ¹². Nobody has measured how leaky the human area postrema is, because the tracer studies that answer that question in mice cannot be done in a person ².

Nobody knows whether the nausea caused by the GLP-1 drugs can in fact be separated from their efficacy in humans, as opposed to in mice. One mouse study says the pathways are anatomically distinct and that intake reduction survives when the aversion pathway is blocked ¹³; a second says that the weight loss the drugs produce and their aversive effect run through the same area postrema population ³⁴. Whether a molecule can separate them in a human brainstem is an open question that only a trial can answer.

Nobody knows what the dorsal motor nucleus is doing in the first stage of Parkinson disease, or why. Braak's staging establishes where the pathology appears; it does not establish what initiates it, whether the constipation that precedes the diagnosis by decades is caused by the same lesion, or whether intervening there would change anything ^26,32.

And nobody has done the experiment that would settle the polyvagal dispute on its own terms, because it is not clear what that experiment would be. The comparative anatomy is not in doubt and the measurement problems are well characterized; what remains is a disagreement about whether a psychological framework can survive the loss of the neuroanatomical premises it was offered with ^27,28,29. A reader who wants to be fair should notice both that the thirty-nine-author appraisal represents an unusually broad expert consensus against the theory's physiological claims ²⁸, and that it is an appraisal rather than a decisive experiment, whose first author wrote the two earlier critiques ^27,48 and several of whose co-authors wrote the primary studies it relies on ^38,45,49. The practical advice is narrower and safer than either side: when a clinician says a patient is in a dorsal vagal state, they are describing something real about the patient and attaching it to a brainstem nucleus that, in mammals, is not the main source of the vagal control of heart rate, although in rats it does help set the excitability and the contractility of the ventricles ^45,46,79,81.

Finally, a gap of a different kind. The complex is studied almost entirely as a feeding and nausea structure, because that is where the money is. Its role in airway reflexes - a guinea-pig study found inflammation and microglial activation in the complex after particulate exposure, alongside heightened cough sensitivity ²⁰ - in biliary sensation ⁷, in the cephalic control of blood sugar ⁶⁴, and in the brainstem end of the inflammatory reflex ^19,65 is comparatively unexamined. The molecular tools now exist to look; the single-cell map of the nodose ganglion published in 2026, with 106,436 cells and 21 neuronal subtypes, is the kind of resource that makes those questions answerable ⁸³.

The tests, one by one

• Heart rate variability and respiratory sinus arrhythmiaRespiratory sinus arrhythmia is the rise and fall of heart rate with breathing, and the high-frequency band of heart rate variability is the number usually extracted from it. It is the measurement on which almost every clinical claim about vagal tone rests ^27,48.Offered as a non-invasive index of parasympathetic drive to the heart, and in polyvagal-informed practice as an index of autonomic state ^60,61Within one person it often tracks cardiac vagal control reasonably well when the caveats are respected ⁴⁸Grossman and Taylor list six caveats: respiratory parameters confound it, between-person associations may be modest even when within-person ones are strong, momentary physical activity biases it, beta-adrenergic tone affects its magnitude, it can dissociate from cardiac vagal tone, and the polyvagal theory's evolutionary account of it is inaccurate and ignores cardiac aliasing ⁴⁸. Grossman later added that treating it as equivalent to vagal tone is a category mistake ²⁷. It says little directly about the dorsal motor nucleus, because in mammals most of the vagal motoneurons that slow the heart lie in the nucleus ambiguus ⁴⁵
• Gastric motility and accommodation testingMeasuring what the efferent limb actually does - the stomach relaxing to receive a meal, the antrum contracting, the stomach emptying - is the functional readout of the complex in a person ^3,5.To investigate nausea, early fullness, vomiting and suspected gastroparesisIt measures the organ, which is the thing the circuit controls; in freely moving rats, the high-frequency band of heart rate variability correlated positively with proximal antral motility at baseline and after cholecystokinin ⁷⁶The same rat study found the correlation broke down after ghrelin or after food, so a cardiac measure is not a substitute for a gastric one under physiological conditions ⁷⁶. The test cannot distinguish a brainstem cause from a nerve, enteric or muscle cause ⁴
• Direct recording from vagal preganglionic neuronsSingle-unit recording with carbon-fiber microelectrodes from the dorsal motor nucleus and the nucleus ambiguus in rats, the only method that reads the output of these neurons as it happens ⁴⁹.To answer what central vagal drive actually does, rather than inferring it from the heartDirect and unambiguous for the neurons recorded; Korsak and colleagues found that both populations were strongly activated during bouts of exercise and that exercise training raised resting activity in both ⁴⁹Animal only. It samples a handful of neurons at a time, and the finding that vagal drive rises rather than falls during exercise challenges the inference, standard in human work, that vagal drive is withdrawn during exercise; it has not been replicated in people ⁴⁹
• Single-nucleus RNA sequencing atlasesSorting the nuclei of individual cells and reading their messenger RNA, then grouping cells by what they express. This is now the main way the complex is described ^8,9.To find out which cell types exist here and which of them a drug acts onHigh resolution and reproducible across laboratories: a harmonized mouse and rat atlas resolved 123 cell identities from more than 180,000 nuclei ⁸, and a cross-species atlas of more than 530,000 cells across rat, mouse and macaque resolved 80 neuronal populations ⁹It reads gene expression, not activity or function, and the human caudal brainstem is represented thinly; conclusions about which cells matter still rest on rodent manipulation ^8,9
• Post-mortem neuropathologyStaining sections of the medulla for alpha-synuclein, the protein that aggregates in Lewy bodies and Lewy neurites. For the dorsal motor nucleus this remains the definitive examination ²⁶.To stage Parkinson disease and to establish where the pathology beginsBraak and colleagues found that the lesions appear first in the dorsal motor nucleus of the glossopharyngeal and vagal nerves and in the anterior olfactory nucleus, and that the brainstem process ascends with little variation between individuals ²⁶It is available only after death, so it cannot time the onset in a living patient, and the staging describes a sequence rather than proving the sequence is causal ²⁶

Where it connects

Sources

83 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Rogers RC, McTigue DM, Hermann GE. Vagovagal reflex control of digestion: afferent modulation by neural and "endoneurocrine" factors.Am J Physiol · 1995 · 268(1 Pt 1):G1-10doi:10.1152/ajpgi.1995.268.1.G1 · PMID 7840189

   Review that names the dorsal vagal complex as the coordinator of gastric, small intestinal and pancreatic function and stresses that it lies in a circumventricular region without a blood-brain barrier, so the reflex circuitry is exposed to humoral as well as neural input. A conceptual synthesis of animal work; it reports no new measurements.

2. 2
   Morita S, Furube E, Mannari T, Okuda H, Tatsumi K, Wanaka A, Miyata S. Heterogeneous vascular permeability and alternative diffusion barrier in sensory circumventricular organs of adult mouse brain.Cell Tissue Res · 2015 · 363(2):497-511doi:10.1007/s00441-015-2207-7 · PMID 26048259

   Tracer study in adult mice giving the size cut-offs: 389-dalton fluorescein isothiocyanate and 3,000-dalton dextran enter the parenchyma, 10,000-dalton dextran stays perivascular, and a dense glial fibrillary acidic protein-positive astrocyte and tanycyte network, which the authors propose, with the outer basement membrane, as a possible alternative diffusion barrier in place of the endothelium. Mouse only; no human measurement of area postrema permeability exists.

3. 3
   Browning KN, Travagli RA. Central nervous system control of gastrointestinal motility and secretion and modulation of gastrointestinal functions.Compr Physiol · 2014 · 4(4):1339-68doi:10.1002/cphy.c130055 · PMID 25428846

   Long review of extrinsic control of the gut. Establishes the gradient of dependence: the intestines can run themselves, the stomach and esophagus depend heavily on vagal input. A review of mostly rodent physiology; gives no human trial data.

4. 4
   Berthoud HR, Munzberg H, Yu S, Morrison CD, Kim A, Neuhuber WL. Anatomy and function of efferent and afferent vagal innervation of the stomach.Auton Neurosci · 2026 · 267:103465doi:10.1016/j.autneu.2026.103465 · PMID 42628243

   Current review of the gastric vagus. States that dorsal motor nucleus cholinergic neurons end on myenteric enteric neurons rather than on gastric effector cells, that intraganglionic laminar endings are the most abundant afferent terminal, and that transcriptomics supports a predominantly labeled-line organization. A review; the coding logic it describes is largely rodent.

5. 5
   Gillis RA, Dezfuli G, Bellusci L, Vicini S, Sahibzada N. Brainstem neuronal circuitries controlling gastric tonic and phasic contractions: a review.Cell Mol Neurobiol · 2022 · 42(2):333-60doi:10.1007/s10571-021-01084-5 · PMID 33813668

   Four decades of one group's rodent work assembled into a circuit scheme: noradrenergic and nitrergic connections for the fundus, a GABAergic connection for the antrum, somatostatin-GABA interneurons setting basal tone, and an endogenous opioid required for the reflex to operate. A review by the authors of the primary studies, so it carries their interpretation.

6. 6
   Berthoud HR, Neuhuber WL. Functional and chemical anatomy of the afferent vagal system.Auton Neurosci · 2000 · 85(1-3):1-17doi:10.1016/S1566-0702(00)00215-0 · PMID 11189015

   The standard catalog of what the vagus senses: esophagus, lower airways, heart, aorta, the whole gastrointestinal tract, liver, portal vein, biliary system and pancreas, but not the spleen, with named terminal structures. A review of tracing and recording studies, mostly rat and cat.

7. 7
   Neuhuber WL, Yu S, Williams J, Burk DH, Morrison CD, Munzberg H, Berthoud HR. Vagal sensory neurons profusely innervate the mouse gall bladder.Auton Neurosci · 2026 · 266:103443doi:10.1016/j.autneu.2026.103443 · PMID 42258898

   Anterograde viral tracing from the nodose ganglion showing dense vagal sensory innervation of the mouse gall bladder, including laminar endings in gall bladder ganglia. Anatomy only: the authors state that the sensory modalities and any functional role remain to be determined.

8. 8
   Hes C, Tomlinson AJ, Michielsen L, Murdoch HJ, Soltani F, Kokoeva MV, Sabatini PV. A unified rodent atlas reveals the cellular complexity and evolutionary divergence of the dorsal vagal complex.Elife · 2025 · 14:RP106217doi:10.7554/eLife.106217 · PMID 41055205

   Harmonizes mouse and rat single-nucleus data into more than 180,000 cells and 123 cell identities at five levels of resolution, and reports species differences including a leptin receptor-expressing neuron population in the rat area postrema. Transcriptomic description; it assigns no functions.

9. 9
   Ludwig MQ, Coester B, Gordian D, et al. A cross-species atlas of the dorsal vagal complex reveals neural mediators of the effects of cagrilintide on energy balance.Nat Metab · 2026 · 8(6):1350-67doi:10.1038/s42255-026-01539-3 · PMID 42260119

   More than 530,000 cells across rat, mouse and macaque caudal brainstem, 80 neuronal populations, with spatial mapping; identifies nucleus of the solitary tract calcitonin receptor and Prlh neurons, conserved to humans, as mediators of the amylin analog cagrilintide but not of semaglutide. Several authors are Novo Nordisk employees and cagrilintide is a Novo Nordisk compound.

10. 10
    Dallaporta M, Pecchi E, Pio J, Jean A, Horner KC, Troadec JD. Expression of leptin receptor by glial cells of the nucleus tractus solitarius: possible involvement in energy homeostasis.J Neuroendocrinol · 2009 · 21(1):57-67doi:10.1111/j.1365-2826.2008.01799.x · PMID 19094094

    First description of leptin receptors on glial fibrillary acidic protein-positive glia at the area postrema and nucleus of the solitary tract border in rats, with unbranched processes radiating from the fourth ventricle wall, and tracer evidence that these cells may form a diffusion barrier. Rat; the barrier role is proposed, not proven.

11. 11
    Spaeth AM, Kanoski SE, Hayes MR, Grill HJ. TrkB receptor signaling in the nucleus tractus solitarius mediates the food intake-suppressive effects of hindbrain BDNF and leptin.Am J Physiol Endocrinol Metab · 2012 · 302(10):E1252-60doi:10.1152/ajpendo.00025.2012 · PMID 22374757

    Shows that hindbrain leptin raises brain-derived neurotrophic factor in the dorsal vagal complex and that a TrkB antagonist in the medial nucleus of the solitary tract attenuates leptin's suppression of food intake in rats. Pharmacological microinjection; does not establish the physiological route of circulating leptin.

12. 12
    New LE, Wang N, Smith HE, Birks R, Afridi SK, Griffiths JC, Hains R, Johnston J, Filippi BM. Insulin evokes release of endozepines from astrocytes of the NTS to modulate glucose metabolism in male rats.Mol Metab · 2025 · 101:102255doi:10.1016/j.molmet.2025.102255 · PMID 40976282

    A complete mechanistic chain: insulin receptors on nucleus of the solitary tract astrocytes, endozepine release, action at the benzodiazepine site of the GABA-A receptor, control of hepatic glucose production, and rescue of high-fat-diet-induced insulin resistance by endozepine. Male rats only; no female data and no human work.

13. 13
    Huang KP, Acosta AA, Ghidewon MY, et al. Dissociable hindbrain GLP1R circuits for satiety and aversion.Nature · 2024 · 632(8025):585-93doi:10.1038/s41586-024-07685-6 · PMID 38987598

    Two-photon imaging in behaving mice showing that individual hindbrain GLP-1 receptor neurons are tuned either to nutritive or aversive stimuli, that nucleus of the solitary tract activation gives satiety without aversion while area postrema activation gives aversion, and that drug efficacy survives inhibition of the aversion pathway. Mouse; the therapeutic implication is a hypothesis.

14. 14
    Teixidor-Deulofeu J, Blid Skoldheden S, Font-Girones F, Fejes A, Ruud J, Engstrom Ruud L. Semaglutide effects on energy balance are mediated by Adcyap1 neurons in the dorsal vagal complex.Cell Metab · 2025 · 37(7):1530-1546.e6doi:10.1016/j.cmet.2025.04.018 · PMID 40409256

    Ablation of Adcyap1-expressing area postrema and nucleus of the solitary tract neurons largely reversed semaglutide's effects on food intake and body weight in lean and in subchronically treated obese mice, with preferential fat loss and only a modest effect on conditioned taste aversion. Mouse; ablation is permanent and non-selective within the labeled population.

15. 15
    Gutgesell RM, Khalil A, Liskiewicz A, et al. GIPR agonism and antagonism decrease body weight and food intake via different mechanisms in male mice.Nat Metab · 2025 · 7(6):1282-98doi:10.1038/s42255-025-01294-x · PMID 40301583

    Single-nucleus sequencing showing opposing transcriptional effects of GIP receptor agonism and antagonism in the dorsal vagal complex, with antagonism resembling GLP-1 receptor signaling. Male mice only. Several authors are Eli Lilly employees, and the compounds concerned are of commercial interest to that company. A Publisher Correction has been issued (Nat Metab 2025;7(6):1299; PMID 40360757; doi 10.1038/s42255-025-01308-8). The Eli Lilly affiliations of Finan, Sloop and Samms are confirmed in the PubMed record.

16. 16
    Province HS, Hayes NW, Leong NA, Tang E, Lorch CM, Pekerman A, Xia JL, Beutler LR. GIP receptor agonism suppresses inflammation-induced aversion and food intake via distinct circuits.Cell Rep · 2026 · 45(4):117116doi:10.1016/j.celrep.2026.117116 · PMID 41860867

    Separates the two effects of GIP receptor agonism: dorsal vagal complex receptors were required for the acute anorectic effect, while the anti-aversive effect ran through parabrachial calcitonin gene-related peptide neurons. Mouse; interleukin-1 beta is a model of inflammation, not a disease.

17. 17
    Fejzo M, Rocha N, Cimino I, et al. GDF15 linked to maternal risk of nausea and vomiting during pregnancy.Nature · 2024 · 625(7996):760-7doi:10.1038/s41586-023-06921-9 · PMID 38092039

    Combines mass spectrometry of a labeled GDF15 variant, human genetics and mouse experiments to argue that fetal GDF15 production and maternal prior sensitivity jointly set the risk of hyperemesis gravidarum, with beta-thalassemia as a natural experiment. The desensitization mechanism is shown in mice, not in pregnant women.

18. 18
    Fujii R, Kawai H, Kai M, Kondo M. Integrated behavioral and biological assessment reveals 5-HT3 receptor-independent nausea/malaise-like responses in cisplatin-treated mice.Biochem Biophys Res Commun · 2026 · 819:153812doi:10.1016/j.bbrc.2026.153812 · PMID 42025017

    5-HT3 receptor knockout mice developed the full cisplatin malaise-like state, with increased c-Fos and GFRAL double-positive cells in the area postrema and nucleus of the solitary tract and plasma GDF15 elevated for at least a day. Suggests a 5-HT3-independent route; it does not show that the 5-HT3 pathway is unimportant for vomiting in species that vomit.

19. 19
    Bourhy L, Moigneu C, Dupin A, Delelis E, Levy J, Lledo PM, Sharshar T, Lepousez G. Vagal subcircuits differentially regulate sepsis-induced sickness and anxiety-related behaviors.Cell Rep · 2026 · 45(9):117896doi:10.1016/j.celrep.2026.117896 · PMID 42658674

    Cecal ligation and puncture model with chemogenetic gain and loss of function. Activating dorsal vagal complex neurons worsened sickness behavior and produced lasting anxiety-related behavior while activating nodose neurons did not; subdiaphragmatic vagotomy reduced acute brain activation and partly attenuated anxiety. Transient inhibition at either site had no effect, which the authors note complicates a simple necessity claim.

20. 20
    Zhang S, Long L, Chai S, Lin M, Lu H, Liu X, He Y, Dong R, Chen Z. Melatonin inhibits the activation of microglia and cough sensitivity of guinea pigs exposed to PM2.5.Histol Histopathol · 2025 · 40(8):1253-60doi:10.14670/HH-18-850 · PMID 39659195

    Measures interleukin-1 beta and tumor necrosis factor alpha directly in the dorsal vagal complex of guinea pigs after particulate exposure, with microglial activation and heightened cough sensitivity, all attenuated by melatonin. Small animal study of an airway endpoint; cited here as one of the few direct measurements of cytokines inside the complex.

21. 21
    Belmont-Rausch DM, Ludwig MQ, Bentsen MA, et al. Semaglutide attenuates neuroinflammation in male mice.Nat Commun · 2026 · 17(1)doi:10.1038/s41467-026-74038-4 · PMID 42265098

    Lipopolysaccharide model showing that semaglutide suppresses inflammatory transcriptional signatures in microglia, endothelial cells and pericytes, and identifying GLP-1 receptor-expressing dorsal vagal complex neurons that regulate anti-inflammatory genes. Male mice only. Most authors are Novo Nordisk employees or work at the Novo Nordisk Foundation Center for Basic Metabolic Research; the claimed relevance to Alzheimer disease is a transcriptional overlap, not a clinical result.

22. 22
    Feetham CH, Ai M, Culotta I, et al. Downstream interaction by glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide agonism is required for synergistic effects on body weight.Mol Metab · 2025 · 99:102214doi:10.1016/j.molmet.2025.102214 · PMID 40681104

    Fluorescently labeled tirzepatide penetrated the brain mainly at circumventricular organs; both GLP-1 and GIP receptors in the dorsal vagal complex were required for the effects of systemic agonists; GIP receptor agonism alone directly activated a distinct GABAergic population in the area postrema, and the full synergy on body weight required central amygdala neurotensin neurons. Mouse. Several authors are Eli Lilly employees and the company makes tirzepatide.

23. 23
    Carvas AO, Leuthardt A, Kulka P, et al. Cagrilintide lowers bodyweight through brain amylin receptors 1 and 3.EBioMedicine · 2025 · 118:105836doi:10.1016/j.ebiom.2025.105836 · PMID 40609154

    In RAMP1/3 knockout mice, cagrilintide's weight-lowering potency was impeded and area postrema c-Fos activation by the drug fell by 57 percent against wild-type, showing dependence on amylin receptors built from the calcitonin receptor and RAMP1 and RAMP3. Mouse, 8 animals per group; funded by an investigator-led Novo Nordisk consortium grant with Novo Nordisk employees among the authors.

24. 24
    Wang ZY, Lian H, Cai QQ, Song HY, Zhang XL, Zhou L, Zhang YM, Zheng LF, Zhu JX. No direct projection is observed from the substantia nigra to the dorsal vagus complex in the rat.J Parkinsons Dis · 2014 · 4(3):375-83doi:10.3233/JPD-130279 · PMID 24613863

    Cited for two things: it is one of four PubMed records using the minority spelling dorsal vagus complex anywhere in the record, and one of only two carrying it in the title. By retrograde and anterograde tracing it also shows that the substantia nigra does not project directly to the complex in rats, even though destroying the nigra changes the complex's chemistry; both tracers tested that one direction, so the reverse projection was not examined. Rat anatomy.

25. 25
    PubMed index query for the exact phrases 'dorsal vagal complex' and 'dorsal vagus complex', and for the Medical Subject Heading expansion of 'area postrema' and 'nucleus tractus solitarius'.National Library of Medicine · query run 19 October 2026

    Primary document, not indexed in PubMed: a database query rather than an article. Returned about 740 records for dorsal vagal complex and 4 for dorsal vagus complex, only 2 of them with the minority phrase in the title, and showed that the search engine expands area postrema and nucleus tractus solitarius into Medical Subject Headings while matching dorsal vagal complex only as free text. Counts drift as records are added. URL: https://pubmed.ncbi.nlm.nih.gov/?term=%22dorsal+vagal+complex%22

26. 26
    Braak H, Del Tredici K, Rub U, de Vos RAI, Jansen Steur ENH, Braak E. Staging of brain pathology related to sporadic Parkinson's disease.Neurobiol Aging · 2003 · 24(2):197-211doi:10.1016/s0197-4580(02)00065-9 · PMID 12498954

    The staging scheme placing the first alpha-synuclein lesions in the dorsal motor nucleus of the glossopharyngeal and vagal nerves and the anterior olfactory nucleus, with an ascending brainstem course. Post-mortem series; it establishes a topographical sequence, not a cause, and the scheme has been contested for cases that do not fit.

27. 27
    Grossman P. Fundamental challenges and likely refutations of the five basic premises of the polyvagal theory.Biol Psychol · 2023 · 180:108589doi:10.1016/j.biopsycho.2023.108589 · PMID 37230290

    Sets out the five premises as stated by the theory's author in 2011, argues that each is untenable or highly implausible on the available literature, and makes the category-mistake argument about respiratory sinus arrhythmia. A single-author review; it reanalyzes no primary data and its force depends on the comparative literature it cites.

28. 28
    Grossman P, Ackland GL, Allen AM, et al. Why the polyvagal theory is untenable: an international expert evaluation of the polyvagal theory and commentary upon Porges, S.W. (2025).Clin Neuropsychiatry · 2026 · 23(1):100-12doi:10.36131/cnfioritieditore20260110 · PMID 41768017

    Thirty-nine experts were invited by the first author and all but one accepted, so the paper has 39 authors - Grossman and 38 co-authors; the one who declined said they were unfamiliar with the theory. It appraises five elements of the theory and concludes that the theory is untenable on existing neurophysiological and evolutionary evidence. An expert appraisal, not an experiment; the co-author list was assembled by the critique's organizer, and it includes the authors of several of the primary studies the critique relies on, among them Edwin W. Taylor [49], Winfried Neuhuber [46], and Alexander Gourine and Gareth Ackland [51].

29. 29
    Porges SW. When a critique becomes untenable: a scholarly response to Grossman et al.'s evaluation of polyvagal theory.Clin Neuropsychiatry · 2026 · 23(1):113-28doi:10.36131/cnfioritieditore20260111 · PMID 41768026

    The reply in the same issue, arguing that the critique evaluates a reconstructed proxy of the theory, conflates neuroanatomy with neurophysiology, reduces theory to measurement, and substitutes phylogenetic continuity for functional organization. It does not present new physiological data; it is an argument about what would count as a refutation.

30. 30
    Tache Y, Adelson D, Yang H. TRH/TRH-R1 receptor signaling in the brain medulla as a pathway of vagally mediated gut responses during the cephalic phase.Curr Pharm Des · 2014 · 20(16):2725-30doi:10.2174/13816128113199990578 · PMID 23886382

    Review of the thyrotropin-releasing hormone pathway into the dorsal motor nucleus, including the finding that blocking the peptide or its receptor in the brainstem abolishes sham-feeding-induced gastric acid secretion in rats. Rodent mechanism; the human cephalic phase is inferred by analogy.

31. 31
    Hogan DL, Ainsworth MA, Isenberg JI. Review article: gastroduodenal bicarbonate secretion.Aliment Pharmacol Ther · 1994 · 8(5):475-88doi:10.1111/j.1365-2036.1994.tb00319.x · PMID 7865639

    Used here only for its human physiology: sham feeding raises bicarbonate secretion in both stomach and duodenum, the gastric component mediated by cholinergic vagal fibers and the duodenal component apparently not. An older review; the primary studies are small human experiments.

32. 32
    Ueki A, Otsuka M. Life style risks of Parkinson's disease: association between decreased water intake and constipation.J Neurol · 2004 · 251 Suppl 7:vII18-23doi:10.1007/s00415-004-1706-3 · PMID 15505750

    94 patients with Parkinson disease: 78.7 percent reported constipation, 71.1 percent met the study's definition, and in 44.6 percent of the constipated patients constipation preceded motor disturbance by 18.1 plus or minus 18.8 years. A small single-center case-control study relying on recall across decades; the water-intake finding is an association and the authors call for prospective work.

33. 33
    Wilding JPH, Batterham RL, Calanna S, et al. Once-weekly semaglutide in adults with overweight or obesity.N Engl J Med · 2021 · 384(11):989-1002doi:10.1056/NEJMoa2032183 · PMID 33567185

    STEP 1: 1,961 adults, 68 weeks, mean weight change minus 14.9 percent against minus 2.4 percent, and 4.5 percent against 0.8 percent discontinuing for gastrointestinal events. Funded by Novo Nordisk with company employees among the authors. It measures outcomes, not mechanism, and says nothing directly about the brainstem.

34. 34
    Yacawych WT, Wang Y, Zhou G, et al. A single dorsal vagal complex circuit mediates the aversive and anorectic responses to GLP1R agonists.J Clin Invest · 2026 · published 15 September 2026doi:10.1172/JCI193743 · PMID 42742987

    Mouse study using neuron silencing and receptor restoration. Nucleus of the solitary tract GLP-1 receptor neurons restrained ordinary feeding but were not required for weight loss on a GLP-1 receptor agonist, while area postrema GLP-1 receptor neurons mediated both the weight-lowering and the aversive effects - which the authors say prevents separating nausea from weight loss at the circuit level. Contradicts the satiety-aversion separation of Huang 2024. Several authors are Novo Nordisk employees or work at the Novo Nordisk Foundation Center for Basic Metabolic Research.

35. 35
    Fejzo M, Wang X, Tan Q, et al. Multi-ancestry genome-wide association study of severe pregnancy nausea and vomiting.Nat Genet · 2026 · 58(4):810-20doi:10.1038/s41588-026-02564-4 · PMID 41981316

    10,974 cases and 461,461 controls across four ancestry groups; ten loci including GDF15, IGFBP7, PGR and GFRAL and six new ones, with opposing maternal and fetal effects for GDF15. Association genetics: it identifies loci, not mechanisms, and the functional claims rest on expression mapping. A Publisher Correction has been issued (Nat Genet 2026;58(5):1186; PMID 42050060; doi 10.1038/s41588-026-02608-9); readers should consult the corrected version.

36. 36
    Shapiro RE, Miselis RR. The central organization of the vagus nerve innervating the stomach of the rat.J Comp Neurol · 1985 · 238(4):473-88doi:10.1002/cne.902380411 · PMID 3840183

    The anatomical reason for treating the nuclei as one complex: dorsal motor nucleus gastric motoneurons send numerous dendrites into the subnuclei of the solitary tract that receive gastric vagal afferent terminals, which the authors propose as loci for monosynaptic vagovagal interaction, and a small number penetrate the ependyma of the fourth ventricle or enter the area postrema. Rat tracing study.

37. 37
    Zhang C, Kaye JA, Cai Z, Wang Y, Prescott SL, Liberles SD. Area postrema cell types that mediate nausea-associated behaviors.Neuron · 2021 · 109(3):461-472.e5doi:10.1016/j.neuron.2020.11.010 · PMID 33278342

    Single-nucleus atlas of the area postrema with cell-specific manipulation, identifying excitatory neurons whose activation induces nausea-related behaviors and mapping long-range excitatory but not inhibitory projections. Mouse; nausea is inferred from behavior because mice cannot vomit.

38. 38
    Neuhuber WL, Berthoud HR. Functional anatomy of the vagus system: how does the polyvagal theory comply?Biol Psychol · 2022 · 174:108425doi:10.1016/j.biopsycho.2022.108425 · PMID 36100134

    Anatomical review written to test the theory against tracing data. Finds the classical three-nucleus account incomplete: vagal afferents also target the spinal trigeminal and paratrigeminal nuclei and the solitary nucleus integrates somatic as well as visceral input. A review by two anatomists; see the corrigendum.

39. 39
    Neuhuber WL, Berthoud HR. Corrigendum to 'Functional anatomy of the vagus system: how does the polyvagal theory comply?'Biol Psychol · 2023 · 179:108554doi:10.1016/j.biopsycho.2023.108554 · PMID 37059643

    Published erratum to the 2022 review. No abstract is available in the PubMed record, so the substance of the correction cannot be stated here; readers citing the 2022 paper should obtain it.

40. 40
    Groarke JD, Crawford J, Collins SM, et al. Ponsegromab for the treatment of cancer cachexia.N Engl J Med · 2024 · 391(24):2291-2303doi:10.1056/NEJMoa2409515 · PMID 39282907

    Phase 2 randomized, double-blind trial in 187 patients with cancer cachexia and serum GDF15 of at least 1,500 pg/mL; the GDF15-neutralizing antibody increased weight at 12 weeks by a median of 1.22, 1.92 and 2.81 kg over placebo at 100, 200 and 400 mg. The strongest human evidence that circulating GDF15 drives appetite loss; funded by Pfizer, which makes the drug, with Pfizer employees among the authors. It blocks the pathway rather than targeting GFRAL as an obesity drug would.

41. 41
    MacDonald AJ, Holmes FE, Beall C, Pickering AE, Ellacott KLJ. Regulation of food intake by astrocytes in the brainstem dorsal vagal complex.Glia · 2020 · 68(6):1241-54doi:10.1002/glia.23774 · PMID 31880353

    The founding demonstration that astrocytes here are part of the feeding circuit: a 65 percent increase in glial fibrillary acidic protein-immunoreactive cells after 12 hours of high-fat chow, and an 84 percent reduction in dark-phase feeding on chemogenetic activation, without any effect on locomotion. Mouse; chemogenetic activation is supraphysiological by design.

42. 42
    Ozkaya KS, Yalcin C, Haar EE, Bhagat R, Browning KN. Brainstem astrocytes regulate estrus-dependent oscillations in food intake.J Neurochem · 2026 · 170(6):e70498doi:10.1111/jnc.70498 · PMID 42313475

    Ties dorsal vagal complex astrocytes to a normal physiological rhythm: astrocyte complexity rises with high estrogen, chemogenetic inhibition abolishes the cycle-linked oscillation in food intake, and the effect runs through NMDA receptors on dorsal motor nucleus neurons. Rat; the estrous cycle is not the human menstrual cycle and the extrapolation is not tested.

43. 43
    Barbot M, Lebrun B, Barbouche R, Gaige S, Tonetto A, Abysique A, Troadec JD. Glial plasticity and metabolic stability after knockdown of astrocytic Cx43 in the dorsal vagal complex.Cells · 2025 · 14(21):1694doi:10.3390/cells14211694 · PMID 41227339

    A negative result worth recording: knocking down connexin 43 in dorsal vagal complex astrocytes changed astrocyte and microglial morphology but did not change the metabolic phenotype on standard or high-fat, high-sugar diets. The buffering interpretation offered by the authors is plausible but is the reading a null result invites.

44. 44
    Guillebaud F, Girardet C, Abysique A, Gaige S, Barbouche R, Verneuil J, Jean A, Leprince J, Tonon MC, Dallaporta M, Lebrun B, Troadec JD. Glial endozepines inhibit feeding-related autonomic functions by acting at the brainstem level.Front Neurosci · 2017 · 11:308doi:10.3389/fnins.2017.00308 · PMID 28611581

    Shows diazepam-binding inhibitor in ependymocytes, tanycyte-like cells and astrocytes at the area postrema and nucleus of the solitary tract interface in rats, and that its peptide fragments injected into the fourth ventricle trigger c-Fos in the complex, reduce food intake and inhibit the swallowing reflex. Rat; pharmacological doses into the ventricle, not physiological release.

45. 45
    Taylor EW. The evolution of efferent vagal control of the heart in vertebrates.Cardioscience · 1994 · 5(3):173-82PMID 7827254

    Source of the key comparative numbers: 8 percent of vagal preganglionic neurons outside the dorsal vagal motor nucleus in elasmobranch fish, all of them cardiac; more than 30 percent outside in mammals; 80 percent of cardiac vagal motoneurons in the nucleus ambiguus in the cat; and the axolotl experiment. No DOI is listed in the PubMed record for this 1994 article.

46. 46
    Taylor EW, Jordan D, Coote JH. Central control of the cardiovascular and respiratory systems and their interactions in vertebrates.Physiol Rev · 1999 · 79(3):855-916doi:10.1152/physrev.1999.79.3.855 · PMID 10390519

    The long review of cardiorespiratory integration across vertebrates, including the partial relocation of vagal preganglionic neurons, particularly cardiac ones, from the dorsal motor nucleus into the nucleus ambiguus. A review; the primary data are from many species and decades.

47. 47
    Gourine AV, Machhada A, Trapp S, Spyer KM. Cardiac vagal preganglionic neurones: an update.Auton Neurosci · 2016 · 199:24-8doi:10.1016/j.autneu.2016.06.003 · PMID 27396874

    Short review arguing that distinct groups of cardiac vagal preganglionic neurons control nodal tissue, which sets heart rate and atrioventricular conduction, and the ventricular myocardium, whose contractility and excitability they modulate. Useful for separating the heart-rate question from the ventricular one.

48. 48
    Grossman P, Taylor EW. Toward understanding respiratory sinus arrhythmia: relations to cardiac vagal tone, evolution and biobehavioral functions.Biol Psychol · 2007 · 74(2):263-85doi:10.1016/j.biopsycho.2005.11.014 · PMID 17081672

    The measurement critique in full: six named caveats, the cardiac aliasing problem, and the argument that the polyvagal account misdescribes the evolution of vagal heart-rate control. Importantly, the authors also conclude the measure often reflects cardiac vagal tone reasonably when the complexities are handled. A review.

49. 49
    Korsak A, Kellett DO, Aziz Q, Anderson C, D'Souza A, Tinker A, Ackland GL, Gourine AV. Immediate and sustained increases in the activity of vagal preganglionic neurons during exercise and after exercise training.Cardiovasc Res · 2023 · 119(13):2329-41doi:10.1093/cvr/cvad115 · PMID 37516977

    Single-unit carbon-fiber recordings from nucleus ambiguus and dorsal vagal motor nucleus preganglionic neurons in rats. Both populations were strongly activated during exercise and training raised resting activity in both; the authors state the data argue against central vagal withdrawal during exercise. Rat; not replicated in humans.

50. 50
    Swanson LW, Hartman BK. Biochemical specificity in central pathways related to peripheral and intracerebral homeostatic functions.Neurosci Lett · 1980 · 16(1):55-60doi:10.1016/0304-3940(80)90100-7 · PMID 6189002

    One of the earliest PubMed records using the phrase dorsal vagal complex, here in the abstract: oxytocin-stained fibers from the hypothalamic paraventricular nucleus innervate the locus coeruleus and the dorsal vagal complex in the rat, and adrenergic fibers project back. Immunohistochemistry and horseradish peroxidase tracing; a short report.

51. 51
    Swanson LW, Kuypers HG. The paraventricular nucleus of the hypothalamus: cytoarchitectonic subdivisions and organization of projections to the pituitary, dorsal vagal complex, and spinal cord as demonstrated by retrograde fluorescence double-labeling methods.J Comp Neurol · 1980 · 194(3):555-70doi:10.1002/cne.901940306 · PMID 7451682

    Cited for the term rather than the content: a 1980 paper with dorsal vagal complex in its title, which is why this page does not claim the name was coined in the 1990s. Double-label retrograde tracing in the rat.

52. 52
    Leslie RA. Neuroactive substances in the dorsal vagal complex of the medulla oblongata: nucleus of the tractus solitarius, area postrema, and dorsal motor nucleus of the vagus.Neurochem Int · 1985 · 7(2):191-211doi:10.1016/0197-0186(85)90106-8 · PMID 20492915

    A 1985 review of the region's transmitters whose title names the complex and all three of its members, showing that the grouping was settled by then. A review of animal work.

53. 53
    Fox EA, Powley TL. Longitudinal columnar organization within the dorsal motor nucleus represents separate branches of the abdominal vagus.Brain Res · 1985 · 341(2):269-82doi:10.1016/0006-8993(85)91066-2 · PMID 4041795

    Shows that each branch of the subdiaphragmatic vagus arises from its own topographically distinct longitudinal column of cells in the dorsal motor nucleus of the rat, so the output nucleus is organized by target organ. Fluorescent tracing; rat.

54. 54
    Porges SW. Orienting in a defensive world: mammalian modifications of our evolutionary heritage. A Polyvagal Theory.Psychophysiology · 1995 · 32(4):301-18doi:10.1111/j.1469-8986.1995.tb01213.x · PMID 7652107

    The founding statement of the theory, cited here so that the page states the claims from their source rather than only from critics. A review and theoretical paper; it reports no new experiment.

55. 55
    Porges SW, Doussard-Roosevelt JA, Portales AL, Greenspan SI. Infant regulation of the vagal 'brake' predicts child behavior problems: a psychobiological model of social behavior.Dev Psychobiol · 1996 · 29(8):697-712doi:10.1002/(SICI)1098-2302(199612)29:8<697::AID-DEV5>3.0.CO;2-O · PMID 8958482

    The paper that introduces the vagal brake, defined as vagal reactivity to challenge and measured as respiratory sinus arrhythmia. Data from 24 infants: those who had difficulty decreasing vagal tone during a social and attention task at 9 months had significantly more behavior problems at 3 years. A small observational study by the theory's author.

56. 56
    Porges SW. The polyvagal theory: phylogenetic substrates of a social nervous system.Int J Psychophysiol · 2001 · 42(2):123-46doi:10.1016/s0167-8760(01)00162-3 · PMID 11587772

    The three-stage hierarchy, including the primitive unmyelinated visceral vagus that responds to threat by depressing metabolic activity and is associated with immobilization behaviors - the origin of the clinical dorsal vagal state. A review.

57. 57
    Partosoedarso ER, Abrahams TP, Scullion RT, Moerschbaecher JM, Hornby PJ. Cannabinoid1 receptor in the dorsal vagal complex modulates lower oesophageal sphincter relaxation in ferrets.J Physiol · 2003 · 550(Pt 1):149-58doi:10.1113/jphysiol.2003.042242 · PMID 12879865

    Gives the numbers for a vagovagal reflex: 30 mL gastric distension relaxed the sphincter by 70 percent, blunted by delta-9-tetrahydrocannabinol and reversed by a cannabinoid-1 antagonist, with receptor staining on intrinsic brainstem elements rather than vagal afferents. Decerebrate ferret preparation; a mechanism, not a clinical result.

58. 58
    Porges SW. The polyvagal perspective.Biol Psychol · 2007 · 74(2):116-43doi:10.1016/j.biopsycho.2006.06.009 · PMID 17049418

    An explicit statement of the theory, published in the same volume of the journal as Grossman and Taylor's measurement critique. A review.

59. 59
    Porges SW. The vagal paradox: a polyvagal solution.Compr Psychoneuroendocrinol · 2023 · 16:100200doi:10.1016/j.cpnec.2023.100200 · PMID 38108034

    The theory's most specific physiological claim about the dorsal motor nucleus: that in the fetus and preterm infant the ventral vagus is not fully functional or myelinated, so dorsal cardioinhibitory pathways may dominate, and that preterm apnea and bradycardia are hypothetically mediated through chronotropic dorsal vagal pathways. A review stating a hypothesis, not a test of it.

60. 60
    Porges SW. Polyvagal theory: current status, clinical applications, and future directions.Clin Neuropsychiatry · 2025 · 22(3):169-84doi:10.36131/cnfioritieditore20250301 · PMID 40735382

    The theory's own current statement by its author: hierarchical autonomic states, the ventral vagal complex, neuroception, co-regulation, dissolution, and clinical application across trauma, chronic pain, autism, developmental disorders and mood disorders. A narrative review; it presents no new data and is the paper the 2026 expert evaluation was written to assess.

61. 61
    Porges SW. Polyvagal theory: a journey from physiological observation to neural innervation and clinical insight.Front Behav Neurosci · 2025 · 19:1659083doi:10.3389/fnbeh.2025.1659083 · PMID 41035859

    A companion restatement combining the scientific case with the author's personal account of the theory's origins. Narrative review by the originator; useful for establishing what is claimed, not as independent evidence.

62. 62
    Geisler CE, Chichura KS, Orativskyi O, et al. Hindbrain octadecaneuropeptide gliotransmission as a therapeutic target for energy balance control without nausea or emesis.Sci Transl Med · 2025 · 17(808):eadu6764doi:10.1126/scitranslmed.adu6764 · PMID 40700519

    Hindbrain-targeted octadecaneuropeptide, a glial peptide of the dorsal vagal complex, caused weight loss and improved glucose clearance in rats; blocking central signaling blunted the response to GLP-1 receptor agonists; a peripherally given derivative reduced weight without pica in rats or emesis in the musk shrew, and without changes in heart rate, temperature or activity. Rat, mouse and shrew; not given to a person.

63. 63
    Dai D, Zhang B, Zhang X, et al. A hypothalamic-vagal pathway mediates stress-induced feeding suppression and gastric dysfunction.Neuron · 2026 · published 11 September 2026doi:10.1016/j.neuron.2026.08.011 · PMID 42727580

    Acute stressors activated neuropeptide Y1 receptor neurons of the hypothalamic paraventricular nucleus that synapse directly on cholinergic neurons of the dorsal vagal complex; activating the pathway suppressed feeding and disordered motility, and inhibiting it rescued stress-induced slowing of gastric emptying. Mouse circuit mapping; the nearest thing in the physiological literature to a dorsal vagal stress response, and it acts on the gut.

64. 64
    Li C, Ngurari S, Ye JJ, Toda C. Cephalic phase response in systemic glucose metabolism: how taste triggers glucose homeostasis.J Oral Biosci · 2026 · 68(5):100813doi:10.1016/j.job.2026.100813 · PMID 42721548

    Current review arguing that taste-driven cephalic phase responses are assembled across the rostral nucleus of the solitary tract, the dorsal vagal complex and vagal parasympathetic output, and can blunt post-meal hyperglycemia before absorption. A review of rodent and human work; the authors state the central mechanisms remain incompletely elucidated.

65. 65
    Kelly MJ, Breathnach C, Tracey KJ, Donnelly SC. Manipulation of the inflammatory reflex as a therapeutic strategy.Cell Rep Med · 2022 · 3(7):100696doi:10.1016/j.xcrm.2022.100696 · PMID 35858588

    Review of the inflammatory reflex: vagal and splenic nerve signaling, acetylcholine from choline acetyltransferase-expressing T cells, and alpha-7 nicotinic receptors on macrophages, with the translational trials to date. A review co-authored by the pathway's principal proponent.

66. 66
    Furube E, Ohgidani M, Tanaka Y, Miyata S, Yoshida S. A high-fat diet influences neural stem and progenitor cell environment in the medulla of adult mice.Neuroscience · 2024 · 559:64-76doi:10.1016/j.neuroscience.2024.08.034 · PMID 39209104

    Shows that one week of a high-fat diet affects neural stem and progenitor cells in the area postrema and central canal more than four weeks does, raises astrocyte density in the solitary and dorsal motor nuclei and microglial density in the area postrema and solitary nucleus, and that longer exposure produces mild medullary inflammation. Male mice; descriptive histology.

67. 67
    Ujhelyi A, Kucharz RS, Rekling JC. Cholecystokinin induces Ca2+ transients in brainstem neurons, astrocytes and central vagal terminals in jugular/nodose ganglia-brainstem slice cultures.Eur J Neurosci · 2026 · 64(5):e70686doi:10.1111/ejn.70686 · PMID 42723213

    Introduces a slice culture that keeps the jugular and nodose ganglia attached to the brainstem through intact vagal rootlets, and shows that cholecystokinin and substance P evoke calcium waves in brainstem astrocytes while muscarine and thyrotropin-releasing hormone activate neurons without recruiting astrocytes. A preparation paper; the selectivity is in vitro.

68. 68
    Geisler CE, Applebey SV, Chichura KS, et al. Sex differences in central endozepine expression and regulation of appetite.Physiol Behav · 2025 · 299:115003doi:10.1016/j.physbeh.2025.115003 · PMID 40553808

    Female rats carried more diazepam-binding inhibitor protein throughout the dorsal vagal complex and were more sensitive to the anorectic effect of octadecaneuropeptide than males, on both chow and high-fat diet. Rat; a sex difference in a system studied almost entirely in males.

69. 69
    Castellanos EA, Schwerdtfeger LA, Smith BN, Tobet SA. Microglia alter autonomic nucleus neuronal activation after peripheral cytokine challenge.Am J Physiol Cell Physiol · 2026 · 330(3):C683-C694doi:10.1152/ajpcell.00810.2025 · PMID 41660994

    A single intraperitoneal injection of tumor necrosis factor alpha raised c-Fos counts in the dorsal vagal complex and ventrolateral medulla of mice, concentrated lateral to the area postrema, and minocycline into the fourth ventricle reduced both microglial staining and that activation. Mouse; one cytokine, two hours.

70. 70
    Hermann A, Koppl T, Wagner A, Shojaei A, Wohlmuth B, Aydin R, Cyron CJ, Miftahof R. A 3D-1D-0D multiscale model of the neuro-glial-vascular unit for synaptic and vascular dynamics in the dorsal vagal complex.J Math Biol · 2025 · 92(1):3doi:10.1007/s00285-025-02317-7 · PMID 41324715

    A computational framework coupling neurons, astrocytes, smooth muscle and a vascular network in this region. It generates simulations, not measurements, and should be treated as a way of organizing hypotheses about neurovascular coupling here.

71. 71
    Brown JM, Bentsen MA, Jorgensen AM, et al. Single-cell profiling identifies obesity-disrupted brain glucose sensing.Nat Commun · 2026 · 17(1)doi:10.1038/s41467-026-77116-9 · PMID 42811021

    75,787 nuclei from arcuate nucleus, ventromedial hypothalamus and dorsal vagal complex of control-diet and high-fat-diet mice after intravenous glucose or saline; the conserved astrocyte and oligodendrocyte response to hyperglycemia was largely absent or reversed after high-fat feeding. Male mice; a transcriptional snapshot, not a functional test.

72. 72
    Lopez-Cruz A, Figueredo Burgos NS, Hakimi AM, et al. Functional segregation of body-brain signals in the area postrema.bioRxiv · 2026 · preprint 2026.06.15.732473doi:10.64898/2026.06.15.732473 · PMID 42453677

    Preprint, not peer reviewed. Reports that GFRAL neurons are activated by dietary fat independently of GDF15 and are required for fat satiation, and that GIP receptor neurons which inhibit them are activated by sugar. If confirmed it would give the nausea receptor an ordinary nutritional role; treat as provisional.

73. 73
    Beyazgul S, Laleh SS. Polyvagal theory and neonatal sleep: a review of autonomic regulation and implications for sudden infant death syndrome.Int J Dev Neurosci · 2025 · 85(6):e70050doi:10.1002/jdn.70050 · PMID 40898438

    Cited only as evidence of how far the framework has spread into clinical fields. It is a narrative review that summarizes associations between vagal activity and neonatal sleep and proposes the polyvagal frame as a lens; it presents no new data and its own conclusion is that further research is warranted.

74. 74
    Roelofs K. Freeze for action: neurobiological mechanisms in animal and human freezing.Philos Trans R Soc Lond B Biol Sci · 2017 · 372(1718):20160206doi:10.1098/rstb.2016.0206 · PMID 28242739

    Review describing freezing as behavioral inhibition with parasympathetically dominated heart-rate deceleration, dependent in rodents on amygdala projections to the periaqueductal gray. Cited for the reality of the phenomenon, not for any brainstem assignment; a review.

75. 75
    Kozlowska K, Walker P, McLean L, Carrive P. Fear and the defense cascade: clinical implications and management.Harv Rev Psychiatry · 2015 · 23(4):263-87doi:10.1097/HRP.0000000000000065 · PMID 26062169

    Clinical review of the defense cascade from arousal through fight, flight and freezing to tonic and collapsed immobility as responses of last resort, mediated through the amygdala, hypothalamus, periaqueductal gray and both sympathetic and vagal nuclei. A review, and a different account of collapse from the polyvagal one.

76. 76
    Meister AL, Jiang Y, Doheny KK, Travagli RA. Correlation between the motility of the proximal antrum and the high-frequency power of heart rate variability in freely moving rats.Neurogastroenterol Motil · 2019 · 31(8):e13633doi:10.1111/nmo.13633 · PMID 31119854

    Simultaneous antral strain-gauge and electrocardiogram recording in rats. Found a positive correlation at baseline and after cholecystokinin but not after ghrelin or food, and states explicitly that cardiac and gastrointestinal vagal outputs arise from the nucleus ambiguus and the dorsal motor nucleus respectively. Small groups of 6 to 8 animals per condition.

77. 77
    McIntyre KM, Sloan RP. Perspective on self-regulatory capacity and the vagus nerve: have we wandered too far?Biol Psychol · 2025 · 200:109106doi:10.1016/j.biopsycho.2025.109106 · PMID 40819735

    Critical review of work treating vagal activity as a readout of self-regulation, under both the polyvagal theory and the neurovisceral integration model, describing confusion around constructs, misleading claims about the relation of high-frequency heart rate variability to vagal activity, and errors in neuroanatomy. A review from a group independent of Grossman's.

78. 78
    Bambha VP, Franklin BE, Sahoo M, et al. Toward a biologically grounded understanding of autonomic function in developmental science.Neurosci Biobehav Rev · 2026 · 191:106984doi:10.1016/j.neubiorev.2026.106984 · PMID 42749237

    Surveys 316 developmental studies published over six and a half years that use respiratory sinus arrhythmia, and finds their psychological interpretations resting on the century-old binary model of the autonomic nervous system. A review with a literature survey; it generates no new physiological measurement.

79. 79
    Machhada A, Ang R, Ackland GL, et al. Control of ventricular excitability by neurons of the dorsal motor nucleus of the vagus nerve.Heart Rhythm · 2015 · 12(11):2285-93doi:10.1016/j.hrthm.2015.06.005 · PMID 26051529

    In anesthetized rats, silencing dorsal vagal motor nucleus neurons shortened the ventricular effective refractory period, lowered the threshold for triggered ventricular tachycardia and prolonged the corrected QT interval, evidence of tonic parasympathetic control of ventricular excitability from this nucleus. Rat, with a mouse synuclein model; about the ventricles, not heart rate.

80. 80
    Machhada A, Marina N, Korsak A, Stuckey DJ, Lythgoe MF, Gourine AV. Origins of the vagal drive controlling left ventricular contractility.J Physiol · 2016 · 594(14):4017-30doi:10.1113/JP270984 · PMID 26940639

    Silencing dorsal vagal motor nucleus neurons increased left ventricular contractility in anesthetized rats, and mapping placed the responsible cells in the caudal part of the left nucleus. Rat; anesthetic-dependent, as the authors note.

81. 81
    Machhada A, Trapp S, Marina N, et al. Vagal determinants of exercise capacity.Nat Commun · 2017 · 8:15097doi:10.1038/ncomms15097 · PMID 28516907

    Genetic silencing of the dorsal vagal motor nucleus population dramatically impaired exercise capacity in rats, while optogenetic recruitment enhanced cardiac contractility and prolonged exercise endurance. Direct evidence that this nucleus supports cardiac performance rather than producing collapse. Rat.

82. 82
    Iwashyna TJ, Ely EW, Smith DM, Langa KM. Long-term cognitive impairment and functional disability among survivors of severe sepsis.JAMA · 2010 · 304(16):1787-94doi:10.1001/jama.2010.1553 · PMID 20978258

    Prospective cohort from a nationally representative US survey: among 516 survivors of severe sepsis, moderate to severe cognitive impairment rose by 10.6 percentage points (odds ratio 3.34, 95 percent confidence interval 1.53 to 7.25) and new functional limitations averaged 1.57 in those with none before, with declines persisting at least 8 years. Observational and in an older population, mean age 76.9 years; it establishes the human burden, not its mechanism.

83. 83
    Cheng S, Dowsett GKC, Rainbow K, Norton M, Roberts AG, Phuah P, Bewick GA, Lam BYH, Yeo GSH, Murphy KG. NodoMap: a single-cell and spatial transcriptomic atlas of the mouse nodose ganglion.Cell Press Blue · 2026 · 1(4):100072doi:10.1016/j.cpblue.2026.100072 · PMID 42483274

    106,436 cells and 21 neuronal subtypes in the mouse nodose ganglion, with spatial mapping and left-right differences. Cited here as the resource that makes the afferent side of the complex tractable; it is mouse and it describes the ganglion, not the brainstem.