Deglutition and the Regulation of the Swallow Motor Pattern

Abstract

Despite centuries of investigation, questions and controversies remain regarding the fundamental genesis and motor pattern of swallow. Two significant topics include inspiratory muscle activity during swallow (Schluckatmung, i.e., “swallow-breath”) and anatomical boundaries of the swallow pattern generator. We discuss the long history of reports regarding the presence or absence of Schluckatmung and the possible advantages of and neural basis for such activity, leading to current theories and novel experimental directions.

Introduction

Deglutition (swallow) is a critical motor action necessary for calorie intake and survival. It is a primitive behavior found in organisms ranging from single cells to mammals (1–6). In modern text, it was first described by William Harvey (7) in his publication on the motion of the heart and blood. Leake’s 1970 translation (8 of Harvey’s 1653 work stated, “… in swallowing; lifting the tongue and pressing the mouth forces the food to the throat, the larynx and the epiglottis are closed by their own muscles, the gullet rises and opens its mouth like a sac, and receiving the bolus forces it down by its transverse and longitudinal muscles. All these diverse movements, carried out by different organs, are done so smoothly and regularly that they seem to be a single movement and action, which we call swallowing” (p. 48).

As vertebrates have become more complex, the action of swallow has evolved in mammals to include three distinct phases (oral, pharyngeal, and esophageal). This idea was first introduced by Magendie (9) in 1822, who noted, “Though apparently simple, deglutition is by far the most complicated of all the muscular actions, which assist in digestion” (p. 201).

The oral phase is divided into two stages (preparation of the bolus and propulsion of the bolus into the pharynx). It is voluntary and can be modulated by taste [carried by facial (VII), glossopharyngeal (IX), and vagus nerves (X)] (10, 11), food texture, and activation of thermal/irritant receptors [carried by the trigeminal nerve (V)] (12). The pharyngeal phase is a reflex but is modulated by ascending and descending pathways (13–16). The esophageal phase moves the bolus into the stomach via peristaltic contractions and is primarily controlled by the autonomic nervous system (17–19). This review will focus on the pharyngeal phase of swallow, herein termed “swallow.”

Accepted Canon for the Basic Motor Pattern of Swallow

Swallow is initiated when the tongue retracts to direct the bolus into the oropharynx. The velopharyngeal port (connection point of nose and throat) closes to build pressure to propel the bolus toward the esophagus (13–16). The pharynx and larynx elevate simultaneously. As seen in the electromyograms in FIGURE 1A, the submental muscles contract, moving the hyoid bone and larynx superior and anterior under the tongue; this is defined by the onset of the mylohyoid (part of the leading complex of muscles, which also includes thyrohyoid) (13–16). During laryngeal elevation, the upper esophageal sphincter (UES) relaxes to open the esophagus, and the vocal folds and aryepiglottic folds adduct (as seen with thyroarytenoid contraction) to prevent material from entering the lower airway. Finally, as the inferior pharyngeal constrictor closes (thyropharyngeus), the bolus passes into the esophagus and then the UES closes. In mammals, this action takes ∼500 ms regardless of species.

FIGURE 1.

Muscle activation during swallowing in a cat A: upper airway sequential muscle activation during swallow in the cat. The beginning of the pharyngeal phase of swallow is marked by the onset of the mylohyoid (leading complex) and diaphragm electromyographic (EMG) activity (red arrows) and coincident upper esophageal sphincter (UES) relaxation. The end of the pharyngeal phase is marked by the end of thyropharyngeus (inferior pharyngeal constrictor) activity and the resumed tone in the UES. The anatomical rendering illustrates how upper airway muscles are activated in a rostral to caudal progression during swallow. This sequence is imperative for proper bolus movement through the pharynx and into the esophagus. B: identification of phase of breathing using diaphragm and thyroarytenoid EMG activity: 1) inspiration (pink) is defined as the onset to the peak of diaphragm activity; 2) E1 (blue) (i.e., early expiration, post-I, or yield) is defined as the peak of diaphragm activity to the end of the thyroarytenoid burst; and 3) E2 is defined as the end of E1 to the beginning of inspiration. C: example of swallow (green box) during midexpiration (E2) and during E1 (halting ongoing inspiration). Both swallows demonstrate costal and bilateral crural diaphragm EMG activity consistent with Schluckatmung; however, the second swallow has larger motor bursts.

While swallow can occur or span across any phase of breathing (FIGURE 1B), in most mammals swallow has an expiratory phase preference (20–24). FIGURE 1C displays an electromyogram recording of one swallow during late expiration (termed E2) and another during the transition from inspiration to expiration (termed E1, post-I, or “yield”). The swallow motor pattern involves the bilateral activation of dozens of muscles that must be precisely coordinated. These are innervated primarily by cranial nerves (V, VII, X, IX, and XII: trigeminal, facial, vagus glossopharyngeal, and hypoglossal, respectively). Swallow muscles are inherently multifunctional and are involved in a wide array of oromotor, respiratory, and other behaviors. These behaviors must be coordinated precisely with each other and with swallow. Due to overlapping or shared motor neuron pools, there is likely a great deal of overlap in premotor regions common to multiple behaviors, rather than dedicated circuits for producing distinct behaviors (25–30).

Schluckatmung and the Continuing Controversy

As is often the case for complex motor behaviors, debates remain regarding the accepted understanding of features of the swallow motor pattern. One controversy that has continued for over 200 years is whether or not chest wall muscles and/or the diaphragm, which are active during inspiration, also operate multifunctionally by participating during swallow. While this component of the swallow pattern has been described and “named” by research groups since the 1800s, this hypothesis has not persisted into the modern literature.

The respiratory physiologist Rosenthal (31) was the first to report, and Arloing (32) provided additional substantial evidence, that diaphragm activity is an active part of swallow and creates a negative deflection that he called “Schluckatmung,” which means “swallow-breath” in German. Rosenthal (1861) first described it in his studies on diaphragm movement with superior laryngeal nerve stimulation (31); in 1865, both Bidder (33) and Blumberg (34) concluded that the movement of the diaphragm was concurrent with swallow, but they considered it passive.

In 1870, Waller and Prévost demonstrated that the diaphragm was an active component of the swallow motor pattern because its action remained after the trachea and esophagus were transected but was diminished after phrenic nerves (which supply the diaphragm) were cut (35). Physiologists Meltzer (36) and his mentor Kronecker (37–39), among the first to use esophageal manometry in humans, demonstrated that the Schluckatmung was accomplished by central connections between the swallow and breathing pattern generators, but hypothesized that the action represented expiratory movement of the thorax. Steiner (40) independently confirmed that contraction of the diaphragm was an active component of the swallow pattern and also that it occurred with or without swallow apnea (i.e., laryngeal/glottic closure) and during respiratory suppression by morphine (40). Subsequently, Marckwald outlined his own and others’ investigations on this topic (41). Marckwald’s studies in rabbits determined that 1) diaphragm action follows the mylohyoid by 20–30 ms (as recorded with levers attached to the mylohyoid, larynx, and diaphragm); 2) diaphragm activity begins before laryngeal/glottic closure; 3) thoracic inspiratory muscles also participate during Schluckatmung; 4) Schluckatmung is dependent on central respiratory circuits; and 5) during prolonged apnea, Schluckatmung remains (41). Marckwald also posed the most evident fundamental question of the controversy: why would diaphragm activation during swallow not cause aspiration, let alone be helpful?

Negative Pressure

The controversy resurfaced, as described by Bosma in his 1957 review. At that time, fluoroscopy was the primary method employed in swallow investigations, and the definition of the pharyngeal phase of swallow was still being considered (42). Barclay, a renowned radiologist and the namesake for Barclay-Baron disease, had collaborated with the physicist Anrep to also combine fluoroscopy with manometry, and their experiments were outlined in a 1933 book on the digestive tract (43). Previous publications (44–48) had reported negative pressure between 16 and 18 cmH₂O, which is described as necessary for pharyngeal clearance. Barclay scrutinized the balloon introduced into the esophagus for measurement, noting that the action of the esophagus would be positive if the balloon were overinflated (43). This was confirmed by Allison’s recording of positive pressure in excised esophagi (49). Barclay’s work was largely accepted and was presented by Negus (president of the Section of Laryngology) in 1942 (6). In Bosma’s review, he cited a study by Saunders and colleagues (50) to supposedly refute Barclay’s claims, but the explanation in that manuscript does not satisfactorily explain the negative pressure.

Negative pressure during swallow was later described by Sokol et al. (51) in 1966, who noted that there were sub-atmospheric pressures (i.e., sucking) in the pharynx and upper esophagus, but attributed it to upper airway movements or esophageal elongation. A particularly important study in 1970 tracked esophageal pressure before and after repair of a diaphragmatic hernia and demonstrated that negative pressure was only accomplished when the diaphragm could dissent properly during swallow (52). Beginning in the 1980s, the otolaryngologist McConnel (53–64) published a number of reports establishing that negative pressure is a significant component in the normal physiology of swallow. He described this action as the “hypopharyngeal suction pump,” and in a 1988 study demonstrated that there are two major forces on the bolus: a ∼17-mmHg positive pressure from above and ∼20-mmHg negative pressure from below, which he mechanistically ascribed to laryngeal elevation (55). McCulloch, an otolaryngologist, more recently began publishing an extensive body of work using high-resolution manometry (65–71). By 1993, Castell and Castell (72) had attributed the negative pressure during the esophageal phase of swallow solely to relaxation of the upper esophageal sphincter, and this explanation has prevailed in McCulloch’s reports.

While several publications from the midcentury era had investigated pressures during swallow, especially using human manometry and fluoroscopy, they did not often use the term Schluckatmung nor did they posit that diaphragmatic activity could be the source of pressure deflections. Ingelfinger’s 1958 review of esophageal motility discusses negative pressure during swallow with no mention of chest wall muscle activity or Schluckatmung (18). There are, however, a few exceptions. Vantrappen and Hellemans (73) reported in 1968 that negative esophageal pressure during swallow was always accompanied by negative pleural pressure. In 1988, Kennedy and Kent (74) demonstrated negative pressure during swallow briefly following laryngeal elevation, and they used the term Schluckatmung in reference to contraction of hypopharyngeal musculature. Howard et al. (75) in 1989 also demonstrated a negative 5- to 10-mmHg pressure in the cervical esophagus during swallow and described Schluckatmung as a short inspiration that precedes swallow. Finally, the report of Jacob et al. (76) in 1989 most closely approximates the original definition of the term Schluckatmung but notes that it has not been “adequately explained.”

Action of Inspiratory Muscles

Early work on the regulation of diaphragm activity during swallow used levers to track movement (41). In the 1950s, Odanaka was the first to demonstrate via electromyographic (EMG) recording that the sternal diaphragm was active during swallow induced by infusion of water in the mouth of cats (77, 78). Shortly afterward, Doty and Bosma (1) published their investigation of EMG activity during swallow in cat, dog, and monkey. They reported that 80% of swallow in dog and monkey occurred during inspiration, with inspiratory diaphragm activity continuing for 50–200 ms into the subsequent activity of the leading complex of swallow muscles. This work in dog was extended in 1964, showing negative pressure in the trachea at the beginning of swallow that transitioned to negative pressure in the esophagus (79). In 1975, Sumi presented at a developmental symposium and later published work (80) showing that, in kittens, the swallow diaphragmatic activity was always present and was actually larger compared to breathing (81).

This has also been observed in infants by the neonatologist Thach and others (82, 83), and in lambs by Harding and colleagues (84, 85), although the representative swallow bursts there were not larger compared to breathing. In infants, swallow must also be coordinated with other rhythmic oromotor behaviors. Suckling and swallow are established prenatally well before respiratory rhythms, and nonnutritive suckling can occur in parallel with respiration (86, 87). Oral pressure and jaw movements are coordinated during nutritive but not during nonnutritive suckling (86), and when swallows occur during a masticatory or bottle-feeding sequence in humans, they usually do so with the jaws closely approximated (86, 88, 89). Because of the extensive coordination required during suckling, there may be additional considerations necessary for understanding Schluckatmung regulation during development.

Recent evidence of Schluckatmung in adult humans also comes from the work of investigators in Sweden in 2009 and Turkey in 2012. In both studies, EMGs were placed in the costal diaphragm (lateral to the sternum) (90, 91). Collectively, they demonstrate swallow-related diaphragmatic activity during inspiration, expiration, and the transition from inspiration to expiration. Ceborg et al. (90) also reported diaphragm activity when there was no airflow at the mouth [i.e., swallow apnea (92, 93)]

Central Schluckatmung Regulation

A vast volume of investigations into the central regulation of breathing has classically identified two distinct populations of diaphragmatic premotor neurons in the brainstem (94–103) (FIGURE 2). The first is the ventral respiratory column (VRC; FIGURE 2A), which runs as a rostral/caudal column on each side of the ventrolateral brainstem; the second is a region that has historically been known as the dorsal respiratory group (DRG). The DRG has bulbospinal phrenic projection neurons in cat (99, 104, 105); it is in the caudal medulla and overlaps with the nucleus tractus solitarius (NTS) and nearby reticular formation (RF) (94–103). The current putative respiratory network includes numerous additional brainstem regions. Usually, control of breathing investigations are performed in preparations that are decerebrate, paralyzed, and artificially ventilated (fictive behavior can be recorded) (106–111) or anesthetized (30). In swallow studies, swallow is evoked by two general methods: 1) electrical stimulation of the superior laryngeal nerve (SLN) at a fixed frequency (106–111), or 2) water infusion into the oropharynx (30, 112–116). In most studies that record neurons in the VRC, the cerebellum is also removed. These conditions present effects that may potentially confound the observation and subsequent understanding of Schluckatmung regulation.

FIGURE 2.

Representation of the rat brain stem A: a dorsal view representation of the rat brainstem identifies regions involved in breathing and swallow. The orange shaded area indicates the dorsal respiratory group (DRG), and the green area indicates the ventral respiratory column (VRC) and pontine respiratory group (PRG). The gray shaded region bounded by the dashed line indicates the classically accepted region for the swallow pattern generator (dorsal and ventral swallow groups). Pink areas indicate medullary medial (central) reticular formation, and yellow area indicates medullary median (midline) reticular formation (raphé). The nucleus tractus solitarius (NTS) receives peripheral afferent information important for both breathing and swallow. Motor nuclei innervating muscles involved in both breathing and swallow include those for trigeminal (MoV), facial (VII nuc), vagal (NA), hypoglossal (XII nuc), and phrenic nerves (PMN; via spinal nerves). B: a schematic (cat) shows that feedback from the upper airway and cervical regions can centrally affect Schluckatmung. Purple dashed arrow represents efferent drive from the brainstem respiratory regions (VRC and DRG) to phrenic and intercostal/parasternal motoneurons located in the cervical and thoracic spinal cord. Afferent feedback from the larynx and trachea centrally inhibits diaphragm/parasternal activity during swallow (red arrows), while activation of pharyngeal and esophageal mechanoreceptors stimulates diaphragm/parasternal activity during swallow (green arrows). By disrupting these connections, cervical spinal cord injury reduces Schluckatmung activity. C: a magnified brainstem schematic view shows the central regions thought to be important for swallow. In green/red are the peripheral afferent and higher-order brain inputs that would stimulate/modify swallow. In grey are the classically accepted regions for the swallow pattern generator: the dorsal swallow group (DSG) in the NTS, and below that the ventral swallow group (VSG). The pink region is the approximate area of the medullary reticular formation. In blue are the motor nuclei for the nerves innervating the muscles (efferent outputs in purple) that produce the swallow pattern. This is adapted from Jean’s (13) classic trisynaptic model for the swallow reflex pathway (here, each synapse is labeled with a red number). In this model, 1) the afferent swallow stimulus enters the DSG (which is contained mostly within the NTS); 2) the interneuronal DSG triggers the swallow command and sends it to the VSG; and 3) the interneuronal VSG (contained within the caudal reticular formation) distributes the command to the cranial motor nuclei that innervate the swallow muscles. We propose that the VSG should be expanded to include a greater portion of the medullary reticular formation, which is an important premotor region for orofacial behaviors. We also propose that there may be more than three synapses involved in the swallow reflex pathway, particularly among interneurons in the reticular formation and their connections. The ventral respiratory column (VRC; in green) and dorsal respiratory group (DRG; in orange) have also not been included in the classic swallow circuit, but are involved because activation of the diaphragm during swallow requires inspiratory neuron activity. Finally, motor regions of the cervical and thoracic spinal cord are also likely involved because they innervate inspiratory muscles active during swallow (e.g., diaphragm, parasternal, intercostals). V, trigeminal nerve; KF, Kolliker-Fuse nucleus; LPBr, lateral parabrachial nucleus; MPBr, medial parabrachial nucleus; MoV, motor nucleus of the trigeminal nerve; A5, noradrenergic group 5; VII nuc, facial motor nucleus; pFRG/RTN, parafacial respiratory group/retrotrapezoid nucleus; sp5, spinal trigeminal tract; BötC, Bötzinger complex; preBötC, pre-Bötzinger complex; rVRG, rostral ventral respiratory group; cVRG, caudal respiratory group; NTS, nucleus of the solitary tract; AP, area postrema; XII nuc, hypoglossal motor nucleus; NA, nucleus ambiguus; NRA, nucleus retroambiguus; VRC, ventral respiratory column; PRG, pontine respiratory group; PMN, motor nucleus of the phrenic nerve.

Marckwald (41) was the first to investigate the participation of the VRC in Schluckatmung. He reported that, following brainstem lesions, which eliminated respiration in the rabbit, swallows were still produced but they lacked diaphragmatic activity. He also noted suppression of Schluckatmung activity in decerebrate preparations in which cerebellectomy was additionally performed. Hukuhara and Okada (116) and Sumi (117) recorded from putative premotor VRC neurons that demonstrated a burst of action potentials at the beginning of the swallow, followed by a period of inactivity. Both groups used decerebrate animals, and swallow was stimulated by water infusion. In contrast, Saito and colleagues (115) used SLN stimulation and found little to no activity of VRC neurons during fictive swallow. Nakazawa and colleagues (111) performed a similar study of fictive swallow using SLN stimulation and obtained the same result, with the exception of a VRC inspiratory neuron that was only active during Schluckatmung.

Gestreau and colleagues (110) performed intracellular recordings of DRG neurons premotor to phrenic motoneurons (diaphragm). While most were briefly activated during fictive swallow induced by SLN stimulation, they were much less active during swallow than during breathing. In contrast, our work in anesthetized cats demonstrated that many putative premotor neurons were more active during swallow than breathing (30). In a recent study by Hashimoto et al. (118) in in situ paralyzed rat, about one-third of swallow neurons recorded in the region of the DRG had respiratory-related activity, most of which was inspiratory. We have also found, in anesthetized preparations, that Schluckatmung is significantly suppressed when even just the cerebellum is removed (119) or when swallow is induced by SLN stimulation rather than oral water infusion (120). In the study where we compared SLN stimulation to water infusion, we speculated that this suppression of Schluckatmung during SLN-stimulated swallow may represent a protective mechanism to decrease negative pressure across the larynx and thus reduce the chance of aspirating the bolus. This is because evoking swallow with electrical SLN stimulation represents an aberrant condition, as laryngeal stimulation would only occur when food or liquid entered the airway. SLN stimulation produces facilitated trains of swallows and concomitant apnea (inhibition of inspiratory activity), while water infusion stimulates swallows that can occur during any phase of respiration. We have also observed preswallow oral behaviors in response to water infusion, but not SLN stimulation (120). Our results support those of another study in cat by Ono and colleagues (1998) in which they recorded inspiratory caudal reticular formation neurons (some were in/near the DRG and VRG) that are premotor to the hypoglossal motor nucleus (XIIn); some of these neurons also had dual projections to the phrenic motor nucleus. One-sixth of the inspiratory neurons also showed Schluckatmung activity during swallow, which was importantly observed as a result of oral water infusion but not SLN stimulation.

Expansion of the Classic Theory on the Importance of Inspiratory Activity During Swallow

We have formed our present working hypothesis upon the basis of both historical literature and newer experiments. We theorize that the Schluckatmung action of inspiratory muscles (diaphragm, parasternal, etc.) during swallow is necessary to produce a negative intra-thoracic pressure that helps guide the bolus into the esophagus during the pharyngeal phase of swallow (FIGURE 3). To answer Marckwald’s original question, the action of these muscles must be precisely regulated to minimize ongoing aspiration risk. Noteworthy features include the following:

FIGURE 3.

Mechanics of pressure regulation during swallow A: activity of chest wall muscles (e.g., parasternal) and diaphragm expand the thoracic cavity generating negative intrathoracic pressure during inspiration. The trachea and esophagus both reside in the thoracic cavity and are under the same pressure conditions. According to the dual-valve hypothesis (121), the state of the larynx and upper esophageal sphincter determine if air/material moves from the upper airway into the lungs or esophagus. B: inspiration (blue) and swallow (red) in an anesthetized cat. C: mean + SD of transdiaphragmatic pressure (Pdi). The Pdi is the difference between stomach pressure subtracted from esophageal pressure, i.e., a pressure differential (122). These results demonstrate that diaphragm activity during swallow produces a measurable force and does not just constitute “bracing.”

1. During conditions of increased aspiration risk (e.g., stimulation of laryngeal afferents), the Schluckatmung is significantly suppressed. This has been observed in experiments using SLN stimulation (120), and in recordings of subglottic positive pressure during swallow after laryngectomy by direct measurement (123) and esophageal manometry (61, 62);

2. Inspiratory muscle activity during swallow (as measured by EMG) increases with pharyngeal or esophageal simulation (124, 125);

3. Cervical spinal cord injury significantly suppresses Schluckatmung activity (126);

4. Inspiratory premotor neurons in the DRG and VRC participate in the regulation of Schluckatmung (106–111, 127);

5. As Schluckatmung appears to play a larger role in infants, development may represent an important factor (80, 82–85); and

6. Despite swallow predominately occurring during expiration, ongoing expiratory drive is suppressed during swallow (30, 128).

The Swallow Pattern Generator

In the definitive review on deglutition published by Jean (13) in 2001, he describes the “current” knowledge of swallow circuitry with a limited number of central interactions. This simple brainstem circuit consists of sensory processing in the nucleus tractus solitarius (NTS), swallow motor command in the nucleus ambiguus (NA), and activation of motoneurons across various cranial nerve motor nuclei (FIGURE 2C). The swallow pattern generating region is traditionally described with two main groups. Jean’s dorsal swallow group (DSG) is in/near the caudal NTS and surrounding reticular formation (RF), which is also the approximate region of the dorsal respiratory group (DRG); the DSG initiates the swallow command, containing the neurons that trigger, shape, and time the swallow motor pattern. Ventral and somewhat lateral to the DSG, within the RF and just above the NA, is the ventral swallow group (VSG); neurons in the VSG distribute the swallow command to motor nuclei (FIGURE 2). There are few direct projections from the NTS to oral motor nuclei (129), and Jean and colleagues propose that this implicates involvement of interneurons in a trisynaptic swallow reflex pathway. Surprisingly, we have not made major leaps in the understanding of swallow circuitry in several decades (130).

Kleinfeld and colleagues propose an expanded model for the overall organization of behavioral orofacial control, including swallow (131–133). In this architecture, nested anatomical feedback loops are distributed across multiple levels in the sensorimotor system and can interact with multiple pattern generators. Muscles (e.g., masseter, tongue) are driven by motor nuclei (e.g., trigeminal motor nucleus: MoV; hypoglossal motor nucleus: XIIn), which are driven by premotor nuclei (e.g., peritrigeminal, hypoglossal reticular formation), some of which can be influenced by “pre-premotor” nuclei such as the respiratory pattern-forming and modulating regions (e.g., pre-Bötzinger complex, parafacial respiratory group, raphe nuclei, Kolliker-Fuse nucleus; FIGURE 2). They refer to these pre-premotor regions as “pre²motor,” to reflect the fact that the respiratory oscillator(s) can modulate other orofacial premotor oscillators, such as those for whisking and sniffing. These other orofacial pattern generators (chewing, whisking, licking) and their associated premotor regions are found primarily within the medullary intermediate reticular formation (IRt). The reticular formation (RF) is a large and loosely connected group of nuclei that spans the brainstem, and the medullary RF nuclei have also been implicated in swallow function (134–142). In the simplest model of this general circuit for orofacial control, sensory information enters the NTS and then information travels to premotor areas in the more lateral medullary RF (including IRt), and/or in the respiratory premotor regions, and then command for reflexive oromotor behaviors would be produced by motoneuronal pools.

A study by Sugiyama et al. (214) that examined projections of neurons active during swallow provided insight regarding bilateral synchrony. Local projections of these neurons within the NTS form a local neuronal circuit, which likely comprises the DSG. There were also reciprocal connections between in the NTS and the RF, with projections to the contralateral RF. The authors postulate that there is some overlap between the DSG and the VSG. While the classic swallow regions include part of the caudal medullary RF, more rostral portions of the RF have not been included in the traditional models of swallow circuitry. Despite a few reports that supported rostral RF involvement in swallow (134, 137, 139, 143), and despite much supporting anatomical evidence (discussed below), the lack of swallow microelectrode recordings in this region has likely prevented its inclusion in the common trisynaptic swallow circuit models (130, 135).

Jaw elevating/closing and jaw depressing/opening motor neurons, which alternate rhythmically during chewing, are found in distinct subdivisions of the rat trigeminal motor nucleus (MoV) (137), and other motor nuclei also participate in mastication. During swallow, the jaw is typically closed and stabilized by jaw elevators (89, 144, 145). However, during the early pharyngeal phase of swallowing, jaw depressor muscles may also be activated to raise the larynx (146). This is supported by a study in humans in which jaw opening and closing muscles contracted concurrently at the end of mastication and the beginning of pharyngeal swallow (147). The authors proposed a simplified model in which feed-forward inhibition from separate jaw-opening and jaw-closing pattern generators converges on the swallow pattern generator concurrently and sums to “convert” mastication into a swallowing reflex via postinhibitory rebound; this hypothetical model includes afferent feedback from jaw muscle spindles and the bolus. A study in rats also demonstrated swallow suppression during rhythmic jaw movement and speculated a role for jaw muscle spindles, chewing pattern generators, and the amygdala in swallow suppression during mastication (148).

An elegant tracing study by Cunningham and Sawchenko (137) focused on anatomical connections between an extended region of the caudal medullary RF (overlapping in part with the VSG) and swallow motor nuclei. Anterograde labeling identified extensive bilateral fibers projecting from the caudal RF to targets throughout the entire medullary RF. Bilateral labeling ascended rostro-ventrally via the RF to myotopically innervate specific subdivision of swallow cranial nerve motor nuclei: MoV, VIIn, and XIIn (137). Other work had previously shown this (149) and that the DSG region of the NTS connects to the IRt, which provides premotor input to the facial nerve (150, 151). In the Cunningham and Sawchenko study (137), swallowing regions along with jaw-opening and jaw-closing MoV regions were simultaneously innervated by neuronal populations in the medullary RF. They speculated that this may be important for mastication but that this collateral pathway was especially ideal for jaw stabilization and laryngeal protection during swallow. The same study found that motoneurons for tongue protractor and retractor muscles in XIIn also receive robust input from the medullary RF, which would aid in the simultaneous activation of those muscles at the onset of swallow. However, this is likely a complex process involving independent coordinated action of distinct motor pools during swallow. The XIIn motoneurons appear to receive broad excitatory synaptic input from the respiratory pattern generator, resulting in coactivation of protruder and retractor muscles during respiration (152). The Cunningham and Sawchenko study (137) focused on anterograde tracing from the dorsal caudal medullary RF (and NTS), and retrograde tracing from the swallow motor nuclei. The focus on the caudal RF is due to the caudal location of Jean’s classic DSG and VSG. However, the retrograde labeling from the motor nuclei was concordant with the anterograde labeling, identifying premotor neurons throughout nearly the entire range of the medullary RF, not just the limited region of the caudal RF. This suggests that the more rostral RF may not merely function as a conduit for a monosynaptic connection between a caudal VSG and swallow motor nuclei but may itself be a relay site containing additional swallow premotor neurons.

A role for the more rostral medullary RF in swallow is also supported by recent work. In a new study using optical imaging of an isolated brainstem to identify active neurons, we discovered that a large population of rostral medullary IRt neurons (especially in the parafacial region) were active during fictive swallow but not during breathing (29). In that report, we speculated extensively about the potential multifunctional role of IRt neurons in coordinating respiration and different oromotor behaviors. Activity of some RF neurons (including IRt) promotes mastication and feeding during hunger, and under “hunger” conditions, inhibitory neurons in this region appear to simultaneously regulate somatic and sympathetic motor systems (139). This region is thought to be premotor for chewing, and it contains both inhibitory and excitatory neurons that innervate jaw masticatory motoneurons (136, 153–157).

In a tracing study, Stanek et al. (156) showed extensive overlap of jaw-closing and tongue-protruding premotor neurons in the brainstem RF including IRt (extending from the caudal to the rostral areas), especially in the parafacial region. They also showed that groups of muscles with a common action (such as jaw closing or opening) had shared premotor neurons, sometimes bilaterally. The premotor pools were intermixed but never double labeled for different muscles. This raises the possibility that bilateral orofacial behaviors rely on shared premotor pools to produce highly coordinated symmetric movements involving coactivation of target motoneurons for multiple muscles. The authors speculate that these are suckling neurons because the tracing was performed from postnatal days 1–8. Chewing appears at day 12 in rat (158), but the tracing was also replicated at day 8–15, so the same circuit likely switches from suckling to chewing during this developmental period.

Some caudal RF neurons are responsive not only during swallow but also during licking or masticatory movements (142, 159). In the study by Travers et al. (142), lick-rhythmic neurons were found in within the range of the classic DSG and VSG, and in Probst’s nucleus just lateral to the hypoglossal motor nucleus (XIIn). These neurons were additionally found in direct premotor regions of the more rostral RF. Many of the lick-rhythmic neurons were also active during other motor behaviors such as swallow or oral rejection. Certain regions of the more rostral medullary RF contained orosensory neurons or neurons with mixed sensory and motor properties. Those results are consistent with an established view (160) that the more lateral regions of the RF (e.g., parvocellular RF) are sensory, and more medial regions (e.g., IRt) are premotor. The authors postulated that the RF premotor neurons could provide a common substrate for coordinating a variety of ororhythmic alternating behaviors that are activated by multifunctional groups of muscles innervated by multiple motor nuclei. They further speculated that respiratory-related neurons and lick-rhythmic neurons constitute mainly separate populations and that the pattern generators for ingestion and respiration do not necessarily overlap; however, they would be recruited together in some fashion during a threat to airway protection (142).

In the farthest caudal range of the medullary RF is the nucleus retroambiguus (NRA), which is directly caudal to the NA. The NRA has many connections with the NA, the respiratory column, and respiratory parafacial and pontine nuclei (161). The ventral part of the RF in this caudal area is not integral to respiratory pattern generation, but it serves as a relay for respiratory modulation. The NRA region is thought to play a role in active expiration, postinspiration, vocalization (162), and any expulsive behaviors that require increased intra-abdominal or intrathoracic pressure (such as cough, sneeze, or vomit). When isoguvacine (GABA_A agonist/pharmacological lesion) was microinjected into the NRA, swallow motor pattern was not affected, but there was an increase in gating of SLN inputs as demonstrated by a loss of sensitivity to SLN stimulus (163).

The medullary RF likely also influences the diaphragm, as neurons that project to the phrenic motor nucleus are found in the medullary RF, the NTS, and the dorsal and ventral respiratory groups in cats, in keeping with the multifunctional actions of the diaphragm (164, 165). In a study by Macron et al. (166), they demonstrated phrenic nerve afferent projections to the medullary RF, and hypothesized that these interacted with other respiratory and nonrespiratory RF inputs to coordinate various behaviors (166).

Ono and colleagues (167) also showed in cat that aproximately one-fifth of inspiratory-related neurons in the caudal medullary RF (including the NRA) that directly project to XIIn also projected to the phrenic motor nucleus in the spinal cord. The authors of this study speculate that this special population of dual projection premotor neurons are active during behaviors that require synchronized co-contraction of tongue and diaphragm muscles during not only inspiration but also in their responses to other stimuli. They extended this in a subsequent study in which they demonstrated activity of some of the XII premotor neurons during both swallow and respiration, including some with Schluckatmung activity, and concluded that these neurons were responsible for the tongue transition from respiration to ingestion or rejection (127).

Thus, additional portions of the RF should be included in the central swallow network, including RF regions rostral to the classic VSG region of the caudal medulla (168). Given the large number and anatomical distribution of motoneurons throughout the brainstem that participate in swallow, and the multifunctional roles of these motoneurons, the collective area required for premotor coordination would be quite large. It appears to span the majority of the medullary RF, rather than the relatively small caudal RF region that has been traditionally thought to comprise the VSG. Respiratory nuclei in the pons may also play an important role in swallow (141, 169), though it is perhaps modulatory (e.g., swallow-breathing and/or postinspiratory coordination or relay from higher brain centers) rather than generative. For example, oral administration of capsaicin (sensed via the trigeminal nerve in the pons) improves swallow function in older adults and after neurologic injury (170–173). New investigations have provided increasing evidence that the respiratory network is distributed throughout the brainstem and that its constituents have multifunctional roles (25–29), and the techniques used in those studies could be employed to further investigate these other roles. The possibility of a distributed conditional swallow oscillator that subserves other oromotor functions has also been more recently acknowledged (130).

Future Directions

With the advantages of modern experimental tools and an ever-expanding group of basic and clinical scientists, we believe that the controversy over Schluckatmung could be resolved in this generation. Because an extraordinary number of neurodegenerative and neurotraumatic diseases/disorders result in dysphagia and its associated morbidity and mortality, it is imperative to pursue a thorough understanding of the regulation and underlying circuitry of all components of the swallow motor pattern.

First, the interpretation of manometry studies without accompanying EMGs would greatly benefit from a concentration on mechanics of the thoracic cavity. This might prompt functional MRI investigations focusing on movement of the lower thoracic cavity during swallow. Second, swallow researchers collectively should consider the experimental conditions that are used to make classifications of swallow pattern generation and to determine mechanisms across model systems. For example, while SLN stimulation robustly triggers repetitive swallow, use of this tool may result in insufficient assumptions about the lack of participation of a vast array of brainstem neurons in swallow pattern generation. Finally, we should reevaluate the size and breadth of the brainstem swallow pattern generator and consider including regions such as those that control swallow-related activation of inspiratory muscles and a larger range of the medullary reticular formation.

The eloquent studies by Bieger (174–177) and others (178, 179) into the pharmaceutical manipulation of swallow and the profound effect of serotonin (5-HT) agents implicate the raphé nuclei, which are part of the median reticular formation. The net influence of 5-HT on any single cell/region varies widely and depends on the relative presence and distribution of both inhibitory and excitatory 5-HT receptor subtypes. Serotonin often modulates neuron responses to other stimuli (e.g., by affecting glutamate release) and is thought to act in a gain-setting role (180). The known action of serotonin on respiration, for example, is complex and varies by anatomical region and receptor type; it can have direct or indirect effects, inhibitory or facilitatory (181).

Finally, because activation of the diaphragm requires VRC inspiratory neuron activity, the fact that inspiratory muscles (e.g., diaphragm and parasternal) and medullary inspiratory neurons are active during swallow (30, 116–118) necessitates the inclusion of the cervical and thoracic spinal cord in consideration of swallow network models. Future investigations should include evaluations of the specific parts of the VRC that are active during swallow and Schluckatmung, such as the pre-Botzinger complex. The newly identified postinspiratory complex (PiCo), which is found in the rostral medullary RF at the border of the NA, is thought to be a hub for laryngeal and postinspiratory regulation, including during behaviors such as swallow and vocalization (139, 162, 182, 183). Schluckatmung should also be evaluated during experimental conditions in which respiratory drive is increased, such as hypoxia and hypercapnia.

Extensive literature describes the effects of cervical spinal cord injury on breathing-related phrenic nerve and diaphragm recruitment, including work by Porter (184) Goshgarian (185, 186), Sieck and Mantilla (187, 188), Mitchell (189–191), Reier and Lane (192, 193), Fuller (194, 195), Alilain and Silver (196, 197), and others (198–203). However, this knowledge cannot be directly translated to predict changes in swallow-related diaphragm recruitment. While swallow motoneurons reside within the brainstem, our work along with the work of others (16, 21, 22, 31, 32, 41, 119, 120, 128, 204) demonstrates the importance of nonvagal afferent feedback relayed through ascending spinal pathways to maintain homeostatic regulation of swallow. Cervical spinal cord injury impairs swallow (205–213), perhaps through disruption of afferent/efferent balance. We have recently discovered that lateral cervical hemisection in cat bilaterally depresses diaphragm activity during swallow and increases inspiratory laryngeal drive (126); however, the vital knowledge of how the local spinal cord circuits regulate motor drive during swallow remains mostly unexplored. All of these future lines of investigation would benefit greatly from extensive electrophysiological swallow experimentation.

Conclusions

The pharyngeal phase of swallow is a highly regulated behavior that is dependent on precise muscle activation and efficient mechanics. The vast number and variety of conditions and diseases that can cause dysphagia implies complexity and vulnerability of swallow. While the role of the upper airway during swallow has been extensively studied, a full description of investigations on the role of inspiratory muscle activity during swallow could be contained within this review. This element is an area primed for exploration and resolution.