Anatomy and function of efferent and afferent vagal innervation of the stomach

Abstract

The stomach is one of the major targets of the vagus nerve. Vagal cholinergic parasympathetic (efferent) neurons are located in the dorsal motor nucleus of the vagus (DMV) in the caudal brainstem. They do not directly innervate gastric effector cells, but instead a large portion of enteric neurons located in the myenteric plexus, which act as muscle motor neurons and function-specific pattern generators that can regulate specific motility and secretion programs in coordination with physiological needs and environmental conditions. Vagal sensory (afferent) neurons located in the nodose ganglia (NG) innervate all layers of the stomach wall, where they pick up mechanical and chemical signals to inform the brain and effectuate vago-vagal reflexes. The most abundant terminal structures, intraganglionic laminar endings (IGLEs), are in intimate contact with myenteric plexus neurons. Besides sensing gastric tension, they are likely also chemosensors and provide a link between the enteric nervous system and the brain. Transcriptomic analyses of vagal efferent and afferent neurons allowed making great progress on deciphering their respective coding logic, with a predominantly labeled line organization emerging as the general principle. Because of their distinct molecular fingerprints, separate populations of vagal efferents and afferents with unique morphologies and functions have become experimentally and therapeutically accessible. However, the full coding logic also depends on the organization of the peripheral interface with the enteric nervous system, as well as the central interface represented by 2^nd and higher order sensory neurons and pre-autonomic neurons upstream of the DMV, which are only starting to be understood.

Graphical abstract

1. Introduction

The term “gut brain axis” is often used to describe the bidirectional communication between the central nervous system and the gut. Its popularity is illustrated by the fact that more than half of the 12,000 papers using gut brain axis as key word indexed on PubMed have been published in the last three years. The term was originally used in connection with studies on gastrointestinal functions and the physiological determinants of ingestive behavior - the unprecedented rise in popularity was fueled by the re-discovery of the gut microbiome (Cryan et al., 2020) and the ascent of neuromodulation as a therapeutic tool (Aljeradat et al., 2024). Gut brain communication is also an important component in the new-found enthusiasm for interoception (Bonaz et al., 2021; Ceunen et al., 2016; Craig, 2002) and has been implicated in a large spectrum of diseases including obesity, diabetes, functional gastrointestinal disorders (Drossman and Hasler, 2016), Parkinson, dementia, schizophrenia, and autism (Cheshire, 2012; Cryan et al., 2020).

This non-systematic review focuses on a limited domain of gut-brain communication, the functional anatomy of vagal innervation of the stomach. It summarizes recent progress in deciphering the anatomical and functional coding logic of both vagal motor fibers to the stomach and vagal sensory fibers from the stomach. It explores the relationship of vagal innervation of the stomach with the enteric nervous system and potential central neural pathways responsible for homeostasis, allostasis, and interoception. For the rich literature on more physiological aspects of the neural control of the stomach, the reader is referred to several informative reviews (Browning and Travagli, 2014; Browning et al., 2026; Furness et al., 2026; Sharkey and Mawe, 2023).

General Principles and pathways of gut-brain communication

Stomach and brain communicate with each other via neural and humoral pathways (Fig. 1). Neural pathways are important for rapid and discrete communication, while humoral signaling pathways are slower but with widespread systemic effects. Neural signaling from the stomach to the brain is accomplished by vagal and spinal afferents. Vagal afferents consist of pseudounipolar neurons with cell bodies located bilaterally in the nodose ganglia that send peripheral axons through the vagus nerve to the stomach and central axons to the nucleus tractus solitarii in the brainstem. Spinal afferents consist of neurons located in the (T3-T13) dorsal root ganglia that send peripheral axons through the splanchnic nerves via the celiac ganglion to the stomach and central dendrites to the dorsal horn of the thoracic spinal cord, where they synapse on brain-projecting dorsal horn neurons (Munzberg et al., 2023).

Figure 1. Schematic diagram showing neural and humoral communication pathways between the stomach and the brain.

See text for details. Abbreviations: BBB, blood brain barrier; CVOs, circumventricular organs; DMV, dorsal motor nucleus of the vagus; NTS, nucleus tractus solitarii; IGLEs, intraganglionic laminar endings; IMAs, intramuscular arrays; MucE, mucosal endings.

Neural signaling from the brain to the stomach is accomplished by classic parasympathetic and sympathetic nervous system pathways, with parasympathetic (vagal) preganglionic neurons located in the dorsal motor nucleus of the vagus (DMV) directly projecting to gastric enteric neurons, and sympathetic preganglionic neurons located in the intermediolateral column of the thoracic spinal cord projecting via the splanchnic nerves and postganglionic neurons in the celiac ganglia to the stomach. Adhering to long-standing autonomic nomenclature (Loewy, 1990; Powley, 2012; Saper, 2002), we define vagal postganglionic neurons as organ-intrinsic neurons receiving direct vagal preganglionic input, which happen to be enteric neurons in the stomach. However, these neurons are integral parts of the ENS, with their own inputs, outputs, and functions.

Stomach and brain can also communicate with each other via humoral signaling. Ghrelin is the best known gastric humoral signal to the brain, although ghrelin can also signal via vagal afferents to the brain (Date et al., 2002). but there are likely other humoral factors produced by the stomach, such as cytokines and miRNAs (DelValle, 1997). In return, the brain can also affect gastric function via the HPA axis (Nardone and Compare, 2014).

2. Anatomy and functions of the stomach

2.1. History and major functions

The importance of the stomach was recognized long ago. Aristoteles (384-322 BC) recognized its function at the start of digestion (“De partibus animalium”) and Galenus (129-210 CE) described “vertical” (longitudinal) and “horizontal” (circular) muscle fibers in its wall which enabled for specific motility (May, 1968). He was also aware of the significance of the vagus nerve, both afferent and efferent, for gastric functions (Kuhn, 1881).

The major functions of the stomach are: 1) storage of large meals, 2) churning and trituration to mechanically break down food particles, 3) secretion of gastric acid and enzymes to chemically digest food, 4) regulated, slow release of chyme (predigested food) into the small intestines, 5) generating signals that reach the brain and other organs to control food intake, and 6) early defense of harmful microbes and toxins. Each of these functions critically depends on neural mechanisms and pathways.

2. 2. Anatomy of the stomach and its sphincters

There are large variations in the gross anatomy and functional subunits (non-glandular and glandular) of the vertebrate and even mammalian stomach. Among omnivores such as rodents, pigs and humans, these differences are less pronounced. Here we consider mainly stomach innervation in mice, rats, guinea pigs and humans.

The greater part of the stomach which consists of fundus, corpus, and antrum lies in the left upper abdomen, with the fundus nestling against the left cupula of the diaphragm, while the pylorus is found slightly right of the midline (Fig. 2). From the esophago-gastric junction, also referred to as cardia, the greater curvature extends around the fundus and continues along the left caudal contour, the lesser curvature along the right cranial contour to the pylorus. The gastric arteries, derived from the celiac trunk or its branches, accompanying veins and regional lymphatics follow these curvatures: right and left gastric arteries in the lesser, right and left gastroepiploic arteries in the greater curvature. Short branches from the splenic artery supply the lateral portions of the fundus. Accompanying veins ultimately drain into the portal vein. The ventral and dorsal surfaces of the stomach are covered with visceral peritoneum which enables frictionless gliding against the ventral abdominal wall and visceral surface of the liver on one hand, and the ventral surface of the pancreas and the retroperitoneum on the other. Peritoneal duplicatures connect the stomach with the liver (lesser omentum), the spleen (gastro-splenic ligament), the diaphragm (gastrophrenic ligament) and in human the transverse colon (greater omentum and gastrocolic ligament).

Fig. 2. Vagal access pathways to the stomach and neighboring organs.

Semi-schematic ventral view loosely based on rodent and human dissection studies, showing vagus nerve branches and subbranches (yellow), the ventral surface of the stomach and distal esophagus (gray), portions of the liver (purplish), pancreas (beige), gall bladder and major bile ducts (green), hepatic portal vein (blue) and vena cava (dark blue), as well as the major arteries originating from the abdominal aorta (brown) on the bottom.

The smooth muscle tunica muscularis consists of an outer longitudinal and inner circular layer, the ganglionated myenteric plexus sandwiched between them. The fibrae obliquae represent a third inner incomplete layer forming a sling around the left circumference of the esophago-gastric junction and extending their branches towards the antrum. The sling fibers are contiguous with the circular muscle fibers of the abdominal esophagus, forming, together with clasp fibers on the right circumference, the lower esophageal sphincter (LES)(Liebermann-Meffert et al., 1979). The pyloric sphincter between the antrum and the duodenal bulb represents a thickened circular muscle layer. The submucosa accommodates shearing forces between the muscularis and mucosa and distributes blood vessels to the latter (Williams, 1980). The submucosal plexus is almost devoid of ganglia in rodents but contains small ganglia in the human stomach (Furness, 2006).

In human, 80 percent of the gastric mucosa are of the oxyntic, hydrochloric acid and pepsin secreting, type and 20 percent of the mucus-producing antral type. In rodents the oxyntic mucosa is only found distal to the limiting ridge; thus, the fundus and part of the corpus are lined with squamous epithelium. In human, the cardia region is equipped with a special mucosa of mucus producing cells. Besides acidproducing oxyntic (parietal) cells, pepsinogen-secreting chief cells and mucus-producing cells, the oxyntic mucosa harbors histamine-producing enterochromaffin-like (ECL) cells while gastrin-producing G-cells are found in the antral mucosa. Both types of mucosa harbor somatostatin-producing D-cells and a variety of entero-endocrine (EE) cells producing serotonin and several peptides, among them cholecystokinin, ghrelin and atrial natriuretic peptide. These local endocrine/paracrine cells are neurally coordinated and act as a network to stimulate (histamine, gastrin) or inhibit (somatostatin) acid production (for review see (Schubert and Rehfeld, 2019).

2. 3. Gastric pathologies and surgeries

Besides functional gastric disorders such as gastroparesis which will be dealt with in another chapter of this special issue, structural pathologies, most important ulcers and malignancies, are the domain of gastric surgery (reviews in (Gillison, 2007). Theodor Billroth pioneered gastric surgery with his famous B I (1881) and B II (1885) resections of the distal half to two-thirds of the stomach. These were originally conducted in carcinoma patients but became also standard procedures for treatment of gastric and duodenal ulcers. The B II resection including a gastro-jejunostomy was later modified, e.g., by Roux’s Y-anastomosis. The idea that gastric hypersecretion promoted development of ulcers and the notion that vagotomy reduced secretion of acidic gastric juice prompted vagotomy for treatment of gastric ulcers (review in (Gillison, 2007).

Vagotomy with its various refinements, for example selective proximal vagotomy (Holle et al., 1972) sparing the vagal branch to the antrum (Latarjet’s nerve) was very popular during the 1970s and 1980s but almost abandoned after the discovery of Helicobacter pylori and its key role in the pathogenesis of gastroduodenal ulcers, gastritis and gastric malignomas (Marshall and Warren, 1984). However, more or less extensive gastrectomy is still the therapy of choice in gastric cancer. A special indication for surgery are stenosing processes of the pylorus, in particular congenital pyloric stenosis. Various techniques of pyloroplasty were developed since Weber-Ramstedt’s myotomy in pylorospasm (1912) which is sometimes required upon vagotomy with lesion of antral branches (Holle et al., 1972).

The repertoire of classical gastric surgery, also in laparoscopic technique, enjoyed a revival with the advent of bariatric surgery. In particular, Roux’s Y anastomosis, gastric sleeve resection and gastric bypass strategies are employed for this indication (for reviews see (Mason, 2005); Gillison, 2007 #1454;Shilton, 2025 #1429;Sundbom, 2014 #1430}.

3. Vagal efferent innervation of the stomach

3.1. Gross anatomy

After penetrating the diaphragm, the ventral/anterior/left and dorsal/posterior/right vagal trunks continue to descend towards the gastric cardia along the distal esophagus (Fig. 2). In human, a single ventral and dorsal trunk is found in 27 and 40 percent, respectively; in the remaining individuals, up to six fascicles were counted (Loeweneck, 1993).

The anterior (ventral) subdiaphragmatic trunk, which is immediately visible in rodents when opening the peritoneal wall with a ventral approach, first gives off the common hepatic branch to the right, which in rodents travels along the hepatoesophageal artery towards the hepatic artery proper (Fig. 2). In human, the common hepatic branch is embedded between the leaflets of the lesser omentum. At the hepatic artery proper, the common hepatic branch divides into the gastroduodenal and hepatic proper branches and mingles with the periarterial plexus surrounding the hepatic artery (Berthoud et al., 1992). The larger of these two subbranches, the gastroduodenal branch further divides into sub-branches that travel along the right gastric, right gastroepiploic, and duodeno-pancreatic arteries to innervate parts of the antrum, pylorus, pancreas, and proximal duodenum (Berthoud et al., 1991a; Berthoud et al., 1991b; Donahue et al., 1988; Kong et al., 2019; Wang and Powley, 2007). The hepatic branch proper innervates the hepatic portal vein, gall bladder, major bile ducts, and the hepatoportal space (Berthoud et al., 1992). In addition, the rat common hepatic branch provides some sensory fibers to the fundic part of the stomach (Phillips et al., 1997).

Slightly more distal to the common hepatic branch point, the anterior trunk gives off the anterior/accessory celiac branch which wraps around the esophagus to join the dorsal celiac branch and travels along the left gastric, common hepatic, and celiac arteries (against the direction of blood flow) towards the celiac ganglia and superior mesenteric artery (Wang and Powley, 2007). In humans, the subdiaphragmatic vagi only become visible after lifting up the left lobe of the liver beneath the esophageal serosa.

The anterior gastric branch continues across the lower esophageal sphincter on the anterior/ventral side of the stomach towards the pylorus, staying close to the lesser curvature and giving off numerous branches to the fundus, corpus, and antrum. In humans, this distribution is organized into 3-5 major branches that further split into numerous smaller branches at the level of the fundus, corpus and antrum. Specifically, the most distal major branch, the nerve of Latarjet, further divides into sub-branches supplying the corpus and antrum, the latter also called Crow’s foot (Loeweneck, 1993) (Fig. 2).

The posterior (dorsal) subdiaphragmatic trunk, which on a ventral approach becomes visible after turning the stomach and distal esophagus slightly to the left, splits into a posterior celiac branch which, together with the anterior celiac branch, travels along the left gastric artery towards the celiac ganglia and superior mesenteric artery and its ganglion and plexus where they intermingle with sympathetic fibers originating in these ganglia (Berthoud et al., 1991a; Wang and Powley, 2007). In the rat and mouse, this split is often embedded in a fat pad that needs to be carefully dissected for clear visualization.

The posterior gastric branch continues to descend near the lesser curvature of the stomach and supplies the dorsal surface of the stomach in a branching pattern similar to the anterior gastric branch.

The lower esophageal sphincter is innervated by fine branches originating from both the anterior and posterior gastric branches (Loeweneck, 1993).

In general, nerves (vagal and non-vagal) are closely following blood vessels (Boekelaar et al., 1985; Wang and Powley, 2007). For the supply of the stomach with sympathetic and dorsal root spinal afferent fibers that project through the celiac ganglia, the paths thus lead from the celiac arterial trunk either via 1) left gastric 2) common hepatic → right gastric, 3) common hepatic → gastroduodenal, 4) common hepatic → gastroduodenal → right gastroepiploic, or 5) splenic artery → left gastroepiploic arteries (Fig. 2). However, for the vagal supply, the entry point is not the celiac artery, but the left gastric artery (anterior and posterior gastric branches) and common hepatic artery (common hepatic branch).

3. 2. Microscopic anatomy of vagal motor fibers and terminals

Based on electron microscopic analyses in rats, the anterior and posterior subdiaphragmatic trunks contain a total of ~8,000 and ~10,000 fibers, with ~30% efferents and ~70% afferents (Prechtl and Powley, 1990). In humans the number of fibers in the abdominal vagi is highly variable and roughly 10 times higher than in the rat (Havton et al., 2021). In the subdiaphragmatic vagus, more than 95 percent of axons are unmyelinated (99.5% in rat (Prechtl and Powley, 1990); 97% in human (Havton et al., 2021), encompassing both preganglionic efferents and visceral afferents.

A majority of subdiaphragmatic vagal fibers innervate the stomach with the remainder innervating the small and large intestines, pancreas, bile system, and hepato-portal space. After the major division points near the lesser curvature, the gastric subbranches penetrate the longitudinal muscle layer to profusely innervate the entire myenteric plexus that extends throughout the stomach.

Before the ascent of modern neural tracing techniques, studies relying on a staining technique for degenerating nerve fibers showed such fibers were present in all layers of the stomach wall 2 days after vagotomy in cats (Feher and Vajda, 1983). Anterograde tracing of vagal preganglionics with injections of suitable tracers such as the carbocyanine dyes DiI and DiA, Biotinylated dextran-amine (BDA), or horseradish peroxidase wheat-germ agglutinin (WGA-HRP) into the DMV of rats and Cre-dependent AAV9-CAG-FLEX-tdTomato in Chat-IRES-Cre mice showed that the target of innervation are enteric neurons primarily in the myenteric ganglia and to a much lesser extent in submucosal ganglia (Berthoud et al., 1991b; Berthoud et al., 1990; Holst et al., 1997; Tao et al., 2021). Vagal efferent neurons do thus not directly innervate gastric and intestinal effector cells and tissues such as smooth muscle and glands (Berthoud et al., 2025). After entering the myenteric plexus, vagal preganglionic axons typically course through numerous ganglia and frequently collateralize. In well-labeled cases, all gastric myenteric ganglia and neurons were supplied by varicose vagal axons and highly varicose terminals often wrap around individual enteric neurons (Berthoud et al., 1991b; Berthoud et al., 1990; Holst et al., 1997; Tao et al., 2021). (Fig. 3). The rat gastric myenteric plexus is estimated to consist of about 200,000 to 400,000 neurons in tens of thousands of ganglia (Di Natale et al., 2022; Jarvinen et al., 1999; Phillips et al., 2004). Given that each of the rat gastric branches contains approximately 2000 efferent axons (Prechtl and Powley, 1990), this is suggesting that each vagal preganglionic axon supplies as many as 20 myenteric ganglia.

Fig. 3. Vagal preganglionic efferents from the DMV innervate enteric neurons in the stomach.

A: Anterogradely labeled vagal preganglionic efferents (brown) make close contacts with nearly all enteric neurons (cuprolinic blue) in the myenteric plexus of the rat stomach (Adapted from (Powley et al., 2019) with permission).

B: Higher power image showing varicose vagal efferent axons (brown) surrounding Nos1+ enteric neurons in the rat gastric myenteric plexus (Adapted from (Powley et al., 2019) with permission).

C: Triple labeling using dextran-Texas Red anterograde tracing of vagal efferents (red) combined with alpha-synuclein (green) and NOS (blue) immunohistochemical labeling of myenteric neurons in rat stomach. Note yellow varicosities indicating co-localization of alpha-synuclein with dextran-Texas Red in putative vagal efferent terminals surrounding NOS+ neurons (Adapted from (Walter et al., 2009) with permission).

D: Cervical vagal stimulation-induced cFos expression (red) in NOS+ (white) and NOS− myenteric plexus neurons in rat stomach (Adapted from (Zheng and Berthoud, 2000) with permission.

E: tdTomato (purple) labeled vagal preganglionic terminals (purple) surrounding NOS+ neurons in the myenteric plexus of mouse stomach (Adapted from (Tao et al., 2021) with permission)

The excitatory character of vagal input to gastric myenteric neurons was demonstrated with electrophysiological recording (Schemann and Grundy, 1992) and with c-Fos immunohistochemistry following electrical stimulation of the cut peripheral end of gastric vagal branches in guinea pigs or the cervical vagus nerve in rats (Zheng and Berthoud, 2000), respectively. (Fig. 3D) These studies demonstrated that the proportion of enteric neurons receiving excitatory functional vagal input was nearly 100%, at least near the lesser curvature (Schemann and Grundy, 1992).

3. 3. Functions of gastric vagal efferents and the vagal-enteric interface

Studying effects of vagal stimulation on gastric functions has a long history. For example, electrically stimulating the peripheral end of the cut cervical vagus, Langley noted that: “Usually the primary effect of vagus stimulation whether anesthetics alone were given, or these with curari, or atropine, or with both, was contraction. But at times distinct dilation of the sphincter was obtained. I conclude then that the body of the stomach and the pylorus receive inhibitory as well as motor fibres from the vagus” (Langley, 1921; Langley, 1898)

Given that vagal preganglionics exclusively innervate enteric neurons, understanding the downstream connectivity and functions of specific enteric neurons is one key component to comprehend how the efferent vagus affects gastric functions. In his “Command Neuron” hypothesis, JD Wood proposed that “local integrative circuits of the ENS are organized for program operations independent of input from the CNS” and that “subsets of neural circuits are preprogrammed for control of distinct patterns of gut behavior in each effector system and for the coordination of activity of multiple systems” (Wood et al., 1999). Although intuitively attractive, there is little experimental evidence for this hypothesis. Wood further argued that “because vagal efferents are relatively sparse compared to the many enteric neurons, rather than controlling individual enteric motor neurons, messages transmitted by parasympathetic efferent fibers are command signals for the activation of expanded blocks of integrated circuits positioned in the gut wall” (Wood et al., 1999), and in his Fig. 3, it is suggested that only a few specialized interneurons receive direct vagal efferent input. This interpretation was supported at the time by an anterograde labeling study that found very sparse vagal efferents in the gastric myenteric plexus (Kirchgessner and Gershon, 1989). However, subsequent studies demonstrating a much more widespread input achieved by massive collateralization, particularly in the stomach (Berthoud et al., 1990; Powley et al., 2019; Schemann and Grundy, 1992; Tao et al., 2021) rather suggests that most types of enteric motor neurons receive vagal efferent input. The idea that vagal efferent inputs modulate intrinsic gastric motor patterns and programs rather than control specific functions is still valid, but how this is achieved must lie in the uniqueness of vagal preganglionic neurons regarding their peripheral collateralization pattern and/or differential central inputs, rather than in the uniqueness of innervated neurons (Berthoud et al., 1990; Powley et al., 2019; Schemann and Grundy, 1992; Tao et al., 2021) (Fig. 4).

Fig. 4. The vagal-enteric interface.

Schematic diagram showing peripheral interface between vagal efferent neurons from the DMV in the brain and the gastric enteric nervous system. All major types of motor neurons arising from gastric myenteric ganglia (excitatory [Ach] and inhibitory [NOS] muscle motor neurons; neurons innervating mucosal effectors (parietal cells [Ach], chief cells [GRP], gastrin cells [GRP] and somatostatin cells connected to the gastrin cells) are innervated by vagal preganglionic terminals. The excitatory and inhibitory muscle motor neurons are thereby innervated by two distinct labeled lines. Whether the neurons innervating the mucosa and the excitatory and inhibitory muscle motor neurons in the lower esophageal and pyloric sphincters and the vasodilator neurons in the submucosal plexus are also innervated by distinct labeled lines of vagal efferents is not yet known. (Modified after (Furness et al.,2020)

Enteric motor neurons in the gastric myenteric plexus have been characterized by their morphology (Anetsberger et al., 2018; Brehmer, 2021; Brookes et al., 1998), electrophysiological properties (Carbone et al., 2014; Schemann et al., 2001), transmitter/peptide immunohistochemistry (Anetsberger et al., 2018; Beck et al., 2009; Furness et al., 1991; Tonini et al., 2000), and most recently, their transcriptomic profile (Millett et al., 2025; Shi et al., 2025)(Kim & Lowell, unpublished). Briefly, they include two types of muscle motor neurons, excitatory (ChAT+) and inhibitory (NOS+), four types of neurons innervating mucosal effector cells (parietal cells, chief cells, gastrin cells and somatostatin cells) expressing various combinations of transmitters and peptides (ChAT, VIP, GRP, ENK, TAC), as well as vasodilator neurons (Costa et al., 2000; Furness et al., 2020) (Fig. 4). Importantly. there is evidence that each of these subtypes is innervated by vagal preganglionic neurons from the DMV. Anatomical evidence is based on experiments using immunohistochemistry for neurons expressing cholinergic, nitrergic, vasoactive intestinal peptide (VIP)-ergic, and gastrin releasing peptide (GRP)-ergic phenotypes combined with anterograde tracing (Berthoud, 1995; 1996; Jarvinen et al., 1999; Tao et al., 2021) or c-Fos activation following electrical stimulation of the vagus nerve (Berthoud et al., 2001b). By using a sparse anterograde labeling strategy, it was possible to demonstrate that individual vagal preganglionic axons entering the myenteric plexus either selectively innervate Nos1+ or Nos1− neurons (Jaffey et al., 2023). Selective input to nitrergic and cholinergic enteric neurons by different populations of vagal motor neurons was also suggested by a number of studies measuring gastric motility patterns after stimulation of the vagus nerve or the DMV (for review see (Travagli and Anselmi, 2016).

Early immunohistochemical studies demonstrated the presence of neurons and fibers staining for GABA (Jessen et al., 1986) and GABA-transporters (Fletcher et al., 2002) in the intestinal myenteric plexus of guinea pigs and rats, as well as glutamatergic neurons in the gastric myenteric plexus (as reviewed in Tsai (Tsai, 2005). Similarly to what has recently been reported in the mouse small and large intestines (Hamnett et al., 2025; Liu et al., 2024) some of us, using genetics-guided viral approaches confirmed the presence of a few glutamatergic neurons with long projections and also GABAergic neurons in the gastric myenteric plexus (Kim and Lowell, unpublished), and, although not confirmed, it is likely that these neurons also receive direct inputs from vagal preganglionics.

Transcriptomic analysis of murine cholinergic DMV neurons confirmed the existence of two separate populations of DMV neurons exclusively innervating either gastric Nos-1 or cholinergic myenteric plexus neurons in the murine glandular stomach, with DMV neurons innervating enteric ChAT neurons expressing the molecular marker CCK, while neurons innervating enteric Nos-1 neurons expressing prodynorphin (Tao et al., 2021). A subsequent study with more refined transcriptomic analysis of the DMV and molecular marker-based anterograde tracing revealed at least four additional molecularly distinct subpopulations of vagal preganglionic efferents with projections to the stomach (Kim & Lowell unpublished). These six stomach-innervating DMV subtypes are not only distinguished by their transcriptomic profiles but also by their spatial distributions along the rostro-caudal and medio-lateral axes of the DMV and extent of stomach innervation, with many of them displaying preferential innervation of the proximal or distal stomach. With the advent of single-cell transcriptomics, molecular identities of subpopulations of enteric neurons regulating distinct gastric motor and secretory functions are also being revealed (Millett et al., 2025; Shi et al., 2025) (Kim & Lowell unpublished). There is now growing evidence that each molecularly distinct population of vagal preganglionic efferents engage a defined group of function-specific enteric neurons, supporting the existence of parallel, labeled-line vagal efferent pathways that control dedicated gastric functions. Molecular identities of vagal preganglionics that regulate lower esophageal and pyloric sphincter relaxations have also been identified (Kim & Lowell unpublished).

Mathematical modeling approaches to the neural control of gastric motor functions also conclude that assuming a coordinated interplay among extrinsic inputs via the vagus and sympathetic nervous system, the enteric nervous system, Interstitial Cells of Cajal (ICC), and smooth muscle cells best fits observed gastric motility behavior (Fernandes and Kothare, 2025). Interestingly, generating spatio-temporal motility maps from video recordings of the in vitro perfused guinea-pig stomach with an intact vagal nerve supply showed that vagal stimulation triggered anally propagating antral contractions and enlargement of the fundus. While the strength of contractions and relaxation responses were frequency-dependent, the interval between contractions remained the same 5-6/min typical for gastric slow waves, indicating that vagal input impinges on intrinsic enteric neural circuits that have a modulatory role in the myogenic mechanism underlying slow-wave peristalsis, rather than directly on gastric musculature (Berthoud et al., 2002).

The other main function of vagal input is gastric acid secretion. Vagal efferents in the two gastric branches are most important for gastric acid secretion in the rat, but subbranches of the common hepatic branch innervating the antrum via the right gastric and gastroepiploic arteries also make a small but significant contribution (Berthoud et al., 1991a; Berthoud et al., 1986). Are vagal efferents stimulating gastric acid secretion distinct from vagal efferents with motor effects? Studies with electrical vagal stimulation in cats found that based on stimulation thresholds, different nerve fiber groups must be mediating a) enhanced gastric motility and b) secretion of hydrochloric acid and pepsinogen, as well as vasodilatation and inhibition of motility (Martinson, 1965).

The latest genetics-based approaches showed that selective activation of gastric myenteric Grp+/Vip+, but not Grp+/Tac1+/Penk+ neurons increased plasma gastrin levels and lowered stomach pH in mice, suggesting the presence of functionally distinct subpopulations (Millett et al., 2025). Thus, vagal inputs to VIP+ and VIP+/GRP+ enteric neurons (Berthoud, 1996) are most likely responsible for the stimulation of gastric acid and enzyme secretion via parietal, chief, gastrin, and somatostatin cells. VIP-positive nerve fibers were also found near ghrelin secreting cells (Hunne et al., 2019) but there are conflicting reports about efferent vagal effects on ghrelin secretion (Veedfald et al., 2018). Additional studies in Cre-mouse lines that allow selective stimulation or silencing of specific populations of vagal preganglionics should clarify these issues.

Vagal efferents controlling the pyloric sphincter have been recently characterized with focal chemical stimulation in the DMV and retrograde PRV tracing (Richardson et al., 2023). They found that glutamatergic stimulation in the rostral DMV results in contraction, while stimulation in the caudal DMV resulted in relaxation of the pylorus, suggesting that separate vagal preganglionics are involved. The molecular identities of vagal preganglionics that regulate pyloric and lower esophageal sphincter functions have recently been identified (Kim & Lowell unpublished).

4. Vagal sensory innervation of the stomach

Vagal sensory neurons are pseudounipolar neurons with the cell body located in the bilateral nodose ganglia located just below the scull base. It is estimated that the stomach is innervated by up to 8,000 vagal sensory fibers in the rat (Prechtl and Powley, 1990) and up to 80,000 in humans (Havton et al., 2021) The sensory vagal fibers stay together with vagal preganglionic efferents even in the finest nerve branches, making selective surgical manipulation impossible. Selective manipulations were first made on the basis of chemical susceptibility to the TRPV1 agonist capsaicin (Jancso et al., 1977) and the distinct course of vagal afferents and efferents as they exit the brain (rhizotomy, (Walls et al., 1995). With the ascent of transcriptomics, we now have genetics-guided tools to selectively manipulate (activate or silence) molecularly distinct populations of vagal afferent and efferent fiber populations (Bai et al., 2019; Borgmann et al., 2021; Williams et al., 2016).

4.1. Microscopic anatomy and putative functions of vagal sensory nerve terminals in the stomach

Intraganglionic laminar endings (IGLEs)

Silver-stained, putative afferent nerve endings with profusely arborizing flattened branches in contact with the inner surface of the connective tissue capsule enveloping myenteric ganglia were first reported in dog esophagus by Nonidez et al. (Nonidez, 1946). However, it took another 30 years until they were named intraganglionic laminar endings (Rodrigo et al., 1975) and their vagal sensory identity was indirectly confirmed based on their disappearance following extirpation of nodose ganglion neurons as well as infra- but not supra-nodose vagotomy in cat (Rodrigo et al., 1982).

Using direct anterograde tracing from the nodose ganglia with various tracers, the presence of IGLEs throughout the rat gastrointestinal tract was reported in the 1980ies and nineties, further confirming their sensory nature (Rat: (Berthoud and Powley, 1992; Clerc and Condamin, 1987; Kressel and Radespiel-Troger, 1999; Neuhuber, 1987; Wang and Powley, 2007); Mouse: (Fox et al., 2000). (Fig. 5A, B). The total number of IGLEs in the rat stomach was conservatively estimated to ~4,000, with about 25% of all myenteric ganglia innervated (Berthoud et al., 1997). Finding sensory nerve terminals abundantly in the myenteric plexus was somewhat unexpected; what could they possibly sense? Indirect proof of their mechanosensitive character was provided in the guinea-pig stomach. It was found that the locations of focal stimulation of the stomach muscle wall with van Frey hairs eliciting electrical activity in vagal afferents from the gastric cardia were in close register with anterogradely labeled IGLEs subsequently identified from the same whole-mounted tissue (Zagorodnyuk et al., 2001). These units had low stretch thresholds of <1 mm to circumferential stretch, showed slow adaptation, and correlated closely with accommodation of intramural tension and were thus classified as tension sensors (Zagorodnyuk et al., 2001). Soon thereafter, immunohistochemical studies demonstrated that IGLEs throughout the gastrointestinal tract express the purine receptor subunit 2 (P2X2) (Castelucci et al., 2003; Wang and Neuhuber, 2003).

Fig. 5. Basic morphology and location of vagal sensory endings in the stomach.

A: DiI-labeled (white) intraganglionic laminar endings (IGLEs) in the myenteric plexus of the rat corpus. Confocal images (z-stacks encompassing 10-20 μm) of stomach sections from rats with DiI injections into the nodose ganglia to anterogradely label vagal afferents and ip injections of Fluorogold for a counterstain of enteric neurons.

B: Larger magnification of gastric IGLE (brown) after anterograde tracing with Phal in nodose ganglion revealing their profusely arborizing structure ending in flattened lamina with weakly labeled enteric neurons (blue) in the background (Adapted from (Powley et al., 2019) with permission).

C: DiI-labeled (white) intramuscular array (IMA) in circular muscle layer of rat corpus. Note the close association of vagal afferent fibers with interstitial cell of Cajal in inset.

D: DiI-labeled (white) IMA entering the circular pyloric sphincter muscle (sm). Note entry of parent fiber (arrow) between Brunner’s glands which appear in black after digital subtraction of Fluorogold image.

More recently, transcriptomic analyses of vagal afferent neurons (VANs) and subsequent studies with genetics-guided anterograde tracing in different mouse Cre-lines confirmed the presence of IGLEs in the mouse GI-tract (Borgmann et al., 2021; Lowenstein et al., 2023; Scott et al., 2025; Williams et al., 2016; Zhao et al., 2022) (Fig. 6A). A majority of gastric IGLEs express the genetic marker Glpr1 (glucagon-like peptide receptor-1). Fewer gastric IGLEs are formed by a distinct population of VANs expressing the marker gene Oxtr (Oxytocin receptor) that predominantly form IGLEs in the proximal intestine (Bai et al., 2019), and many of these Glp1r+ and Oxtr+ VANs also express receptors for cholestystokinin (Cckar), serotonin (Htr3a/b), NPY/PYY (Npy2r), and capsaicin (Trpv1) (Bai et al., 2019). The markers Prox2 and Runx3 preferentially label IGLEs in the non-glandular and glandular stomach, respectively, but not in the small intestine (Lowenstein et al., 2023). Finally, about half of VANs forming gastric IGLEs also express Piezo2 (Lowenstein et al., 2023; Zhao et al., 2022).

Fig. 6. Nodose ganglion transcriptomics-based classification of vagal afferent endings and functions.

A: Quantification of mucosal ending- and IGLE-distributions labeled by vGlut2Cre, Nav1.8Cre, and the seven nodose-subtype Cre lines (Adapted from (Bai et al., 2019) with permission).

B: Genetic identity of vagal afferent neurons responsive to various body stimuli. Summary of vagal afferent neuron populations responsive to diverse stimuli, with key marker genes and response patterns determined using RNAscope across multiple visceral organs. Adapted from (Zhao et al., 2022) with permission).

A minority of IGLE-forming VANs express Prox2 (Prospero homeobox protein 2) and Runx3 (runt-related transcription factor 3) and have distinct dynamic mechanosensory properties. In the non-glandular stomach, a subtype of Prox2-expressing VANs (Prox2+, Glp1r+, Rbp4+, Gata3−, Piezo2+) predominantly exhibited slowly adapting properties, with distension-induced firing activity returning to baseline only after removal of the distension stimulus (Type I). In the glandular stomach, a subtype of Runx3-expressing VANs (Runx3+, Slc183+, Pappa2+, P2ry1+, Piezo2−) predominantly exhibited rapidly adapting properties, with firing activity returning to baseline while the stomach was still distended (Type II)(Lowenstein et al., 2023). Furthermore, gastric IGLEs formed by Piezo2+ and Agtr1a⁺ VANs have different preferences between mechanical and chemical stimuli (Zhao et al., 2022).

As already conceived in 1963, some vagal sensory terminals may also detect the neurochemical environment of, or electrical events in myenteric ganglia (Kolossow and Milochin, 1963) and lying on top of myenteric ganglia, IGLEs are in a perfect location to do that. Alternatively, they might modulate the underlying neurons by releasing transmitters or peptides. The presence of small clear and large dense core vesicles as well as synaptic specializations in esophageal and gastric IGLEs (Neuhuber, 1987; Powley et al., 2008) is consistent with these possibilities. An axon reflex like mechanism was suggested by Wei et al. based on observations that some gastric VANs with collaterals innervating the esophagus have local effects in the stomach (Wei et al., 1995).

The finding that a large majority of gastric IGLE-forming VANs express receptors such as Glp1r and Cckar was surprising. It was generally assumed that the natural ligand for these receptors, the intestinal hormones GLP-1 and CCK exert their powerful effects on vagally-mediated satiation by directly acting on vagal afferent terminals in the intestinal mucosa, where their concentration is highest. This view is starting to change. It is conceivable that hormones such as GLP-1, PYY, and CCK secreted in the intestinal mucosa are reaching IGLEs after a long journey via the systemic circulation to enhance the satiety signal from gastric mechanoreceptors. It had already been known for some time that gastric distension-induced single-unit vagal afferent firing in anesthetized rats was enhanced by CCK administration near the stomach (Schwartz et al., 1991). Interestingly, posthoc testing of the recorded units carried out with a blunt-tipped glass rod revealed circumscribed receptive fields mainly in the corpus, consistent with the location and geometry of IGLEs. It should thus be interesting to test for similar effects for GLP-1, PYY, and other satiety hormones by genetics-guided modern approaches including calcium imaging in nodose ganglia with subsequent anterograde tracing of activated units.

The genetics-informed studies also allow selective manipulation of specific populations of VANs such as optogenetic and chemogenetic stimulation, inhibition, and chronic silencing. Selective optogenetic activation of Glp1r+ vagal afferents (representing the majority of gastric IGLEs) caused a small increase (~8%) of gastric pressure (Williams et al., 2016). The fact that the centrally projecting axons of Glp1r+ VANs occupy a ventrolateral position in the NTS including termination sites in the DMV (Williams et al., 2016; Zhao et al., 2022) suggest that this change in gastric pressure is due to a vago-vagal reflex via the brainstem. An alternative explanation is that activated IGLEs directly act on enteric neurons via synapses as mentioned above (Neuhuber, 1987; Powley et al., 2008). Interestingly, selective optogenetic stimulation of either population (Glp1r+ or Oxtr+) of IGLE-forming VANs in freely moving mice robustly decreased fasting-induced food intake (Bai et al., 2019; Scott et al., 2025), suggesting that selective activation of either gastric or proximal intestinal tension sensors contributes to satiation.

Given the abundance of IGLEs in the stomach and throughout the GI-tract, systematic investigation of their molecular mechanisms and functions will be very important. Encouraging results, with reduced food intake and weight gain have been obtained with an ingestible bioelectronic vibrator suggested to stimulate gastric IGLEs (Srinivasan et al., 2023). However, novel in vitro and in vivo techniques allowing for selective recording of IGLE activity under controlled conditions will have to be developed to more selectively manipulate these powerful sensors.

Intramuscular arrays (IMAs)

Based on classical anterograde tracing, another type of characteristic vagal afferent endings, the so-called intramuscular arrays were described in both longitudinal and circular smooth muscle layers of the stomach wall and to a lesser extent in the intestines (Rat: (Berthoud and Powley, 1992; Wang and Powley, 2000), Mouse: (Fox et al., 2000) (Fig. 5C). Originating from fibers coursing through the myenteric ganglia and connectives, they enter either muscle layer and then run parallel to the respective muscle fibers, often for several millimeters in close association with the network of interstitial cells of Cajal, upon which they form multiple spiny appositions or varicosities (Powley et al., 2016; Powley and Phillips, 2011; Wang and Powley, 2000) (Berthoud and Powley, 1992). Ultrastructural evidence for small agranular vesicles, occasionally large granular vesicles and prejunctional thickenings in apposition to ICC-IMA processes suggests communication between ICC and IMAs via synapse-like contacts (Powley et al., 2016; Powley et al., 2008). In the forestomach, a singular concentration of orthogonally crossed IMAs is organized into a lattice (Wang and Powley, 2000) and IMAs display variations in morphology, with one specialization consisting of short terminal processes associated with sphincters and a more widespread form consisting of long, rectilinear processes in the forestomach, along the greater curvature (Wang and Powley, 2000) (Fig. 7A). The lower esophageal sphincter and the gastric sling muscles are innervated by specialized elongated IMAs and its distal antral attachment sites by web-like endings (Powley et al., 2013; Powley et al., 2012; Powley et al., 2016) and the pyloric sphincter is innervated by profusely and narrowly arborizing ring-like IMAs (Kressel et al., 1994; Powley et al., 2014; Powley et al., 2016) (Fig. 7B). Finally, approximately a third of gastric IMAs in the rat are polymorphic and polytopic, with distinct arborizations in the smooth muscle walls and myenteric ganglia (Powley et al., 2016).

Fig. 7. Distribution of IMAs in the rat stomach and sphincters.

A: Distribution of longitudinal muscle (blue) and circular muscle (yellow) IMAs in rat stomach. Scale bar = 8 mm. (Adapted from (Powley et al., 2016) with permission).

B: Locations of sling (blue) and clasp IMAs (red) constituting the specialized stretch sensors near the lower esophageal sphincter, as well as antral IMAs (purple and green) and web-endings (gold triangles) and pyloric IMAs yellow). From Powley et al. (Adapted form (Powley et al., 2013) with permission).

Compared to IGLEs, there appears to be no major/single genetic marker for IMAs. VANs forming IMAs responding only transiently vs. sustained to gastric stretch can be distinguished by their absence or presence of TMC3 (transmembrane channel 3) expression, respectively, but both of these populations also respond to other gastrointestinal stimuli (Zhao et al., 2022) (Fig. 6B). VANs with transient responses to gastric stretch also respond to duodenal stretch and duodenal nutrient infusion, while VANs with sustained responses also respond to duodenal and lung stretch. This suggests that distinct response patterns originate from the same VSNs and response heterogeneity is generated by different subpopulations of VANs. For example, Piezo2+ VANs respond to stretch in the lung, esophagus, stomach and duodenum with similar sustained kinetics (Zhao et al., 2022). Zhao et al. further suggested that “VAN response patterns are not well correlated with their ending structures, and that instead of developing specialized sensory mechanisms for individual body–brain pathways, VANs are organized into modular sensory units to code categorized stimulus modalities across visceral organs. (Zhao et al., 2022).

To summarize, gastric IMAs are a heterogeneous population with considerable morphological specializations and further progress in understanding their specific functions depends on the generation of specific Cre mouse lines that allow selective visualization and experimental manipulation of subpopulations of IMA-forming VANs.

Mucosal endings (MEs)

A third type of vagal afferent endings, mucosal endings were originally identified in the duodenum of rats following DiI anterograde tracing from the nodose ganglia (Berthoud et al., 1995). In the duodenum, mucosal endings densely innervate the crypts of Lieberkuehn and climb all the way to the tip of duodenal villi. In the gastric antral mucosa, which has a different architecture, profusely arborizing terminals densely innervate the base of gastric glands and climb in the lamina propria to reach just underneath the epithelium (Powley et al., 2011) (Fig. 8). Page et al. (Page et al., 2002) used a mouse vagus-gastroesophageal preparation which allowed studying the effects of local stimuli to the gastric mucosal surface on vagal afferent electrical activity. Stroking the mucosa with von Frey hairs generating local pressures of as little as 10 mg significantly increased activity of vagal afferents with a short latency, and a portion of these mechanosensitive vagal afferents also responded to local application of chemical stimuli such as 5-HT, ATP, hydrochloric acid, and bile acids. Although these findings might suggest involvement of vagal mucosal sensory endings, it is more likely that IGLEs were responsible. It was noted that stimulation with van Frey hairs, even with as little as 10 mg, caused observable distortion of the underlying layers, strongly suggesting that IGLEs located in the myenteric plexus were activated (Page et al., 2002). Potential functions of morphologically identified vagal afferent endings are comprehensively discussed in a recent review (Wang et al., 2020), acknowledging that clear evidence specifically on mucosal endings in the stomach are missing. Available behavioral data using pyloric cuffs suggest that gastric satiation is strictly volumetric and intestinal satiation is nutritive, suggesting that there are no direct nutrient sensors in the gastric mucosa (Powley and Phillips, 2004). However, both ghrelin and leptin are produced in the gastric mucosa (Bado et al., 1998) and some vagal afferents express corresponding receptors (Burdyga et al., 2002; Buyse et al., 2001) and could thus indirectly modulate satiety and appetite (Burdyga et al., 2006; Date et al., 2002; Peters et al., 2006).

Fig. 8. Mucosal endings of vagal afferents in rat stomach.

Vagal antral gland afferents. Note the collaterals on the right forming lamellar processes along the basal surfaces of the epithelial walls of the antral glands. On reaching the epithelium that constitutes the luminal surface, the collaterals often form aggregates of terminal varicosities and swellings (arrows). (Adapted from (Powley et al., 2011) with permission).

Transcriptomic analysis of mouse nodose ganglia revealed at least two molecularly distinct (anatomically non-overlapping) populations forming mucosal endings relatively specifically in the stomach (Bai et al., 2019) (Fig. 6A). A population of VANs expressing the marker gene Sst (somatostatin) are highly enriched in the pyloric antrum and nowhere else in the GI-tract. Another population expressing the marker gene Calca (calcitonin-related peptide alpha) forms mucosal endings in the corpus and IMAs in the antrum and colon (Bai et al., 2019). A third population of mucosal endings in the glandular stomach was distinguished by the genetic marker Gpr65, but the large majority of this population formed mucosal endings in the small intestine (Bai et al., 2019; Williams et al., 2016).

Based on experiments on in vitro-perfused guinea pig gastric wall with vagal nerves attached, it was found that they responded to stroking with von Frey hairs of 10-200 mg and to local capsaicin, but not to stretch (Berthoud et al., AJP R1371, 2001). Similarly, rapidly adapting single vagal afferent units presumably innervating the mucosa were previously isolated in cat (Clarke and Davison, 1978).

It is not known whether gastric enterochromaffin cells can activate vagal nerve endings as it has been shown to occur in the small intestine (Bellono et al., 2017; Spencer et al., 2024), where Htr3a and Htr3b are highly expressed in VANs innervating the duodenal mucosa (Zhao et al., 2022).

5. CNS integration of gastric vagal functions.

5.1. Dissemination of sensory vagal information from the stomach in the NTS and brainstem

Knowledge of how vagal sensory information specifically from the stomach is disseminated in the brain comes mainly from retrograde transganglionic tracing studies (Altschuler et al., 1989; Bassi et al., 2022; Marongiu et al., 2025; Norgren and Smith, 1988; Rinaman and Schwartz, 2004; Shapiro and Miselis, 1985; Shin and Loewy, 2009), and studies based on assessing receptor stimulation-induced neural activation with c-Fos (Willing and Berthoud, 1997), electrophysiological recording (Yuan and Barber, 1990), or calcium imaging (Ly et al., 2023; Ran et al., 2022).

Based on early retrograde tracing studies, a rough viscerotopic organization has been reported for centrally projecting axons of vagal afferents, with taste afferents mainly terminating in rostral, gastric afferents in medial, and intestinal afferents in more caudal aspects of the NTS (Altschuler et al., 1989; Norgren and Smith, 1988) (Fig. 9 A, B). These early studies also emphasized the presence of putative synaptic inputs from gastric vagal afferents to the DMV and AP (Rinaman et al., 1989; Shapiro and Miselis, 1985) that was confirmed by electrophysiological recording studies (Yuan and Barber, 1993). A more recent retrograde tracing study in rats which combined organ-specific vagal afferent, with labeling of putative synaptic inputs to NTS neurons found that there was little selectivity between vagal afferents from different viscera, such as the fundic stomach, proximal duodenum, heart and trachea, and that afferents from these different sites made close appositions with the same NTS neurons. Because of this widely distributed and overlapping input of viscerosensory information in the NTS, the study concluded that coding must be determined by the intrinsic properties and projections of 2^nd order NTS neurons (Bassi et al., 2022). Specifically, vagal afferents and putative synaptic inputs from the fundic stomach occupied a large area of the middle aspects of the NTS, with relatively minor differences in the distribution pattern compared with the much weaker overall inputs from the proximal duodenum (Bassi et al., 2022). In stark contrast, selective vagal sensory neuron labeling with a Cre-dependent AAV injected into the nodose ganglia of either Glp1r-ires-Cre or Gpr65-ires-Cre mice revealed strikingly different NTS termination sites. Glp1r-expressing vagal afferents representing mainly gastric IGLEs terminated selectively in the ventral aspects of the NTS, including the DMV, while Gpr65-expresssing vagal afferents representing duodenal mucosal receptors terminated in more dorsal aspects of the NTS, the dorsomedial and commissural subnuclei (Williams et al., 2016; Zhao et al., 2022). Multiple infusions into the mouse stomach wall of rgAAV2/rh10.hM3Dq-HA, a virus mixture expressing a DREADD receptor and preferentially taken up by sensory nerves labelled almost exclusively Glp1r+ neurons in the nodose ganglia and administration of CNO potently suppressed food intake and induced cFos in the NTS (Marongiu et al., 2025). Unfortunately, the study did not determine which NTS neuron type/s were activated.

Fig. 9. Pathways of eating-related vagal sensory input to the brain and dissemination of information in the brain.

A: Schematic diagram showing vagus and other cranial nerve branches carrying sensory information related to ingestive behavior to the brain and major communication pathways in the brain.

B: Termination sites of retrogradely traced vagal afferents in rat (Adapted from (Altschuler et al., 1989) with permission).

C: Peripheral stimulation-induced activity in optically recorded NTS neurons of mice (Adapted from (Ran et al., 2022) with permission).

Early studies measuring activity of 2^nd order NTS neurons with c-Fos immunohistochemistry were limited in that only one stimulus could be applied at a time. Selective activation of gastric mechanoreceptors by gastric balloon distension in rats significantly increased the number of c-Fos expressing neurons in most NTS subnuclei in a dose-dependent manner, with by far the highest activation in the medial NTS (Willing and Berthoud, 1997; Zheng et al., 1999). Compared to gastric balloon distension, duodenal infusion of glucose or intralipid elicited a similar activation pattern in the medial NTS, but in addition also strongly activated the dorsomedial and commissural NTS and area postrema, while gastric balloon distension strongly activated the DMV (Berthoud et al., 2001a).

Using an in vivo two-photon calcium imaging approach in mice with the ability to test several consecutive and diverse sensory stimuli, Ran et al. reported “a striking topographical map” in the brainstem for visceral inputs from the gastrointestinal tract and the larynx, with “neuron location in the brainstem reflecting the site of sensation within the body” (Ran et al., 2022) (Fig. 9 C). Not only were oral and laryngeal sensory sites well-separated from gastrointestinal sites, but within gastrointestinal sites, the stomach (antral balloon distension) was also well-separated from the more distal GI-tract (duodenal balloon distension). Importantly, the number of activated 2^nd order neurons in the NTS by gastric distension was significantly lower in mice with chemogenetic stimulation of inhibitory somatostatin/ GABA neurons in the NTS, an effect that could be acutely reversed by administration of the GABA-A blocker bicuculline. Such inhibitory gating, previously observed with in vitro brain stem slices (Thek et al., 2019) manifested itself when simultaneous mechanical stimulation of the duodenum and stomach resulted in fewer activated NTS neurons compared to each stimulus separate (Ran et al., 2022). Combining c-Fos and glutamate receptor immunohistochemistry, we also found a widespread, general involvement of both NMDA and AMPA receptors in primary afferent signal transmission at the level of the NTS, with no differential recruitment of the examined receptor subtypes by the different gastrointestinal sensory stimuli (Berthoud et al., 2001a).

The NTS harbors a great variety of neurochemically distinct neurons. Historically, catecholaminergic (TH+) NTS neurons, a subset of which also expresses prolactin releasing hormone (Prlh) have received much attention as important mediators of gastrointestinal stimuli induced satiety/anorexia (Bechtold and Luckman, 2006; Rinaman, 2003; Schwartz et al., 1991; Willing and Berthoud, 1997). With the ascent of GLP1R agonist administration for treatment of obesity, NTS neurons expressing glucagon (Gcg) and glucagon-like receptor 1 (Glp1r) became the focus of attention (Holt et al., 2019); for review see (Brierley et al., 2021)). Up to one third of Gcg-expressing neurons in the mid-to-caudal NTS were activated with selective gastric balloon distension in anesthetized rats (Vrang et al., 2003). Recent transcriptomic analyses revealed the presence of several clusters of glutamgateric NTS neurons distinguished by the expression of specific combinations of peptides and receptors (Dowsett et al., 2021; Ludwig et al., 2021), and there are currently major efforts to associate specific roles ingestive behavior and other functions with these subpopulations of NTS neurons (Cheng et al., 2022; Yacawych et al., 2025). Gastric distension-induced satiation in the physiological range is likely mediated by two clusters of glutamatergic NTS neurons, one expressing Th, Calcr (calcitonin receptor), Prlh (prolactin releasing hormone), and Glp1r (GLU11) and another one expressing Gcg, Lepr, and possibly Cck (GLU13 in Cheng et al.). Aversion and anorexia induced by supra-physiological gastric distension and/or toxic luminal stimuli are likely mediated by two distinct clusters of glutamatergic neurons, one expressing Rxfp1 (relaxin family peptide receptor) and Cck (GLU1) and one expressing Gfral (GDNF family receptor alpha-like) and Cck (GLU5 in Cheng et al.).

Genetics-guided studies selectively on Gcg and Prlh neurons, the two main caudal NTS neuron populations responsible for non-aversive satiation with a likely gastric component, revealed interesting differences (Ly et al., 2023). While Gcg neurons responded strongly to mechano- and chemo-sensory feedback from the gut and promoted satiety lasting for tens of minutes, Prlh neurons respond to both slower feedback from the gut and rapid input from the taste system, and their distinct role in satiation seems to be restraining the pace of ingestion (Ly et al., 2023). However, because intragastric catheters were used to deliver nutritive stimuli to the gut, the contribution of gastric vs. intestinal vagal sensors is not clear.

In summary, despite some conflicting findings that are likely to be due to differences in methodology and species, it seems to emerge that 1) most NTS neurons (of a range of molecular identities) receive converging input from many vagal afferents with distinct morphologies, peripheral locations, molecular makeups, and stimulus modalities, and 2) there is a rough spatial segregation between NTS neurons driven by intestinal chemosensory vs. gastric mechanosensory input, with the latter predominantly activating more ventro-lateral aspects of the NTS and the adjacent DMV. Great progress has also been made in deciphering the functional roles of genetically distinct populations of second order NTS neurons, particularly regarding their involvement in the process of satiation, but the specific role of gastric signals remains to be determined.

5.2. Dissemination of gastric vagal sensory information to higher brain areas: Is there gastric-specific interoception?

To reveal higher order sensory neurons in the brain, transsynaptic anterograde tracing with the H129 strain of herpes simplex virus is the method of choice (Barnett et al., 1995). Injection of H129 into the stomach wall of rats revealed a large neural network including the dorsal vagal complex (NTS, DMV AP) and many other medullary nuclei, nuclei in the hypothalamus (PVH, DMH, LH, ARC), the paraventricular thalamic nucleus and areas in the forebrain including amygdala, BNST, septum, zona incerta, hippocampus, perirhinal, cingulate, and agranular insular cortex (Rinaman and Schwartz, 2004), all brain areas that are generally also labeled with classical anterograde tracing from the more caudal aspects of the NTS (Ter Horst et al., 1989) (Fig. 9 A). H129 retrograde tracing also resulted in labeling of pontine catecholamine cell groups including the locus coeruleus (Rinaman and Schwartz, 2004), thought to play an important role in mediating the beneficial effects of vagus nerve stimulation in patients with irretractable epilepsy (Farrand et al., 2023).

Similar brain areas were activated by gastric balloon distension as measured by BOLD fMRI in rats (Min et al., 2011). In humans, neuroimaging has also been used to reveal brain areas receiving gastric sensory input. Distension was associated with activation in sensorimotor cortices and right insula; larger distension also activated the left posterior amygdala, left posterior insula and left precuneus. Responses in the left amygdala and insula were negatively associated with changes in self-reports of fullness and positively with changes in plasma ghrelin concentration, whereas those in the right amygdala and insula were negatively associated with the subject’s body mass index (Wang et al., 2008). However, because the stomach is also innervated by dorsal root afferents (Ma et al., 2023), these studies cannot rule out that some of the effects of gastric balloon distension on brain activity changes are mediated by dorsal root/spinal sensory pathways (Bielefeldt et al., 2005; Ozaki and Gebhart, 2001; Saper, 2002; Wolfson et al., 2023).

This brings up the question of whether there is stomach-specific interoception. Is there a stomach-specific neural network that can be distinguished from other organs, and can this sensation contribute to generating a conscious mental state or feeling? Early studies based on comparing taste, gastrointestinal distension, arterial chemoreception, and respiratory-related sensation suggested that central pathways are viscerotopically organized at all levels of the neuraxis, from brainstem to insular cortex (Cechetto, 1987; Cechetto and Saper, 1987). While it may be true for these widely disparate sensations, it is not clear whether sensations from closely associated organs such as the gut and pancreas, or different sensations from the same organ, such as the stomach and small intestine are anatomically and operationally separated.

Interestingly, the slow-wave gastric rhythm is synchronized to various degrees to brain activity at rest, particularly in primary sensory and motor regions, while in polymodal areas with higher level cognition gastric rhythm is only poorly synchronized (Rebollo and Tallon-Baudry, 2022), suggesting at least subconscious gastric awareness. Surprisingly, increased frontoparietal brain coupling to the gastric rhythm is associated with poorer mental health, spanning anxiety, depression, stress and well-being (Banellis, 2025), suggesting that being in “touch with one’s gut” is not always beneficial (Dalmaijer, 2025). Not only mechanosensory but also humoral information from the stomach can influence cognitive functions. There is evidence that the gastric hormone ghrelin modulates hippocampus-dependent contextual episodic memory formation but not anxiety-like behavior via vagal afferents (Davis et al., 2020). Molecular dynamics, synaptic plasticity, dendritic spine density, and long-term recognition memories were also shown to depend on subdiaphragmatic vagal integrity, but stomach-specific contributions to these effects were not studied (Onimus et al., 2024).

In a large cohort of patients with inflammatory bowel syndrome and functional dyspepsia it was found that a change in sensitivity to gastric balloon distension over time was an independent predictor for the gradual increase in GI symptom severity (Simren et al., 2018) and see (Elsenbruch, 2011) for a recent review). Looking for a less invasive, simpler test of gastric sensitivity, literature using the two-step Water Loading Test (WLT-II) which claims to assess “gastric interoception” and significantly affect ingestive, affective and other behaviors (Ahlich et al., 2023; Cantoni et al., 2025; Salaris et al., 2025; van Dyck et al., 2016). Although gastric sensation is clearly involved in this test, intestinal distension may play a more significant role, as water empties rapidly from the stomach to the small intestine, and mechanosensory input from the small intestine is particularly powerful in affecting ingestive behaviors (Bai et al., 2019). Likewise, in a study assessing the effect of experimental colitis on anxiety-like behaviors mediated by the basolateral amygdala, it was concluded that sensory information from the stomach was critical. However, none of the vagotomy procedures were selective for gastric vagal afferents (Chen et al., 2023). There is also mounting evidence for a role of gastric vagal afferents in pain perception. In mice, perception of noxious gastric distension depended on a glutamatergic pathway via the NTS, LPB, paraventricular thalamus to the prelimbic cortex (Zhang et al., 2024).

Thus, while evidence for general gastrointestinal interoception and its association with psychological health is rapidly growing, there is only questionable evidence for stomach-specific interoception and its neural pathways.

5.3. Brain sites controlling vagal outflow to the stomach: Vago-vagal reflexes

The DMV is a bilateral elongated spindle-shaped nucleus on either side of the 4^th ventricle and central canal, consisting almost exclusively of cholinergic (ChaT+) neurons, which are viscerotopically organized into longitudinal columns (Laughton and Powley, 1987). Projections to the stomach (through the two gastric branches occupy the two medial columns, while projections to the distal duodenum all the way down to the distal colon occupy the lateral two columns. Neurons projecting through the common hepatic/gastroduodenal branch to the antrum, pylorus, and proximal duodenum are found in both columns of the left DMV (Powley et al., 1987).

Directly overlying the DMV is the nucleus tractus solitarii (NTS), where almost all the central axons of vagal afferents terminate (Norgren and Smith, 1988; Powley et al., 2019). The rough rostro-caudal viscerotopic representation in the NTS, with taste input most rostral and intestinal input most caudal (Altschuler et al., 1989; Norgren and Smith, 1988; Ran et al., 2022) and the longitudinal viscerotopic organization of the DMV (Laughton and Powley, 1987) form an orthogonal sensory-motor lattice to efficiently organize vago-vagal reflexes (Powley, 2021). Vagal preganglionic neurons receive massive input from the overlaying NTS for the effectuation of vago-vagal reflexes (Fig. 10). Besides the monosynaptic inputs of primary vagal afferents (Neuhuber and Sandoz, 1986; Norgren and Smith, 1988; Rinaman et al., 1989), 2^nd order sensory neurons in the NTS represent a major source of input to DMV motor neurons for the elaboration of brainstem level gastric vago-vagal reflexes such as the gastric accommodation reflex and the gastric distension-induced gastric acid and pepsinogen secretion reflex. The importance of these reflexes has been recognized already in the early nineteen hundreds and literature is voluminous (For informative reviews see (Browning and Travagli, 2014; Browning et al., 2026; Gillis et al., 2022; Li and Owyang, 2003; Travagli and Anselmi, 2016; Travagli et al., 2006).

Fig. 10. Schematic diagram showing central neural pathways involved in vagal parasympathetic control of the stomach.

Brain nuclei identified based on transsynaptic retrograde tracing with pseudorabies virus (PRV) injections into the stomach of rats and mice. Note that interconnecting pathways in green are inferred based on timed survival studies and may not distinguish direct from indirect pathways. The size of individual nuclei represents approximate number of neurons labeled. Vagal and spinal sensory pathways are shown in blue and purple, respectively. Sensory brain nuclei and pathways potentially linked to motor outflow pathways are shown in blue. Abbreviations: ARC, arcuate nucleus; AP, area postrema; A5, noradrenergic cell group in the pons; BNS, bed nucleus of the stria terminalis; CeA, central nucleus of the amygdala; DH, dorsal horn of spinal cord; DMN, dorsomedial nucleus; DMX dorsal motor nucleus of the vagus; HIP, hippocampal formation; ILC, infralimbic cortex; INS, insular cortex; LHA, lateral hypothalamic area; MeHA, medial habenula; MRN, various medullary reticular nuclei; NTS, nucleus tractus solitarius; OLF, piriform cortex; PAG, periaqueductal grey; PBN, parabrachial complex; PFC, prefrontal cortex; PVH, paraventricular nucleus of the hypothalamus; SCN, suprachiasmatic nucleus; VMN, ventromedial hypothalamus.

5.4. Location and function of higher order gastric pre-autonomic neurons

Information about upstream brain areas controlling vagal efferent outflow to the stomach comes mainly from conventional and transsynaptic retrograde tracing with PRV in the rat. Retrograde tracing of vagal efferents and afferents with injections of cholera toxin-HRP conjugate into the gastric external muscle in rat revealed the morphology of DMV motor neurons with dendrites extending to the overlaying subnucleus gelatinosus of the NTS, where they form synapses with labeled vagal afferent terminals (Rinaman et al., 1989). Transsynaptic labeling of pre-autonomic forebrain neurons achieved by inoculating the ventral stomach wall with pseudorabies virus (PRV) on postnatal day 1, 4, or 8 in rats resulted in age- and length of survival-dependent successively increased labeling in a brainstem → hypothalamus → cortex axis that includes the NTS and ventral medulla, lateral parabrachial nucleus, hypothalamus, amygdala, bed nucleus of the stria terminalis, and visceral cortical areas such as the insular cortex (Rinaman et al., 2000). Similar brain areas containing pre-autonomic neurons were identified with non-selective monosynaptic rabies tracing from Chat+ DMV neurons (Zhu et al., 2024).

PRV retrograde tracing from the anterior wall of the rat stomach also confirmed the presence of stomach-projecting neurons in layer V of the rostral insular cortex, with fewer neurons in the infralimbic and prelimbic prefrontal cortex (Levinthal and Strick, 2020). Stimulation of both of these cortical areas has been shown to elicit gastric motility changes (Hurley-Gius and Neafsey, 1986; Yasui et al., 1991). Since the gastric motor area in the insular cortex is adjacent or even overlapping with the gastric sensory area it is seen as visceral counterpart to the somatic sensory-motor cortex and lays the foundation for conscious perception of visceral signals (Barrett, 2017; Levinthal and Strick, 2020; Saper, 2002). Interestingly, retrograde PRV tracing from the stomach through spinal pathways (in rats with vagotomy) revealed a very different labeling pattern, with infected neurons primarily located in the motor (M1 and M2) and sensory (S1) cortices (Levinthal and Strick, 2020). PRV tracing from the pyloric sphincter, besides heavy labeling in the DMV, resulted in labeling of neurons in the NTS, AP, and PVH (Richardson et al., 2023). However, the fact that the nucleus ambiguous was also heavily labeled suggests that there was significant tracer leakage, as the pyloric sphincter is not innervated by nucleus ambiguous neurons.

Importantly, having identified gastric-specific subclasses of DMV ChAT neurons (see above) it is now possible to use selective retrograde tracing approaches for the identification of pre-autonomic vagal neurons responsible for specific gastric functions.

There is increasing evidence that the vagus nerve acts as a bidirectional conduit for the spreading of Parkinson pathology originating with aggregation and neuron-to-neuron propagation of alpha-synuclein (Borghammer, 2023; Anselmi et al., 2017; Heng et al., 2022; Holmqvist et al., 2014; Montalban-Rodriguez et al., 2024; Ozkaya and Browning, 2026; Rietdijk et al., 2017). Specifically, the gut (periphery)-first type of pathology suggests that alpha-synuclein accumulation in the ENS spreads to dopamine neurons in the substantia nigra via the vagus nerve (Rietdijk et al., 2017) and is characterized by early autonomic and gastrointestinal symptoms, as well as REM sleep behavior disorder (as reviewed in: Borghammer, 2023; Ozkaya and Browning, 2026). In contrast, the brain-first type of pathology suggests that initial aggregation of alpha-synuclein in the olfactory bulb spreads to the substantia nigra and may eventually affect vagal outflow and gastrointestinal functions (Anselmi et al., 2017; Borghammer, 2023; Ozkaya and Browning, 2026).

5.5. The gastric vagus as neuromodulation target

Bilateral subdiaphragmatic/abdominal/truncal vagotomy (TV) used to be an important procedure for the prevention of recurrent gastric ulceration (Gillespie, 1969) and is still used in some patients today (Clapp et al., 2025). As weight loss is one of the major consequences of TV, it was also used as a treatment for morbid obesity (Kral, 1979). Weight loss after TV is due to an inability to empty the stomach, as the inhibitory vagal input to the pyloric sphincter muscle via nitrergic postganglionic enteric neurons is removed. This can easily be prevented by pyloroplasty or providing exclusively semi-liquid diets (Yox et al., 1991). With these precautions, rats and normal weight humans do not lose weight, appear healthy, and appear to live quite normal lives. This underpins the remarkable redundancy in physiological control systems and suggests that gastric functions, as well as functions of other abdominal organs supplied by the vagus nerve, do not critically depend on vagal control. It also suggests that putative interoceptive functions of vagal afferents from these organs (independent from vagal efferent modulation) are not seriously impeding life. Thus, central control of gastric functions should be understood in this context. The brain merely modulates mostly autonomous gastric functions, and it does that partly by setting the efficiency of vago-vagal reflexes at the level of the caudal brainstem.

Given its easy accessibility, the vagus nerve has been the target of numerous neuromodulation approaches as potential therapies to treat obesity and other diseases. For vagal stimulation approaches that are not gastric-specific, the reader is referred to recent reviews (Berthoud and Neuhuber, 2019; Fadel et al., 2023). Several electrical stimulation devices targeting gastric vagal branches or gastric musculature intended to reduce appetite, body weight, and improve glucose handling have recently been put on the market (see (Lebovitz, 2016) for a recent review). The Tantalus (DIAMOND ^™) gastric electrical stimulatory device consists of three pairs of electrodes implanted into the fundal, anterior and posterior antral smooth muscle of the stomach connected to a pulse generator with a charging coil placed into a pocket of abdominal subcutaneous fat. The fundus electrode is activated by food distension and activates the antral electrodes synchronized to slow wave activity (Cheng et al., 2021). Several large clinical trials demonstrated the ability of this device to meaningfully reduce body weight and HbA1c (Lebovitz, 2016; Lebovitz et al., 2015). Similarly, the “Abiliti” device consisting of a trans-gastric sensor near the greater curvature of the fundus/corpus region to detect food entry and triggering a stimulation electrode near the lesser curvature at the point where the nerve of Latarjet divides into the “crow’s foot” was also able to significantly reduce body weight in an obese population (Busetto et al., 2017; Horbach et al., 2015; Miras et al., 2015). The observation that second order sensory vagal neurons in the NTS of rats are selective to the orientation and location of gastric electrical stimulation (Powley et al., 1987) could help further optimization of electrode placements in humans, and advances in finding stimulation parameters for selectively activating vagal sensory fibers (Ahmed et al., 2020) may lead to further improvements in efficacy of such devices in the future.

6. Conclusions & Outlook

The stomach is more heavily innervated by the vagus nerve than any other organ. In rodents, there is anatomical and functional evidence that vagal motor fibers originating in the DMV innervate every ganglion of the gastric myenteric plexus and make synaptic contacts with a large majority of enteric neurons populating these ganglia. Importantly, vagal motor neurons do not directly innervate effector cells and tissues. Instead, they modulate activity of enteric neurons. Excitatory and inhibitory enteric muscle motor neurons are driven by separate populations of vagal motor neurons that express specific marker genes and can be selectively manipulated. Based on transcriptomic analyses of DMV neurons, additional subpopulations of vagal motor neurons with distinct morphologies, molecular identities, postganglionic/enteric neurons, and functions are currently being characterized. Thus, the emerging concept of the coding logic for vagal motor outflow is one of specialized labeled lines, where each line serves a particular gastric function. The “Command Neuron” hypothesis which was based on a much less dense vagal motor innervation of the stomach suggested that special enteric neurons as sole recipients of vagal input acted as highly collateralizing interneurons forming “internuncial” circuits or pattern generators for the initiation of functional programs. The emerging new concept suggests that at least some gastric functions are modulated by vagal efferents directly affecting enteric muscle motor and mucosa innervating neurons. It thus shifts the “command” from enteric to vagal motor neurons. However, the fact that essential gastric functioning is preserved after truncal vagotomy is consistent with the idea that the vagus merely modulates program libraries and reflexes intrinsically organized by the ENS.

One missing link for fully understanding the vagal-enteric interface is the exact axon collateralization pattern of distinct labeled lines of vagal motor neurons. The axon collateralization pattern could largely determine which program or pattern is being modulated. To this end, genetics-based visualization of specific populations of vagal motor neurons in combination with traditional ENS electrophysiology may be a promising approach for future studies.

Gastric vagal motor neurons in the DMV have been well characterized using traditional and geneticsbased approaches, but a full understanding of inputs from pre-autonomic and higher order neurons to function-specific subclasses of DMV neurons is still missing. This will be important for the development of pharmacological and neuromodulation approaches that selectively target function-specific subpopulations of vagal motor outflow to the stomach and other vagally targeted organs and tissues (Zhu et al., 2024).

The stomach is also heavily innervated by vagal sensory neurons located in the nodose ganglia linking mechanical and chemical signals from the stomach directly to the brainstem. There have been significant advances in understanding the coding logic of vagal sensory neurons based on transcriptomic analysis of primary sensory neurons in the nodose ganglia and second order sensory neurons in the brainstem. The most abundant type of vagal sensory neuron forms IGLEs and is identified by the marker gene Glp1r, with a minority expressing Oxtr. IGLEs are embedded between the longitudinal and circular external muscle layers and in intimate contact with myenteric plexus ganglia. IGLEs are thus in a perfect position to detect stretch and tension of the stomach wall, as well as chemical and electrical signals from enteric neurons.

Another type of vagal afferent neuron forms IMAs that run parallel to the longitudinal and circular smooth muscle layers and are in intimate contact with interstitial cells of Cajal. IMAs are thus in an ideal position to sense stretch of the external muscle layer in both directions. Specialized IMAs cluster at strategic locations such as the LES, fundus, antrum, and pylorus, but little is known about their segregation into unique, molecularly distinct populations and their central connectivity, possibly owing to their often polytopic and polymorphic character.

The third type of vagal afferent ending, mucosal endings have a complex architecture and innervate the gastric glands to reach up just underneath the epithelium. They remain virtually unexplored in terms of their adequate sensory stimuli, genetic coding, and central dissemination. Future studies using approaches that include genetics-based visualization and electrophysiological recording from identified single axons and simultaneous application of localized mechanical and chemical stimuli will be necessary to unravel their exact location and sensory functions.

Finally, the specific contribution of function-specific vagal gastric sensory pathways to interoception, feelings, self-awareness, and pain is unclear. Anatomical studies show dissemination of gastric vagal sensory information to a large number of brain areas, similar to the dissemination of vagal sensory information from other viscera. Whether there are gastric-specific and sensory modality-specific pathways to any of these many brain areas remains to be investigated. Recent functional studies suggest that awareness of gastric fill can modulate emotions and may have implications for general well-being, but the methodologies cannot exclude contributions of vagal sensory input from the distal gut.

Highlights.

• Among all innervated organs, the stomach is most heavily innervated by the vagus nerve.

• Vagal efferent, parasympathetic control uses molecularly distinct pathways mediating excitation and inhibition of stomach and pylorus muscle functions.

• Separate populations of vagal sensory neurons have distinct molecular fingerprints and unique morphologies and functions.

• Further research on the organization of the peripheral and central interfaces is needed to fully understand the overall coding logic of both efferents and afferents.