Extensive projections from the midbrain periaqueductal gray to the caudal ventrolateral medulla: a retrograde and anterograde tracing study in the rat

Orchestration of innate and conditioned defensive actions by the periaqueductal gray

The midbrain periaqueductal gray (PAG) has been recognized for decades as having a central role in the control of a wide variety of defensive responses. Initial discoveries relied primarily on lesions, electrical stimulation and pharmacology. Recent developments in neural activity imaging and in methods to control activity with anatomical and genetic specificity have revealed additional streams of data informing our understanding of PAG function. Here, we discuss both classic and modern studies reporting on how PAG-centered circuits influence innate as well as learned defensive actions in rodents and humans. Though early discoveries emphasized the PAG's role in rapid induction of innate defensive actions, emerging new data indicate a prominent role for the PAG in more complex processes, including representing behavioral states and influencing fear learning and memory. This article is part of the Special Issue on "Fear, Anxiety and PTSD".

Electrophysiology as a Tool to Decipher the Network Mechanism of Visceral Pain in Functional Gastrointestinal Disorders

Abdominal pain, including visceral pain, is prevalent in functional gastrointestinal (GI) disorders (FGIDs), affecting the overall quality of a patient's life. Neural circuits in the brain encode, store, and transfer pain information across brain regions. Ascending pain signals actively shape brain dynamics; in turn, the descending system responds to the pain through neuronal inhibition. Pain processing mechanisms in patients are currently mainly studied with neuroimaging techniques; however, these techniques have a relatively poor temporal resolution. A high temporal resolution method is warranted to decode the dynamics of the pain processing mechanisms. Here, we reviewed crucial brain regions that exhibited pain-modulatory effects in an ascending and descending manner. Moreover, we discussed a uniquely well-suited method, namely extracellular electrophysiology, that captures natural language from the brain with high spatiotemporal resolution. This approach allows parallel recording of large populations of neurons in interconnected brain areas and permits the monitoring of neuronal firing patterns and comparative characterization of the brain oscillations. In addition, we discussed the contribution of these oscillations to pain states. In summary, using innovative, state-of-the-art methods, the large-scale recordings of multiple neurons will guide us to better understanding of pain mechanisms in FGIDs.

The PV2 cluster of parvalbumin neurons in the murine periaqueductal gray: connections and gene expression

The PV2 (Celio 1990), a cluster of parvalbumin-positive neurons located in the ventromedial region of the distal periaqueductal gray (PAG) has not been previously described as its own entity, leading us to study its extent, connections, and gene expression. It is an oval, bilateral, elongated cluster composed of approximately 475 parvalbumin-expressing neurons in a single mouse hemisphere. In its anterior portion it impinges upon the paratrochlear nucleus (Par4) and in its distal portion it is harbored in the posterodorsal raphe nucleus (PDR). It is known to receive inputs from the orbitofrontal cortex and from the parvafox nucleus in the ventrolateral hypothalamus. Using anterograde tracing methods in parvalbumin-Cre mice, the main projections of the PV2 cluster innervate the supraoculomotor periaqueductal gray (Su3) of the PAG, the parvafox nucleus of the lateral hypothalamus, the gemini nuclei of the posterior hypothalamus, the septal regions, and the diagonal band in the forebrain, as well as various nuclei within the reticular formation in the midbrain and brainstem. Within the brainstem, projections were discrete, but involved areas implicated in autonomic control. The PV2 cluster expressed various peptides and receptors, including the receptor for Adcyap1, a peptide secreted by one of its main afferences, namely, the parvafox nucleus. The expression of GAD1 and GAD2 in the region of the PV2, the presence of Vgat-1 in a subpopulation of PV2-neurons as well as the coexistence of GAD67 immunoreactivity with parvalbumin in terminal endings indicates the inhibitory nature of a subpopulation of PV2-neurons. The PV2 cluster may be part of a feedback controlling the activity of the hypothalamic parvafox and the Su3 nuclei in the periaqueductal gray.

The orbitofrontal cortex projects to the parvafox nucleus of the ventrolateral hypothalamus and to its targets in the ventromedial periaqueductal grey matter

Although connections between the orbitofrontal cortex (OFC)-the seat of high cognitive functions-the lateral hypothalamus and the periaqueductal grey (PAG) have been recognized in the past, the precise targets of the descending fibres have not been identified. In the present study, viral tracer-transport experiments revealed neurons of the lateral (LO) and the ventrolateral (VLO) OFC (homologous to part of Area 13 in primates) to project to a circumscribed region in the ventrolateral hypothalamus, namely, the horizontally oriented, cylindrical parvalbumin- and Foxb1-expressing (parvafox) nucleus. The fine collaterals stem from coarse axons in the internal capsule and form excitatory synapses specifically with neurons of the parvafox nucleus, avoiding the rest of the hypothalamus. In its further caudal course, this contingent of LO/VLO-axons projects collaterals to the Su3- and the PV2 nuclei, which lie ventral to the aqueduct in the (PAG), where the terminals fields overlap those deriving from the parvafox nucleus itself. The targeting of the parvafox nucleus by the LO/VLO-projections, and the overlapping of their terminal fields within the PAG, suggest that the two cerebral sites interact closely. An involvement of this LO/VLO-driven circuit in the somatic manifestation of behavioural events is conceivable.

Ventrolateral periaqueductal grey matter neurotransmission modulates cardiac baroreflex activity

Baroreflex activity is a neural mechanism responsible for short-term adjustments in blood pressure (BP). Several supramedullary areas, which send projections to the medulla, are able to control this reflex. In this context, the ventrolateral part of the periaqueductal grey matter (vlPAG), which is a mesencephalic structure, has been suggested to regulate the cardiovascular system. However, its involvement in baroreflex control has never been addressed. Therefore, our hypothesis is that the vlPAG neurotransmission is involved in baroreflex cardiac activity. Male Wistar rats had stainless steel guide cannulae unilaterally or bilaterally implanted in the vlPAG. Afterward, a catheter was inserted into the femoral artery for BP and HR recording. A second catheter was implanted into the femoral vein for baroreflex activation. When the nonselective synaptic blocker cobalt chloride (CoCl₂ ) was unilaterally injected into the vlPAG, in either the left or the right hemisphere, it increased the tachycardic response to baroreflex activation. However, when CoCl₂ was bilaterally microinjected into the vlPAG it decreased the tachycardic response to baroreflex stimulation. This work shows that vlPAG neurotransmission is involved in modulation of the tachycardic response of the baroreflex. Moreover, we suggest that the interconnections between the vlPAG of both hemispheres are activated during baroreflex stimulation. In this way, our work helps to improve the understanding about brain-heart circuitry control, emphasizing the role of the autonomic nervous system in such modulation.

The nucleus para-retroambiguus: a new group of estrogen receptive cells in the caudal ventrolateral medulla of the female golden hamster

Receptive female hamsters display very rigid lordotic postures. Estradiol facilitates this behavior via activation of estrogen receptors. In the hamster brainstem estrogen receptor-alpha-immunoreactive neurons (ER-alpha-IR) are present in various brainstem regions including nucleus retroambiguus (NRA) in the caudal ventrolateral medulla (CVLM) and nucleus of the solitary tract. ER-alpha-IR neurons in the CVLM project to the thoracic and upper lumbar cord. However, A1 neurons in this region do not project to the spinal cord, in contrast to overlapping C1 neurons. The question now arises: are ER-alpha-IR cells in the CVLM part of the A1/C1 group, or do they belong to the NRA or do they compose a separate cluster. A study in ovariectomized female hamsters using a combination of double immunostaining and retrograde tracing techniques and measurement of soma diameters was carried out. The results showed that A1/C1 neurons in the CVLM are almost never ER-alpha-positive; neurons inside or bordering the NRA can be divided in two different types: large multipolar and small; the large NRA-neurons, projecting caudally, are neither tyrosine hydroxylase- (TH) nor ER-alpha-IR; the small neurons, bordering the NRA and projecting caudally, are ER-alpha-IR but not TH-IR. From the available evidence and the present findings it can be concluded that the group of small ER-alpha-IR neurons in the CVLM has to be considered as a distinct entity, probably involved in the autonomic physiological changes concurring with successive phases of the estrous cycle. Because the location is closely related to the NRA itself the nucleus is called nucleus para-retroambiguus, abbreviated (NPRA).

Physiology of neuronal subtypes in the respiratory-vocal integration nucleus retroamigualis of the male zebra finch

Learned vocalizations, such as bird song, require intricate coordination of vocal and respiratory muscles. Although the neural basis for this coordination remains poorly understood, it likely includes direct synaptic interactions between respiratory premotor neurons and vocal motor neurons. In birds, as in mammals, the medullary nucleus retroambigualis (RAm) receives synaptic input from higher level respiratory and vocal control centers and projects to a variety of targets. In birds, these include vocal motor neurons in the tracheosyringeal part of the hypoglossal motor nucleus (XIIts), other respiratory premotor neurons, and expiratory motor neurons in the spinal cord. Although various cell types in RAm are distinct in their anatomical projections, their electrophysiological properties remain unknown. Furthermore, although prior studies have shown that RAm provides both excitatory and inhibitory input onto XIIts motor neurons, the identity of the cells in RAm providing either of these inputs remains to be established. To characterize the different RAm neuron types electrophysiologically, we used intracellular recordings in a zebra finch brain stem slice preparation. Based on numerous differences in intrinsic electrophysiological properties and a principal components analysis, we identified two distinct RAm neuron types (types I and II). Antidromic stimulation methods and intracellular staining revealed that type II neurons, but not type I neurons, provide bilateral synaptic input to XIIts. Paired intracellular recordings in RAm and XIIts further indicated that type II neurons with a hyperpolarization-dependent bursting phenotype are a potential source of inhibitory input to XIIts motor neurons. These results indicate that electrically distinct cell types exist in RAm, affording physiological heterogeneity that may play an important role in respiratory-vocal signaling.

Identification of neural pathways involved in genital reflexes in the female: a combined anterograde and retrograde tracing study

The medial preoptic area (MPOA) is important for reproductive behavior in females. However, the descending pathways mediating these responses to the spinal motor output are unknown. The MPOA does not directly innervate the spinal cord. Therefore, pathways mediating MPOA-induced changes in sexual behavior must relay in the brain. The nucleus paragigantocellularis (nPGi) projects heavily to spinal circuits involved in female sexual reflexes and is involved in the tonic inhibition of genital reflexes. However, the periaqueductal gray (PAG) is also important for female sexual behavior. The present study examined the hypothesis that the MPOA output relays through PAG and the nPGi before descending to the spinal cord. We used anterograde and retrograde tracing techniques to examine the descending pathways and relay sites from the MPOA to the spinal cord and the nPGi in the female rat. Injection of biotinylated dextran amine into the MPOA produced dense labeling in specific regions of the PAG and Barrington's nucleus; anterogradely labeled fibers terminated close to neurons retrogradely labeled from the spinal cord in the PAG, Barrington's nucleus, nPGi, lateral hypothalamus and paraventricular nucleus (PVN). Anterogradely labeled fibers and varicosities were also found close to neurons retrogradely labeled from the nPGi in the PAG, lateral hypothalamus and PVN. These results suggest that the major MPOA output relays in the PAG and nPGi before descending to innervate spinal circuits regulating female genital reflexes and that the MPOA plays a multifaceted role in female reproductive behavior through its modulation of PAG output systems.

Activation of the midbrain periaqueductal gray induces airway smooth muscle relaxation

In this study, we examined effects of chemical stimulation of the ventrolateral region of the midbrain periaqueductal gray (vl PAG) on airway smooth muscle tone. We observed that in anesthetized, paralyzed, and artificially ventilated ferrets, vl PAG stimulation elicited airway smooth muscle relaxation. To clarify the mechanisms underlying this observation, we examined the GABA-GABA(A) receptor signaling pathway by 1) examining the expression of GABA(A) receptors on airway-related vagal preganglionic neurons (AVPNs) located in the rostral nucleus ambiguus region (rNA), by use of receptor immunochemistry and confocal microscopy; 2) measuring GABA release within the rNA by using microdialysis; and 3) performing physiological experiments to determine the effects of selective blockade of GABA(A) receptors expressed by AVPNs in the rNA region on vl PAG-induced airway relaxation, thereby defining the role of the GABA(A) receptor subtype in this process. We observed that AVPNs located in the rNA region do express the GABA(A) receptor beta-subtype. In addition, we demonstrated that activation of vl PAG induced GABA release within the rNA region, and this release was associated with airway smooth muscle relaxation. Blockade of the GABA(A) receptor subtype expressed by AVPNs in the rNA by bicuculline diminished the inhibitory effects of vl PAG stimulation on airway smooth muscle tone. These data indicate, for the first time, that activation of vl PAG dilates the airways by a release of GABA and activation of GABA(A) receptors expressed by AVPNs.

Brain afferents to the medullary dorsal reticular nucleus: a retrograde and anterograde tracing study in the rat

The medullary dorsal reticular nucleus (DRt) was recently shown to belong to the supraspinal pain control system; neurons within this nucleus give origin to a descending projection that increases spinal nociceptive transmission and facilitates pain perception [Almeida et al. (1999), Eur. J. Neurosci., 11, 110-122]. In the present study, the areas of the brain that may modulate the activity of DRt neurons were investigated by using of tract-tracing techniques. Injection of a retrograde tracer into the DRt resulted in labelling in multiple areas of the brain. In the contralateral orbital, prelimbic, infralimbic, insular, motor and somatosensory cortices labelling was prominent, but a smaller ipsilateral projection from these same areas was also detected. Strong labelling was also noted in the central amygdaloid nucleus, bed nucleus of stria terminalis and substantia innominata. Labelled diencephalic areas were mainly confined to the hypothalamus, namely its lateral and posterior areas as well as the paraventricular nucleus. In the mesencephalon, the periaqueductal grey, red nucleus and deep mesencephalic nucleus were strongly labelled, whereas, in the brainstem, the parabrachial nuclei, rostroventromedial medulla, nucleus tractus solitarius, spinal trigeminal nucleus, and the parvocellular, dorsal, lateral and ventral reticular nuclei were the most densely labelled regions. All deep cerebellar nuclei were labelled bilaterally. These data suggest that the DRt integrates information from the somatosensory, antinociceptive, autonomic, limbic, pyramidal and extrapyramidal systems while triggering its descending facilitating action upon the spinal nociceptive transmission.

Functional abnormalities of the cervical cord and lower medulla and their effect on pain: observations in chronic pain patients with incidental mild Chiari I malformation and moderate to severe cervical cord compression

OBJECTIVE

Abnormalities of central sensory processing may play a role in the pathogenesis of chronic pain. The Chiari I malformation is a congenital hindbrain anomaly characterized by protrusion of the cerebellar tonsils into the upper cervical canal, with variable effects on the lower brain stem and cervical cord. The purpose of this study was to compare sensory function and pain among patients with chronic pain who had these disorders incidentally diagnosed, to assess the effect on pain in these patients in comparison with those without central nervous system disease.

DESIGN

Retrospective study in which pain, mood, and sensory function in 32 patients with chronic pain who had mild Chiari I malformation were compared with that in 53 patients with chronic pain who had moderate to severe compression of the cervical spinal cord and 52 patients with chronic pain who had no apparent central nervous system disorder. Data had been collected previously as part of standard clinical assessments, including clinical neurological examinations, quantitative sensory testing, pain drawings, and psychometric testing with the Symptom Checklist 90.

PATIENTS

All subjects were patients of a hospital-based pain management practice who had been accepted for treatment over a 5-year period.

RESULTS

Both the Chiari I and cervical compression groups had long tract signs evident on clinical neurological examination. Quantitative sensory testing indicated elevations in the trigeminal territory among patients with Chiari I malformation and on the neck, hands, and feet in both the Chiari I and cervical compression groups. The extent of pain and mood disturbance was greatest in the Chiari I group and least in the group with no central nervous system disorder. Complex regional pain syndrome, fibromyalgia, and temporal mandibular joint disorder were more common among the Chiari I malformation group than among the other groups.

CONCLUSIONS

Quantitative sensory analysis indicates sensory dysfunction associated with Chiari I malformation and cervical cord compression. The pattern of sensory abnormality is consistent with medullary dysfunction among the patients with Chiari I malformation and cervical cord dysfunction among cord compression patients. There were differences in the types and extent of pain and the associated disorders of mood observed among the cohorts defined above. These differences may be partly due to the presence and location of central sensory dysfunction.

Neural pathways underlying vocal control

Vocalization is a complex behaviour pattern, consisting of essentially three components: laryngeal activity, respiratory movements and supralaryngeal (articulatory) activity. The motoneurones controlling this behaviour are located in various nuclei in the pons (trigeminal motor nucleus), medulla (facial nucleus, nucl. ambiguus, hypoglossal nucleus) and ventral horn of the spinal cord (cervical, thoracic and lumbar region). Coordination of the different motoneurone pools is carried out by an extensive network comprising the ventrolateral parabrachial area, lateral pontine reticular formation, anterolateral and caudal medullary reticular formation, and the nucl. retroambiguus. This network has a direct access to the phonatory motoneurone pools and receives proprioceptive input from laryngeal, pulmonary and oral mechanoreceptors via the solitary tract nucleus and principal as well as spinal trigeminal nuclei. The motor-coordinating network needs a facilitatory input from the periaqueductal grey of the midbrain and laterally bordering tegmentum in order to be able to produce vocalizations. Voluntary control of vocalization, in contrast to completely innate vocal reactions, such as pain shrieking, needs the intactness of the forebrain. Voluntary control over the initiation and suppression of vocal utterances is carried out by the mediofrontal cortex (including anterior cingulate gyrus and supplementary as well as pre-supplementary motor area). Voluntary control over the acoustic structure of vocalizations is carried out by the motor cortex via pyramidal/corticobulbar as well as extrapyramidal pathways. The most important extrapyramidal pathway seems to be the connection motor cortex-putamen-substantia nigra-parvocellular reticular formation-phonatory motoneurones. The motor cortex depends upon a number of inputs for fulfilling its task. It needs a cerebellar input via the ventrolateral thalamus for allowing a smooth transition between consecutive vocal elements. It needs a proprioceptive input from the phonatory organs via nucl. ventralis posterior medialis thalami, somatosensory cortex and inferior parietal cortex. It needs an input from the ventral premotor and prefrontal cortex, including Broca's area, for motor planning of longer purposeful utterances. And it needs an input from the supplementary and pre-supplementary motor area which give rise to the motor commands executed by the motor cortex.

Parallel circuits mediating distinct emotional coping reactions to different types of stress

All animals, including humans, react with distinct emotional coping strategies to different types of stress. Active coping strategies (e.g. confrontation, fight, escape) are evoked if the stressor is controllable or escapable. Passive coping strategies (e.g. quiescence, immobility, decreased responsiveness to the environment) are usually elicited if the stressor is inescapable and help to facilitate recovery and healing. Neural substrates mediating active versus passive emotional coping have been identified within distinct, longitudinal neuronal columns of the midbrain periaqueductal gray (PAG) region. Active coping is evoked by activation of either the dorsolateral or lateral columns of the PAG; whereas passive coping is triggered by activation of the ventrolateral PAG. Recent anatomical studies indicate that each PAG column receives a distinctive set of ascending (spinal and medullary) and descending (prefrontal cortical and hypothalamic) afferents. Consistent with the anatomy, functional studies using immediate early gene expression (c-fos) as a marker of neuronal activation have revealed that the preferential activation of a specific PAG column reflects (i) the type of emotional coping reaction triggered, and (ii) whether a physical or psychological stressor was used.

Modulation of defensive behavior by periaqueductal gray NMDA/glycine-B receptor

Glutamate (GLU) associated with glycine, act as co-transmitter at the N-methyl-D-aspartate/glycine-B (NMDA/GLY(B)) receptor. Dorsal periaqueductal gray (dPAG) neurons express NMDA/GLY(B) receptors suggesting a GLU physiological role in mediating the responses elicited by stimulation of this area. Immunohistochemical data provided evidence of a possible correlation among elevated plus-maze (EPM), fear-like defensive behavior, and dPAG activity. The present data show that whereas the NMDA/GLY(B) receptor agonists increased the open-arm avoidance responses in the EPM, the antagonists had the opposite effects. Microinjection of NMDA/GLY(B) receptor agonists within the dPAG during test sessions in the EPM resulted in an enduring learned fear response detected in the retest. Therefore, in addition to the proposed role for the dPAG in panic attacks (escape), these findings suggest that the dPAG can also participate in more subtle anxiety-like behaviors.

The projections of the midbrain periaqueductal grey to the pons and medulla oblongata in rats

It is now established that stimulation of the ventrolateral midbrain periaqueductal grey (PAG) evokes inhibition of nociceptive spinal neurons, which results in analgesia and a powerful attenuation of pain behaviour. It is postulated that the PAG exerts this inhibitory effect on spinal nociceptive functions through the activation of descending serotonergic and noradrenergic pathways that arise from the rostral ventromedial medulla (RVM) and pontine noradrenergic nuclei. To investigate the neuroanatomical substrate of this functional link between the PAG and RVM, as well as the pontine noradrenergic nuclei in the rat, we labelled axons that project from the ventrolateral PAG to various regions of the pons and medulla oblongata using the anterograde tracing substance, Phaseolus vulgaris leucoagglutinin. We demonstrated that some of PAG efferents really do terminate in the RVM and pontine noradrenergic nuclei, but a substantial proportion of them project to the intermediate subdivision of the pontobulbar reticular formation. Combining the axonal tracing with serotonin- and tyrosine-hydroxylase-immunohistochemistry, we also found that, in contrast to previous results, PAG efferents make relatively few appositions with serotonin- and tyrosine-hydroxylase-immunoreactive neurons in the RVM and pontine noradrenergic nuclei; most of them terminate in nonimmunoreactive territories. The results suggest that the ventrolateral PAG may activate a complex pontobulbar neuronal assembly including neurons in the intermediate subdivision of the pontobulbar reticular formation, serotonin- and tyrosine-hydroxylase-immunoreactive and nonimmunoreactive neurons in the RVM and pontine noradrenergic nuclei. This pontobulbar neural circuitry, then, may mediate the PAG-evoked activities towards the spinal dorsal horn resulting in the inhibition of spinal nociceptive functions.

Evidence that hemorrhagic hypotension is mediated by the ventrolateral periaqueductal gray region

Severe hemorrhage lowers arterial pressure by suppressing sympathetic activity. This study tested the hypothesis that the decompensatory phase of hemorrhage is mediated by the ventrolateral periaqueductal gray (vlPAG), a region importantly involved in the autonomic and behavioral responses to stress and trauma. Neuronal activity in the vlPAG was inhibited with either lidocaine or cobalt chloride 5 min before hemorrhage (2.5 ml/100 g body wt) was initiated in conscious, unrestrained rats. Bilateral injection of lidocaine (0.5 microl of a 2% or 1 microl of a 5% solution) into the caudal vlPAG delayed the onset and reduced the magnitude of the hypotension produced by hemorrhage significantly. In contrast, inactivation of the dorsolateral PAG with lidocaine was ineffective. Cobalt chloride (5 mM; 0.5 microl), which inhibits synaptic transmission but not axonal conductance, also attenuated hemorrhagic hypotension significantly. Microinjection of lidocaine or cobalt chloride into the vlPAG of normotensive, nonhemorrhaged rats did not influence cardiovascular function. These data indicate that the vlPAG plays an important role in the response to hemorrhage.

A neural circuit for circadian regulation of arousal

An unknown aspect of behavioral state regulation is how the circadian oscillator of the suprachiasmatic nucleus (SCN) regulates sleep and waking. In this report, we describe the necessary elements for a circuit that provides circadian regulation of arousal. Trans-synaptic retrograde tracing revealed a prominent indirect projection from the SCN to the noradrenergic nucleus locus coeruleus (LC), a brain arousal system. Double-labeling experiments revealed several possible links between the SCN and the LC, including the dorsomedial (DMH) and paraventricular hypothalamic nuclei (PVN), as well as medial and ventrolateral pre-optic areas. Lesion studies confirmed that the DMH is a substantial relay in this circuit. Next, neurophysiology experiments revealed circadian variations in LC impulse activity. Lesions of the DMH eliminated these circadian changes in LC activity, confirming the functionality of the SCN-DMH-LC circuit. These results reveal mechanisms for regulation of circadian and sleep-waking functions.

Micturition evoked by glutamate microinjection in the ventrolateral periaqueductal gray is mediated through Barrington's nucleus in the rat

Neural tracing experiments have demonstrated a direct spinal projection to Barrington's nucleus and a possible indirect pathway to Barrington's nucleus via the periaqueductal gray. We sought to identify the role of the periaqueductal gray matter in micturition in urethane-anesthetized rats. Blockade of micturition by focal injection of cobalt chloride was used to identify sites critical to micturition. These sites were located near the ventral margin of the caudal ventrolateral periaqueductal gray and in Barrington's nucleus. L-Glutamate injections into caudal regions of the periaqueductal gray evoked bladder contraction with coordinated sphincter activation. Additional L-glutamate sites with a similar pattern of response and sites where sphincter activation was produced without bladder contraction were found more rostrally and dorsally in the periaqueductal gray. Activation of bladder contractions by L-glutamate injection in the ventrolateral periaqueductal gray was blocked by prior injection of cobalt chloride into Barrington's nucleus. From these data we propose that ventrolateral periaqueductal gray is functionally important to micturition in the urethane-anesthetized rat. Further, we have shown that a periaqueductal gray to Barrington's nucleus pathway is functionally relevant to central mediation of bladder contraction.

Postnatal changes in Fos-like immunoreactivity evoked by hypoxia in the rat brainstem and hypothalamus

We have recently used Fos expression in adult rats to map neuronal populations activated in the brainstem and hypothalamus during the acute ventilatory response to moderate hypoxia (O(2) 11%). Although present at birth, this response evolves postnatally. The present investigation aimed at a better understanding of these maturational processes by delineating structures that might functionally develop after birth. The developmental pattern Fos expression evoked by hypoxia was analysed in rats aged from 0 to 26 postnatal days. The numbers of Fos positive neurons markedly increased with the age in the medullary areas related to respiratory control during the 2 first postnatal weeks. Thereafter, the response plateaued in the nucleus tractus solitarius and attenuated in the ventral medulla. In the upper brainstem (parabrachial area, central grey) and the hypothalamus (posterior and dorsomedial nuclei, ventral zone), Fos response to hypoxia was absent or weak at birth and increased until late development. The significance of the development of evoked Fos expression in these rostral sites is discussed together with their possible contribution to the maturation of O(2)-sensitive chemoreflex pathways.

Muscle pain activates a direct projection from ventrolateral periaqueductal gray to rostral ventrolateral medulla in rats

Activation of the ventrolateral periaqueductal gray (vlPAG) evokes a reaction of quiescence, immobility, hypotension and bradycardia. Pain of deep somatic or visceral origin also often triggers a reaction of quiescence, immobility, hypotension and bradycardia and further, evokes a selective increase in immediate-early-gene (c-Fos) expression within the vlPAG. Vasodepression evoked from the vlPAG is thought to be mediated by an inhibition of presympathetic neurons within the rostral ventrolateral medulla (RVLM). In this study the prior injection of retrograde tracer into the RVLM was combined with the use of Fos expression as a marker of neuronal activation, to determine if deep (muscle) pain-evoked vasodepression could be mediated by a direct vlPAG-RVLM pathway. It was revealed that intramuscular injection of formalin, in the anaesthetised rat, evoked a significant increase in Fos expression within the caudal vlPAG, and that approximately 25% of the Fos-immunoreactive neurons projected to the RVLM.

Central circuits mediating patterned autonomic activity during active vs. passive emotional coping

Animals, including humans, react with distinct emotional coping strategies to different sets of environmental demands. These strategies include the capacity to affect appropriate responses to "escapable" or "inescapable" stressors. Active emotional coping strategies--fight or flight--are particularly adaptive if the stress is escapable. On the other hand, passive emotional coping strategies-quiescence, immobility, decreased responsiveness to the environment-are useful when the stress is inescapable. Passive strategies contribute also to facilitating recovery and healing once the stressful event is over. Active vs. passive emotional coping strategies are characterised further by distinct patterns of autonomic change. Active strategies are associated with sympathoexcitation (hypertension, tachycardia), whereas passive strategies are associated with sympathoinhibitory patterns (hypotension, bradycardia). Distinct neural substrates mediating active vs. passive emotional coping have been identified within the longitudinal neuronal columns of the midbrain periaqueductal gray region (PAG). The PAG offers then a potentially useful point of entry for delineating neural circuits mediating the different forms of emotional coping and their associated patterns of autonomic activity. As one example, recent studies of the connections of orbital and medial prefrontal cortical (PFC) fields with specific PAG longitudinal neuronal columns are reviewed. Findings of discrete orbital and medial PFC projections to different PAG columns, and related PFC and PAG columnar connections with specific subregions of the hypothalamus, suggest that distinct but parallel circuits mediate the behavioural strategies and patterns of autonomic activity characteristic of emotional "engagement with" or "disengagement from" the external environment.

Monosynaptic projections from the lateral periaqueductal gray to the nucleus retroambiguus in the rhesus monkey: implications for vocalization and reproductive behavior

The periaqueductal gray (PAG) is known to be essential for vocalization and reproductive behavior. The PAG controls components of these behaviors by means of projections to the nucleus retroambiguus (NRA), a group of premotor neurons in the caudal medulla oblongata. In the accompanying study (VanderHorst et al., 2000 [accompanying study]), the NRA and its lumbosacral projections have been identified in the rhesus monkey. The present light and electron microscopical tracing study describes the PAG-NRA pathway in primates. To locate midbrain neurons projecting to the NRA, wheat germ agglutinin horseradish peroxidase (WGA-HRP) was injected into the NRA in six monkeys. To determine the distribution pattern of PAG axons in the medulla oblongata, WGA-HRP was injected into the PAG and adjacent tegmentum in three additional monkeys. In one of these three monkeys, biotinylated dextran amine and cholera toxin subunit b were injected into the lumbosacral cord to retrogradely identify NRA neurons. The results show that a compact group of neurons in the medial part of the lateral PAG at the intercollicular level sends a dense projection to the NRA. The projection is bilateral with a clear ipsilateral predominance. At the ultrastructural level, there are monosynaptic contacts between PAG fibers and NRA neurons, including NRA neurons that project to the lumbosacral cord. The synaptic contacts were primarily asymmetrical and the labeled terminal profiles contained spherical and dense core vesicles. It is concluded that there exists a strong and direct PAG-NRA pathway in the rhesus monkey. Because NRA neurons projecting to the lower lumbar cord are included, the PAG-NRA projection is likely to be involved not only in vocalization but also in other behaviors, such as receptive posture.

Orbitomedial prefrontal cortical projections to distinct longitudinal columns of the periaqueductal gray in the rat

We utilised retrograde and anterograde tracing procedures to study the origin and termination of prefrontal cortical (PFC) projections to the periaqueductal gray (PAG) in the rat. A previous study, in the primate, had demonstrated that distinct subgroups of PFC areas project to specific PAG columns. Retrograde tracing experiments revealed that projections to dorsolateral (dlPAG) and ventrolateral (vlPAG) periaqueductal gray columns arose from medial PFC, specifically prelimbic, infralimbic, and anterior cingulate cortices. Injections made in the vlPAG also labeled cells in medial, ventral, and dorsolateral orbital cortex and dorsal and posterior agranular insular cortex. Other orbital and insular regions, including lateral and ventrolateral orbital, ventral agranular insular, and dysgranular and granular insular cortex did not give rise to appreciable projections to the PAG. Anterograde tracing experiments revealed that the projections to different PAG columns arose from specific PFC areas. Projections from the caudodorsal medial PFC (caudal prelimbic and anterior cingulate cortices) terminated predominantly in dlPAG, whereas projections from the rostroventral medial PFC (rostral prelimbic cortex) innervated predominantly the vlPAG. As well, consistent with the retrograde data, projections arising from select orbital and agranular insular cortical areas terminated selectively in the vlPAG. The results indicate: (1) that rat orbital and medial PFC possesses an organisation broadly similar to that of the primate; and (2) that subdivisions within the rat orbital and medial PFC can be recognised on the basis of projections to distinct PAG columns.

A projection from the ventral tegmental area to the periaqueductal gray involved in cardiovascular regulation

Experiments were done in alpha-chloralose-anesthetized rats to determine a pathway mediating the cardiovascular depressor responses elicited from stimulation of the ventral tegmental area (VTA). The magnitude of the depressor responses elicited by glutamate stimulation (0.1 M/30 nl) of the VTA was examined after neuronal block produced by microinjections of lidocaine into ascending fiber bundles leaving the VTA to innervate the forebrain and thalamus. Bilateral microinjections of 1 microl of 4% lidocaine in the medial forebrain bundle (n = 6) and in the periventricular fibers of the midbrain (n = 5) did not attenuate the depressor response from stimulation of the VTA. Experiments were done using the anterograde tracer biotinylated dextran amine to identify descending projections from the VTA to cardiovascular centers in the brain stem. Examination of the nucleus of the solitary tract, ventrolateral medulla, and A5 catecholaminergic cell group revealed few or no fibers or terminals. Occasional fibers and some terminals were observed in the nucleus of raphe magnus, parabrachial nucleus, and locus ceruleus. A very dense bilateral projection was found to the ventrolateral periaqueductal gray (PAGvl) and dorsal raphe nucleus adjacent to the PAGvl. Bilateral injections of 4% lidocaine (n = 4) or 10 mM cobalt chloride (n = 5) into the PAGvl region attenuated the depressor responses elicited by stimulation of the VTA by approximately 50%. These experiments indicate that the depressor responses elicited from activation of the VTA are mediated in part by a pathway to a cardiovascular depressor area located in the PAGvl.

Brainstem and hypothalamic areas involved in respiratory chemoreflexes: a Fos study in adult rats

The adaptation to hypoxia and hypercapnia requires the activation of several anatomical structures along the neuraxis. In this study, using Fos immunoreactivity, we sought to map neuronal populations involved in chemoreflex networks activated during the responses to moderate hypoxia (O(2) 11%), and hypercapnia (CO(2) 5%) in the brainstem and the hypothalamus of the rat. In the medulla, hypoxia elicited marked and significant staining in the nucleus of the solitary tract (NTS), and in parapyramidal neurons located near the ventral surface, whereas hypercapnia evoked significantly c-fos only near the ventral surface in paraolivar neurons. In contrast, within pontine and suprapontine structures, both hypoxia and hypercapnia evoked similarly Fos immunoreactivity in the lateral parabrachialis area, the central grey, the caudal hypothalamus (dorsomedial and posterior hypothalamic nuclei), and in a ventro-lateral hypothalamic area, extending from the rostral limit of the mammillary nuclei to the retrochiasmatic area. More rostrally, labelling was observed in the paraventricular nucleus of the hypothalamus in response to hypercapnia, and in the supraoptic nucleus in response to hypoxia. These results support the hypothesis that chemoreflexes pathways are not only restricted to medulla and pons but also involved mesencephalic and hypothalamic regions. The parabrachialis area and the central grey may be key relays between caudal and ventral hypothalamic neurons, and medullary neurons involved in the response to hypoxia and hypercapnia.

Noradrenaline in the ventral forebrain is critical for opiate withdrawal-induced aversion

Cessation of drug use in chronic opiate abusers produces a severe withdrawal syndrome that is highly aversive, and avoidance of withdrawal or associated stimuli is a major factor contributing to opiate abuse. Increased noradrenaline in the brain has long been implicated in opiate withdrawal, but it has not been clear which noradrenergic systems are involved. Here we show that microinjection of beta-noradrenergic-receptor antagonists, or of an alpha2-receptor agonist, into the bed nucleus of the stria terminalis (BNST) in rats markedly attenuates opiate-withdrawal-induced conditioned place aversion. Immunohistochemical studies revealed that numerous BNST-projecting cells in the A1 and A2 noradrenergic cell groups of the caudal medulla were activated during withdrawal. Lesion of these ascending medullary projections also greatly reduced opiate-withdrawal-induced place aversion, whereas lesion of locus coeruleus noradrenergic projections had no effect on opiate-withdrawal behaviour. We conclude that noradrenergic inputs to the BNST from the caudal medulla are critically involved in the aversiveness of opiate withdrawal.

Cardiovascular changes following acute and chronic chemical lesions of the dorsal periaqueductal gray in conscious rats

This study was carried out to investigate the effects of chemical lesions of dorsal periaqueductal gray (DPAG) on resting arterial pressure (AP) and heart rate (HR) as well as on cardiac baroreflex of conscious normotensive rats. Lesions were performed by bilateral microinjections of 150 mM NMDA into the DPAG (DPAG-lesion group). Controls were similarly injected with 165 mM NaCl (DPAG-sham group). Animals with chronic lesions confined only to the superior colliculus (SC-lesion group) were also used as controls of DPAG-lesion. Cardiovascular parameters were recorded 1 or 7 days after the microinjections of NMDA in acute and chronic groups, respectively. Cardiac baroreflex was assessed by measuring the HR responses to the intravenous injection of phenylephrine or sodium nitroprusside. Baroreflex was estimated by sigmoidal curve fitting of HR responses. An increased baroreflex gain was observed in chronic DPAG-lesion rats compared to both DPAG-sham (p < 0.01) and SC-lesion (p < 0.05) chronic groups. The chronic DPAG-lesion group showed also an elevation of both the tachycardia (p < 0.05) and bradycardia (p < 0.01) plateaus compared to chronic DPAG-sham rats, while the SC-lesion group showed an elevation of the bradycardia plateau only (p < 0.01). Similar results on baroreflex function were observed following acute lesion of the DPAG, i.e. an increase in baroreflex gain (p < 0.01) and the elevation of both tachycardia (p < 0.05) and bradycardia plateaus (p < 0.01) compared to the acute DPAG-sham group. Resting AP and HR did not differ among the chronic groups. In contrast, the acute lesion of the DPAG produced a reduction in AP (p < 0.01) accompanied by an increase in HR (p < 0.01). The present data suggest that the DPAG is involved in the tonic and reflex control of AP and HR in conscious rats. In addition, the SC seems to contribute to the baroreflex cardioinhibition.

Direct parabrachial nuclear projections to the pharyngeal motoneurons in the rat: an anterograde and retrograde double-labeling study

The parabrachial nucleus consists of several subnuclei which contains autonomic, gustatory, visceral sensory, nociceptive, and respiratory related neurons. We have investigated the direct projections from the rat parabrachial region, including the K olliker-Fuse nucleus, to the pharyngeal motoneurons with an anterograde and retrograde double-tracing technique. The cholera toxin subunit-B was injected into the lower pharynx or the esophagus after injection of biotinylated dextran amine into the ventrolateral parabrachial nuclear region, including the external medial, the external lateral, and the crescent area of the central lateral parabrachial nuclei and into the Kölliker-Fuse nucleus. The anterogradely dextran amine-labeled fibers from these nuclei projected to the semicompact, loose and external formations besides the compact formation of the nucleus ambiguus. Many anterogradely labeled fibers and terminals were found to contact retrogradely cholera toxin-labeled pharyngeal neuronal soma and dendrites in the semicompact formation of the nucleus ambiguus. The medial half of the parabrachial nucleus, including the medial and the medial part of the central lateral parabrachial nuclei, sent a few fibers to the reticular formation just dorsal to the esophageal motoneurons but no fibers to either the pharyngeal or to the esophageal motoneurons. These results suggested that the visceral sensory, gustatory, nociceptive or respiratory related neurons in the parabrachial nucleus project directly to the pharyngeal motoneurons, but there are no parabrachial projections to the esophageal motoneurons in the nucleus ambiguus.

Medullary catecholaminergic projections to the ventrolateral periaqueductal gray region activated by halothane anaesthesia

Under anaesthesia, blood loss and deep pain can evoke a premature, centrally-mediated sympathoinhibition leading to decompensated shock and sometimes even death. The central circuits evoking premature vasodepressor syncope are unknown, although medullary catecholaminergic pathways have been implicated. The ventrolateral periaqueductal gray region is one of only three brain regions in which catecholamine content is increased during halothane anaesthesia. The ventrolateral periaqueductal gray also contains neurons which are selectively activated by blood loss and deep pain, and recent work from our laboratory has suggested that it is a pivotal structure in central sympathoinhibitory circuits. Using retrograde tracing techniques combined with the immunohistochemical detection of: (i) the catecholamine synthetic enzyme, tyrosine hydroxylase and (ii) the protein product of the immediate-early gene c-fos as a marker of neuronal activation; the results of this study indicate that catecholaminergic projections from the A1, C1 and C2 regions of the medulla to the ventrolateral periaqueductal gray are activated by halothane anaesthesia. These data are consistent with the hypotheses that ascending catecholaminergic projections to the ventrolateral periaqueductal gray: (i) are a component of the central neural circuitry responsible for the sympathoinhibitory effects of halothane anaesthesia, and (ii) may contribute to the premature elicitation of vasodepressor syncope following blood loss and deep pain under conditions of anaesthesia.

Central nervous system abnormalities in complex regional pain syndrome (CRPS): clinical and quantitative evidence of medullary dysfunction

OBJECTIVE

Sensory and motor abnormalities are common among patients with complex regional pain syndrome (CRPS). The purpose of the present study was to define and characterize these abnormalities and to develop a hypothesis regarding the area of the central nervous system from which they derive.

DESIGN

Data were acquired from study subjects using clinical examination and quantitative assessment of neurological function. Subjects were divided into four groups. CRPS patients were differentiated into two groups based on the presence or absence of sensory deficit on the face to clinical examination. The other two groups were composed of patients with other chronic pain syndromes and normal individuals without chronic pain or disability. Clinical and quantitative data were compared between groups.

PATIENTS

One hundred forty-five CRPS patients, 69 patients with other pain conditions, and 26 normal individuals were studied.

RESULTS

A high incidence of trigeminal hypoesthesia was observed in CRPS patients. CRPS patients with trigeminal hypoesthesia manifested bilateral deficits of sensory function, with a predominant hemilateral pattern. These patients also manifested bilateral motor weakness with a more prominent hemiparetic pattern. Both sensory and motor deficits were greatest ipsilateral to the painful side of the body. These features differed significantly from those of CRPS patients lacking clinical trigeminal deficit, other pain patients, and normals. A lower cranial nerve abnormality (sternocleidomastoid weakness) and a myelopathic feature (Hoffman's reflex) were more common in CRPS patients with trigeminal hypoesthesia.

CONCLUSIONS

Nearly half of CRPS patients had abnormalities of spinothalamic, trigeminothalamic, and corticospinal function that may represent dysfunction of the medulla. One-third of the remaining CRPS patients had neuroimaging evidence of spinal cord or brain pathology. The majority of CRPS patients in this study have measurable abnormalities of the sensory and motor systems or neuroimaging evidence of spinal cord or brain dysfunction.

The ventrolateral periaqueductal gray projects to caudal brainstem depressor regions: a functional-anatomical and physiological study

The reaction of shock, a precipitous, life-threatening fall in arterial pressure and heart rate, is evoked often by the combination of deep pain and blood loss following traumatic injury. A similar "shock-like" pattern of response can be evoked by excitation of the ventrolateral midbrain periaqueductal gray. Further, ventrolateral periaqueductal gray neurons are selectively activated by deep somatic or visceral pain and haemorrhage. The pathways mediating ventrolateral periaqueductal gray evoked hypotension and bradycardia are not known. In this study, the projections from the ventrolateral periaqueductal gray to "cardiovascular" regions in the caudal medulla of the rat were examined. Injections of the anterograde tracer, biotinylated dextran amine at physiologically-defined, ventrolateral periaqueductal gray depressor sites, revealed strong projections to the caudal midline medulla and to the depressor region of the caudal ventrolateral medulla. Injections of excitatory amino acids established that substantial falls in arterial pressure could be evoked from the ventrolateral periaqueductal gray-recipient parts of the caudal midline medulla. Injections of the retrograde tracer, cholera toxin subunit B at physiologically-defined, depressor sites in the caudal midline medulla and the caudal ventrolateral medulla confirmed the existence of substantial projections from the ventrolateral periaqueductal gray. Although previous studies have emphasized the importance of projections from the ventrolateral periaqueductal gray to the pressor region of the rostral ventrolateral medulla, this study has revealed the existence of strong ventrolateral periaqueductal gray projections to depressor regions within the caudal medulla (caudal midline medulla and caudal ventrolateral medulla) which likely contribute to ventrolateral periaqueductal gray-mediated hypotension and bradycardia.

Axonal collateral-collateral transport of tract tracers in brain neurons: false anterograde labelling and useful tool

It is well established that some neuroanatomical tracers may be taken up by local axonal terminals and transported to distant axonal collaterals (e.g., transganglionic transport in dorsal root ganglion cells). However, such collateral-collateral transport of tracers has not been systematically examined in the central nervous system. We addressed this issue with four neuronal tracers--biocytin, biotinylated dextran amine, cholera toxin B subunit, and Phaseolus vulgaris-leucoagglutinin--in the cerebellar cortex. Labelling of distant axonal collaterals in the cerebellar cortex (indication of collateral-collateral transport) was seen after focal iontophoretic microinjections of each of the four tracers. However, collateral-collateral transport properties differed among these tracers. Injection of biocytin or Phaseolus vulgaris-leucoagglutinin in the cerebellar cortex yielded distant collateral labelling only in parallel fibres. In contrast, injection of biotinylated dextran amine or cholera toxin B subunit produced distant collateral labelling of climbing fibres and mossy fibres, as well as parallel fibres. The present study is the first systematic examination of collateral-collateral transport following injection of anterograde tracers in brain. Such collateral-collateral transport may produce false-positive conclusions regarding neural connections when using these tracers for anterograde transport. However, this property may also be used as a tool to determine areas that are innervated by common distant afferents. In addition, these results may indicate a novel mode of chemical communication in the nervous system.

Discrete subregions of the rat midbrain periaqueductal gray project to nucleus ambiguus and the periambigual region

We investigated the organization of projections from the rat midbrain periaqueductal gray to nucleus ambiguus and the periambigual region using retrograde and anterograde tract tracing techniques. Retrograde tracing results revealed that neurons that project to nucleus ambiguus arise from three discrete, longitudinally organized columns of neurons located in the supraoculomotor central gray, lateral and ventrolateral periaqueductal gray. Anterograde tracing studies demonstrated that projections from these three columns of periaqueductal gray neurons terminate with topographic specificity in nucleus ambiguus and the periambigual region. Double-labelling studies demonstrated that periaqueductal gray neurons terminate in close contiguity to cholinergic neurons in the compact, semicompact, loose and external formations of nucleus ambiguus. The present results suggest that projections from periaqueductal gray to nucleus ambiguus may mediate, in part, certain cardiovascular adjustments and vocalizations produced by stimulation of periaqueductal gray.

Cardiovascular effects of microinjections of opioid agonists into the 'Depressor Region' of the ventrolateral periaqueductal gray region

Microinjections of excitatory amino acids made into the ventrolateral midbrain periaqueductal gray of the rat have revealed that neurons in this region integrate a reaction characterised by quiescence, hyporeactivity, hypotension and bradycardia. Microinjections of both excitatory amino acids and opioids into the ventrolateral periaqueductal gray have shown also that it is a key central site mediating analgesia. The effects of injections of opioids into the ventrolateral periaqueductal gray on arterial pressure and heart rate or behaviour are unknown. In this study we first mapped in the rat the extent of the ventrolateral periaqueductal gray hypotensive region as revealed by microinjections of excitatory amino acids. We found that ventrolateral periaqueductal gray depressor region extended more rostrally than previously thought into the tegmentum ventrolateral to the periaqueductal gray. Subsequently we studied for the first time, the effects of microinjections of mu-, delta-, and kappa-opioid agonists made into the ventrolateral periaqueductal grey depressor region. In contrast to the effects of excitatory amino acid injections, microinjections of the mu-opioid agonist ([D-Ala2,N-Me-Phe4,Gly-ol5]enkephalin) evoked hypertension and tachycardia at approximately 50% of sites. Similar to excitatory amino acid injections, microinjections of both the delta-opioid agonist ([D-Pen2,D-Pen5]enkephalin), and the kappa-opioid agonist ((5,7,8)-(+)-N-Methyl-N-[7-(1-pyrrolidinyl)-1-oxaspiro[4.5]dec-8-y l]-benzeneacetamide) evoked either a hypotension and bradycardia, or had no effect. These results indicate that different opiate receptor subtypes are present on a distinct population of ventrolateral periaqueductal gray neurons, or at different ventrolateral periaqueductal gray synaptic locations (pre- or post-synaptic).

Direct synaptic projections to esophageal motoneurons in the nucleus ambiguus from the nucleus of the solitary tract of the rat

Neurons of the nucleus of the solitary tract (NTS) serve as interneurons in swallowing. We investigated the synaptology of the terminals of these neurons and whether they project directly to the esophageal motoneurons in the compact formation of the nucleus ambiguus (AmC). Following wheat germ agglutinin conjugated horseradish peroxidase (WGA-HRP) injection into the NTS, many anterogradely labeled axodendritic terminals were found in the neuropil of the AmC. The majority of labeled axodendritic terminals (89%) contained round vesicles and made asymmetric synaptic contacts (Gray's type I), but a few (11%) contained pleomorphic vesicles and made symmetric synaptic contacts (Gray's type II). More than half of the labeled terminals contacted intermediate dendrites (1-2 microm diameter). There were no retrogradely labeled medium-sized motoneurons, but there were many retrogradely labeled small neurons having anterogradely labeled axosomatic terminals. A combined retrograde and anterograde transport technique was developed to verify the direct projection from the NTS to the esophageal motoneurons. After the esophageal motoneurons were retrogradely labeled by cholera toxin subunit B conjugated HRP, the injection of WGA-HRP into the NTS permitted ultrastructural recognition of anterogradely labeled axosomatic terminals contacting directly labeled esophageal motoneurons. Serial sections showed that less than 20% of the axosomatic terminals were labeled in the esophageal motoneurons. They were mostly Gray's type I, but a few were Gray's type II. In the small neurons, more than 30% of axosomatic terminals were labeled, which were exclusively Gray's type I. These results indicate that NTS neurons project directly not only to the esophageal motoneurons, but also to the small neurons which have bidirectional connections with the NTS.