Immunohistochemical study of neuropeptides in vagal and glossopharyngeal afferent neurons in the rat

The presence and distribution of multiple neuropeptides in vagal and glossopharyngeal afferent ganglia of the rat were studied using immunohistochemistry. Substance P-, calcitonin-gene related peptide-, cholecystokinin-, neurokinin A-, vasoactive intestinal polypeptide-, and somatostatin-immunoreactive neurons were detected in each visceral afferent ganglion. Neurotensin-immunoreactive cells were not observed. In the nodose ganglion (inferior ganglion of the vagus nerve) occasional immunoreactive cells were scattered throughout the main (caudal) portion of the ganglion with small clusters of cells seen in the rostral portion. The pattern of distribution of the various peptides in the nodose ganglion was similar, with the exception of vasoactive intestinal polypeptide-immunoreactive neurons which exhibited a more caudal distribution. The relative numbers of immunoreactive cells varied, with the greatest numbers being immunoreactive for substance P or vasoactive intestinal polypeptide, and the lowest numbers being immunoreactive for neurokinin A and somatostatin. A build-up of immunoreactivity for each of the peptides, except somatostatin and neurotensin, was detected in vagal nerve fibers of colchicine-injected ganglia. Numerous peptide-immunoreactive cells were also found in the petrosal (inferior ganglion of the glossopharyngeal nerve) and jugular (superior ganglion of the vagus nerve) ganglia. No specific intraganglionic distribution was noted although the relative numbers of cells which were immunoreactive for the different peptides varied considerably. Substance P and calcitonin-gene related peptide were found in large numbers of cells, cholecystokinin was seen in moderate numbers of cells, and neurokinin A, vasoactive intestinal polypeptide and somatostatin were seen in fewer cells. These data provide evidence for the presence and non-uniform distribution of multiple peptide neurotransmitters in vagal and glossopharyngeal afferent neurons. In general, relatively greater numbers of immunoreactive cells were located in the rostral compared with caudal nodose ganglion, and in the petrosal and jugular ganglia compared with the nodose ganglion. Thus, multiple neuropeptides may be involved as afferent neurotransmitters in the reflexes mediated by vagal and glossopharyngeal sensory nerves.     

Neurons of origin of the internal ramus of the rabbit accessory nerve: localization in the dorsal nucleus of the vagus nerve and the nucleus retroambigualis

The neurons of origin of the internal ramus of the rabbit accessory nerve were identified in the dorsal nucleus of the vagus nerve, using bilateral injections of horseradish peroxidase into the inferior vagal ganglion, soft palate, and pharynx, which were preceded by different combinations of the unilateral intracranial severings of the rootlets of the vagus and glossopharyngeal nerves, those of the cranial root of the accessory nerve, and the trunk of its spinal root. The neurons of origin occupied the caudal four-fifths of the dorsal vagal nucleus extending from about 1.0 mm rostral to the obex as far caudally as the second cervical spinal segment, with their number being about half the total number of neurons of the nucleus. Although considerably fewer, they were also located in the nucleus retroambigualis of the caudal half of the first cervical spinal segment and the second segment. Axons of most internal ramus neurons traversed the rootlets of the cranial accessory root. Axons of the few neurons located more caudally than about 1.0 mm caudal to the obex emerged from the upper cervical spinal cord to run along the trunk of the spinal accessory root before finally joining the internal ramus; caudal to the midlevel of the first cervical segment, the dorsal vagal nucleus and the nucleus retroambigualis contained neurons whose axons followed only that course.     

The vagal connection of the carotid sinus

We applied wheat germ agglutinin-horseradish peroxidase (WGA-HRP) to the adventitia of the left carotid sinus of cats after tying the glossopharyngeal and the carotid sinus nerves. Forty-five to 255 neurons were labeled in the rostral pole of the nodose ganglion, through the pharyngeal branch of the vagus nerve. Transganglionic label was present only in the ipsilateral dorsal and medial subnuclei of the nucleus tractus solitarii at levels nearly rostral to the area postrema. These findings show the existence of a novel pathway from the carotid sinus, projecting via the vagus nerve, to the nucleus tractus solitarii.     

Occipital artery injections of 5-HT may directly activate the cell bodies of vagal and glossopharyngeal afferent cell bodies in the rat

The primary objective of this study was to determine whether circulating factors gain direct access to and affect the activity of vagal afferent cell bodies in the nodose ganglia and glossopharyngeal afferents cell bodies in the petrosal ganglia, of the rat. We found that the occipital and internal carotid arteries provided the sole blood supply to the nodose ganglia, and that i.v. injections of the tracer, Basic Blue 9, elicited strong cytoplasmic staining in vagal and glossopharyngeal afferent cell bodies that was prevented by prior ligation of the occipital but not the internal carotid arteries. We also found that occipital artery injections of 5-HT elicited pronounced dose-dependent reductions in heart rate and diastolic arterial blood pressure that were (1) virtually abolished after application of the local anesthetic, procaine, to the ipsilateral nodose and petrosal ganglia, (2) markedly attenuated after transection of the ipsilateral vagus between the nodose ganglion and brain and virtually abolished after subsequent transection of the ipsilateral glossopharyngeal nerve between the petrosal ganglion and the brain, (3) augmented after ipsilateral transection of the aortic depressor and carotid sinus nerves, and (4) augmented after transection of all ipsilateral glossopharyngeal and vagal afferent nerves except for vagal cardiopulmonary afferents. These findings suggest that blood-borne 5-HT in the occipital artery gains direct access to and activates the cell bodies of vagal cardiopulmonary afferents of the rat and glossopharyngeal afferents of undetermined modalities.     

The role of the solitary and paramedian reticular nuclei in mediating cardiovascular reflex responses from carotid baro- and chemoreceptors

1. With dye-filled micro-electrodes single neurones in the medulla of anaesthetized paralysed cats were identified which: (a) fired rhythmically in synchrony with or were modulated by the cardiac cycle, and which ceased firing with occlusion of the ipsilateral common carotid artery (carotid sinus baroreceptor neurones); (b) were excited by stimulation of carotid body chemoreceptors by close intra-arterial injection of lobeline into the thyroid artery (carotid body chemoreceptor neurones).2. Twelve carotid baroreceptor neurones were identified, in thirty-three cats, nine of which were localized in the intermediate area of the nucleus of the solitary tract (NTS) within 1 mm ahead of or behind the obex; three units were located either in the parahypoglossal area or the dorsal portion of the paramedian reticular nucleus (PRN).3. Of the twenty-one carotid chemoreceptor neurones which were identified, thirteen were localized in the NTS, three in the parahypoglossal area and four in the dorsal PRN.4. Bilateral lesions of the paramedian reticular area of medulla destroying the PRN, abolished or reversed the depressor response to electrical stimulation of myelinated fibres of the carotid sinus nerve (CSN), attenuated the depressor response to carotid sinus stretch and augmented the pressor response to chemoreceptor stimulation by lobeline. Such lesions did not significantly alter the reflex heart rate responses.5. Small lesions of the NTS within an area 1 mm rostral to the obex abolished all reflex blood pressure and heart rate responses to electrical stimulation of the CSN or natural stimulation of carotid baro- or chemoreceptors.6. Baroreceptors and chemoreceptors of the CSN project both to the intermediate zone of the NTS and to more medial areas of the medulla, particularly the dorsal PRN and parahypoglossal area.7. The PRN serves to mediate the reflex depressor, but not cardio-vagal, response from myelinated baroreceptors and buffers the pressor responses from chemoreceptors; it may serve as an important area integrating cardiovascular activity descending from forebrain, brain stem and cerebellum with baroreceptor reflexes.8. Cardiovascular reflex responses arising from non-myelinated baroreceptors and all chemoreceptors are mediated by neurones in the intermediate area of the NTS.     

The Carotid Sinus Nerve—Structure,Function, and Clinical Implications

Interest has been renewed in the anatomy and physiology of the carotid sinus nerve (CSN) and its targets (carotid sinus and carotid body, CB), due to recent proposals of surgical procedures for a series of common pathologies, such as carotid sinus syndrome, hypertension, heart failure, and insulin resistance. The CSN originates from the glossopharyngeal nerve soon after its appearance from the jugular foramen. It shows frequent communications with the sympathetic trunk (usually at the level of the superior cervical ganglion) and the vagal nerve (main trunk, pharyngeal branches, or superior laryngeal nerve). It courses on the anterior aspect of the internal carotid artery to reach the carotid sinus, CB, and/or intercarotid plexus. In the carotid sinus, type I (dynamic) carotid baroreceptors have larger myelinated A-fibers; type II (tonic) baroreceptors show smaller A- and unmyelinated C-fibers. In the CB, afferent fibers are mainly stimulated by acetylcholine and ATP, released by type I cells. The neurons are located in the petrosal ganglion, and centripetal fibers project on to the solitary tract nucleus: chemosensory inputs to the commissural subnucleus, and baroreceptor inputs to the commissural, medial, dorsomedial, and dorsolateral subnuclei. The baroreceptor component of the CSN elicits sympatho-inhibition and the chemoreceptor component stimulates sympatho-activation. Thus, in refractory hypertension and heart failure (characterized by increased sympathetic activity), baroreceptor electrical stimulation, and CB removal have been proposed. Instead, denervation of the carotid sinus has been proposed for the “carotid sinus syndrome.” Anat Rec, 302:575–587, 2019. © 2018 Wiley Periodicals, Inc.     

Intermingling of central and peripheral nervous tissues in rat dorsolateral vagal rootlet transitional zones

Summary The morphology of the CNS-PNS transitional zone of adult rat dorsolateral vagus nerve rootlets is uniquely complex. A typical rootlet contains a transitional zone over 300 m long, consisting of a central tissue projection extending distally into each rootlet and a peripheral tissue insertion extending for a longer distance deep into the brainstem. The peripheral tissue insertion is continuous with the peripheral tissue of the free rootlet through channels traversing or running parallel to the central tissue projection. Accordingly, the vagal CNS-PNS interface is topologically much more complex than that found elsewhere. In some rootlets the peripheral tissue in the brainstem constitutes an isolated island deep within the neuraxis. In others, peripheral continuity is established only through a cross connection with the peripheral tissue insertion of a neighbouring rootlet. About one fifth of all vagal myelinated axons alternate between the CNS and PNS tissue compartments. This distinguishes the vagus from all other nerves studied to date. These axons are myelinated by Schwann cells distal to the transitional zone, by oligodendrocytes in the central tissue projection and by one or more short intercalated Schwann internodes further centrally, mostly in the peripheral tissue insertion, where their perikarya commonly form closely apposed aggregates. More than four fifths of all unmyelinated axon bundles alternate between central and peripheral tissue compartments, commonly more than once. In the peripheral tissue insertion axons are enveloped by series of non-myelinating Schwann cells. Schwann processes commonly extend for over 50 (Am into the central compartment at each central-peripheral transition. Around one fifth of peripherally unmyelinated axons have an oligodendrocytic sheath in the central compartment. Of those axons possessing more than one intercalated Schwann internade, over one quarter display alternation of myelinated and unmyelinated segments in the peripheral tissue insertion. Astrocytes in the transitional zone segregate PNS tissue, a role played by sheath cells further peripherally in the vagal rootlets. Astrocytes form the surface limiting membranes of the central tissue projection and the barrier between the peripheral tissue insertion and the surrounding brainstem. The barrier consists only of an attenuated layer of processes. This is deficient in places, where oligodendrocytic myelin sheaths are directly exposed to the endoneurial space of the peripheral tissue insertion and in some instances are apposed to myelinating or non-myelinating Schwann cells. Such communication between the central and peripheral compartments is unique to the vagal transitional zone. The findings are consistent with a range of possible events during development. For example, considerable migration and intermingling of central and peripheral tissues, possible overgrowth of rootlet segments by developing myelencephalic tissue, failure of part of the neural crest to separate from the developing neural tube, and the origin of peripheral tissue insertion Schwann cells from the neural tube, or combinations of these factors.     

An HRP study of location of the rabbit palatopharyngeal motoneurons and peripheral course of their axons

The location of the rabbit palatopharyngeal motoneurons and the peripheral course of their axons were investigated with intramuscular injection of HRP either in the normal animal or in conjunction with intracranial severing of the vagal rootlets. Labeled palatopharyngeal motoneurons were ipsilaterally located within a subdivision of the nucleus ambiguus which is formed by a compact arrangement of the smallest neurons of the nucleus and situated in the rostral of the nucleus. We named that subdivision the compact cell group (CoG). The labeled motoneurons occupied the caudal half of the CoG at a level from about 500 to 1,900 microns rostral to the obex. Labeling was completely abolished by severing the vagal rootlets, indicating that the axons of the palatopharyngeal motoneurons traversed the vagal rootlets.     

Anatomic and topographical characteristics of the correlations between glossopharyngeal nerve and autonomic nervous system in Canidae and Mustelidae

Using macro-microanatomical approach, the study of glossopharyngeal nerve was performed on carcasses of fur animals belonging to Canidae and Mustelidae families (mink, sable, sable polar fox, fox). The species peculiarities of interconnections of tympanic nerve with inner carotid nerve and ear branch of vagus nerve were established. The characteristics of the course and connection of the carotid sinus branch of glossopharyngeal nerve with sympathetic branches of cranial cervical ganglion, were defined. The variants of total absence of vagal pharyngeal branch with significant development of similar branch of glossopharyngeal nerve, were noted.     

Calcitonin gene-related peptide immunoreactive sensory neurons in the vagal and glossopharyngeal ganglia innervating the larynx of the rat

We have examined whether calcitonin gene-related peptide-immunoreactive (CGRP-ir) neurons in the vagal and glossopharyngeal ganglia innervate the larynx. Many CGRP-ir neurons were located mostly in the superior glossopharyngeal–jugular ganglion complex that was fused the superior glossopharyngeal ganglion and the jugular ganglion in the cranial cavity. When Fluorogold was applied to the cut end of the superior laryngeal nerve (SLN) or the recurrent laryngeal nerve (RLN), many Fluorogold-labeled neurons were found in the superior glossopharyngeal–jugular ganglion complex and the nodose ganglion. Double-labeling for CGRP and Fluorogold showed that about 80% of Fluorogold-labeled neurons in the superior glossopharyngeal–jugular ganglion complex expressed CGRP-like immunoreactivity in the case of application to the SLN, and about 50% of Fluorogold-labeled neurons expressed CGRP-like immunoreactivity in the case of the RLN. Only a few double-labeled neurons were found in the nodose ganglion. The number of the Fluorogold-labeled neurons and double-labeled neurons in the superior glossopharyngeal–jugular ganglion complex in the case of the SLN was larger than that in the case of the RLN. These results indicate that sensory information from the larynx might be conveyed by many CGRP-ir neurons located in the superior glossopharyngeal–jugular ganglion complex by way of the SLN and the RLN.     

Vagal nerve endings in visceral pleura and triangular ligaments of the rat lung

The inner thoracic cavity is lined by the parietal pleura, and the lung lobes are covered by the visceral pleura. The parietal and visceral plurae form the pleural cavity that has negative pressure within to enable normal respiration. The lung tissues are bilaterally innervated by vagal and spinal nerves, including sensory and motor components. This complicated innervation pattern has made it difficult to discern the vagal vs. spinal processes in the pulmonary visceral pleura. With and without vagotomy, we identified vagal nerve fibres and endings distributed extensively in the visceral pleura (‘P’‐type nerve endings) and triangular ligaments (‘L’‐type nerve endings) by injecting wheat germ agglutinin‐horseradish peroxidase as a tracer into the nucleus of solitary tract or nodose ganglion of male Sprague–Dawley rats. We found the hilar and non‐hilar vagal pulmonary pleural innervation pathways. In the hilar pathway, vagal sub‐branches enter the hilum and follow the pleural sheet to give off the terminal arborizations. In the non‐hilar pathway, vagal sub‐branches run caudally along the oesophagus and either directly enter the ventral‐middle‐mediastinal left lobe or follow the triangular ligaments to enter the left and inferior lobe. Both vagi innervate: (i) the superior, middle and accessory lobes on the ventral surfaces that face the heart; (ii) the dorsal‐rostral superior lobe; (iii) the dorsal‐caudal left lobe; and (iv) the left triangular ligament. Innervated only by the left vagus is: (i) the ventral‐rostral and dorsal‐rostral left lobe via the hilar pathway; (ii) the ventral‐middle‐mediastinal left lobe and the dorsal accessory lobe that face the left lobe via the non‐hilar pathway; and (iii) the ventral‐rostral inferior lobe that faces the heart. Innervated only by the right vagus, via the non‐hilar pathway, is: (i) the inferior (ventral and dorsal) and left (ventral only) lobe in the area near the triangular ligament; (ii) the dorsal‐middle‐mediastinal left lobe; and (iii) the right triangular ligament. Other regions innervated with unknown vagal pathways include: (i) the middle lobe that faces the superior and inferior lobe; (ii) the rostral‐mediastinal inferior lobe that faces the middle lobe; and (iii) the ventral accessory lobe that faces the diaphragm. Our study demonstrated that most areas that face the dorsal thoracic cavity have no vagal innervation, whereas the interlobar and heart‐facing areas are bilaterally or unilaterally innervated with a left‐rostral vs. right‐caudal lateralized innervation pattern. This innervation pattern may account for the fact that the respiratory regulation in rats has a lateralized right‐side dominant pattern.     

Restoration of reflex ventilatory response to hypoxia after removal of carotid bodies in the cat

In mammals there are two sets of peripheral arterial chemoreceptors, the carotid bodies innervated by the sinus branch of the glossopharyngeal nerve and the aortic bodies innervated by the vagus nerves. The afferent impulse discharge from both receptors increases during hypoxia and there is a reflexly mediated increase in ventilation (hypoxic hyperventilation). In the present study we tested this response by exposing anesthetized cats to decreased inspired O₂ concentration before and up to 315 days after bilateral resection of the carotid bodies. Acutely after removing the carotid bodies, hypoxic hyperventilation was abolished. This observation supports the view that the reflex pathway from the aortic body receptors normally contributes minimally to hypoxic hyperventilation. Subsequently, there was a restoration of hypoxic hyperventilation. Restoration was significant 30–43 days after removing the carotid bodies, it reached 70% of the preoperative value at 93–111 days and was essentially complete in terminal experiments 260–315 days after carotid body resection. In terminal experiments, hypoxic hyperventilation was not affected by recutting the regenerated carotid sinus nerves but was abolished completely by bilateral transection of the cervical vagosympathetic trunks. The restored ventilatory response was due predominantly to an increase in rate of breathing while an increase in tidal volume was predominant before carotid body resection. Resting ventilation breathing room air was not consistently decreased after carotid body resection while expired CO₂ was elevated from day 20 to day 111 and at the preoperative level in terminal experiments.It is concluded that restoration of hypoxic hyperventilation in the cat after carotid body resection is mediated by the reflex pathway from aortic body chemoreceptors. The possible contribution of chemo-receptive regenerated carotid sinus nerve axons was excluded. It is suggested that restoration may be a consequence of the central reorganization of chemoreceptor afferent pathways consequent to interruption of the carotid body reflex pathway and that as a result the ‘gain’ of the aortic body ventilatory chemoreflex is enhanced.     

On the direct and crossed components of reflex responses to unilateral stimulation of the carotid body chemoreceptors in the dog

1. In dogs under chloralose-urethane anaesthesia the chemoreceptors of the two carotid bodies were separately stimulated.2. The distribution of three primary reflex responses to carotid body stimulation was studied: parasympathetic bradycardia, sympathetic vasoconstriction, and increase in somatic phrenic nerve activity.3. The reflex bradycardia evoked by either carotid body was mediated by both vagus nerves, but when either vagus was blocked a greater response could be obtained from the contralateral than from the ipsilateral carotid body.4. The reflex vasoconstriction evoked by either carotid body was seen in both hind limbs, with no predominance in either limb.5. The reflex increase in phrenic nerve activity evoked by either carotid body was seen in both phrenic nerves, with no predominance in either nerve.     

Nucleus ambiguus of the rabbit: Cytoarchitectural subdivision and myotopical and neurotopic representations

The cytoarchitectural subdivisions of the nucleus ambiguus of the rabbit and its myotopical and neurotopical representations were investigated with HRP labeling. The nucleus was subdivided into the compact cell group (CoG), the medial and lateral scattered cell groups (SGm and SGl), and the diffuse cell group (DiG). The CoG was formed by esophageal, pharyngeal constrictor, and palatal motoneurons in the rostral half of the nucleus. The SGm and SGl were located medial and lateral to the CoG, respectively, in the rostral one-third of the nucleus. Stylopharyngeal and cricothyroid motoneurons were located in the most rostral one-fifth of the SGm and the remaining four-fifths, respectively, whereas the SGl was not labeled with HRP injections into the palatal, pharyngeal, esophageal, and laryngeal muscles. The DiG was formed by recurrent laryngeal motoneurons in the caudal two-thirds of the nucleus. Neurons of origin for the glossopharngeal nerve occupied the stylopharyngeal region, with a few of them scattered in the CoG and SGl. Neurons giving rise to axons in the superior laryngeal nerve occupied the cricothyroid region, with a few of them scattered in the pharyngeal constrictor region; whereas the pharyngeal vagal branch originated from the pharyngeal constrictor and palatal regions. Neurons of the DiG, SGl, and esophageal region contributed to the infranodosal vagus nerve; esophageal fibers of the recurrent laryngeal nerve originated from the dorsal esophageal region. Laryngeal fibers of the recurrent laryngeal nerve originated from the DiG, the caudal neurons of which had axons traversing the cranial accessory root. © 1993 Wiley-Liss, Inc.     

Calcitonin gene-related peptide immunoreactivity in the nucleus of the tractus solitarius and the carotid receptors of the cat originates from peripheral afferents.

The presence and distribution of the calcitonin gene-related peptide was studied, using immunohistochemical techniques, in carotid receptors, in the nodose and glossopharyngeal ganglia and in the nucleus tractus solitarii of the cat. Seventy-seven per cent of the 42% of the nodose ganglion cells were labeled. Fine, sparsely branched immunoreactive terminal axonal arborizations were found in the carotid body; they disappeared after petrosal ganglionectomy. The intense immunoreactivity present in fibers in the commissural, medial, interstitial, gelatinosus, dorsal, intermediate and rostral gustatory subnuclei of the nucleus tractus solitarius was drastically reduced after removal of the ipsilateral nodose and petrosal ganglia. The central distribution of the immunoreactive axons, the morphology of the terminals in the carotid receptors and their dependence on an intact peripheral innervation are consistent with the idea that in the cat the calcitonin gene-related peptide is present in a high proportion of the primary visceral afferents, most of them unmyelinated.     

Pituitary adenylatecyclase-activating polypeptide-immunoreactive nerve fibers in the rat epiglottis and pharynx

The distribution of pituitary adenylatecyclase-activating polypeptide-immunoreactive (PACAP-IR) nerve fibers was studied in the rat epiglottis and pharynx. PACAP-IR nerve fibers were located beneath the mucous epithelium, and occasionally penetrated the epithelium. These nerve fibers were abundant on the laryngeal side of the epiglottis and in the dorsal and lateral border region between naso-oral and laryngeal parts of the pharynx. PACAP-IR nerve fibers were also detected in taste buds within the epiglottis and pharynx. In addition, many PACAP-IR nerve fibers were found around acinar cells and blood vessels. The double immunofluorescence method demonstrated that distribution of PACAP-IR nerve fibers was similar to that in CGRP-IR nerve fibers in the epithelium and taste bud. However, distributions of PACAP-IR and CGRP-IR nerve fibers innervating mucous glands and blood vessels were different. The retrograde tracing method also demonstrated that PACAP and CGRP were co-expressed by vagal and glossopharyngeal sensory neurons innervating the pharynx. These findings suggest that PACAP-IR nerve fibers in the epithelium and taste bud of the epiglottis and pharynx which originate from the vagal and glossopharyngeal sensory ganglia include nociceptors and chemoreceptors. The origin of PACAP-IR nerve fibers which innervate mucous glands and blood vessels may be the autonomic ganglion.     

Electron microscopic and electrophysiological studies on the carotid body following intracranial section of the glossopharyngeal nerve

1. The innervation of carotid body Type I cells has been investigated in seventeen cats. At a sterile operation the glossopharyngeal and vagus nerve roots were cut intracranially on one side.

2. From 1½ to 378 days after the operation the carotid bodies were fixed in situ and prepared for electron microscopy. Nerve endings on Type I cells were found to degenerate with a prolonged time course. In each cat there was a decrease in the number of nerve endings on the operated side as compared with the non-operated side.

3. Before the carotid bodies were fixed, recordings were made from chemoreceptor, and baroreceptor, afferent fibres in the sinus nerve on the operated side. The chemoreceptors responded in the usual way to changes in arterial O₂, CO₂ and pH; the injection of cyanide evoked a brisk response.

4. It is concluded that the nerve endings on Type I cells are efferent rather than afferent and the cell bodies of their axons are probably in the brain stem.