Search

The respiratory rhythm is generated by the interaction of oscillators disparately distributed throughout the pons, medulla, and spinal cord. According to the classic model, the interaction amongst preBötzinger complex (preBötzC) spontaneously bursting preinspiratory units and Bötzinger complex (BötzC) expiratory cells generates the principal respiratory rhythm, thence relayed caudally to the pattern generating elements and premotoneurons of the rostral and caudal divisions of the ventral respiratory group and bulbospinal units of the dorsal respiratory group. Rhythm and pattern generating elements in the ventrolateral medulla receive powerful phasic and tonic modulatory inputs from diencephalic structures, midbrain, Kölliker-Fuse, and parabrachial nuclei, retrotrapezoid nucleus, parafacial respiratory group, ventrolateral metencephalon, nucleus tractus solitarius, and brainstem reticular formation, collectively shaping the normal eupneic discharge. Empirical and computational studies have generated models of respiratory rhythmogenesis and pattern formation variously predicated upon pacemaker, network, or hybrid pacemaker network mechanisms to explain oscillatory behavior and regularity. Network mechanisms critically require the integrity and functionality of inhibitory synaptic neurotransmission. The operation and contribution of inhibitory elements in respiratory rhythm generation and pattern formation are well demonstrated empirically and incorporated in computational network and hybrid models of breathing. Fast inhibitory synaptic neurotransmission utilizes GABAAergic and glycinergic mediated activation of receptor linked chloride conductances, generating an inwardly directed flux of chloride ions mediating membrane voltage hyperpolarization and is required to generate eupneic respiratory patterns in vivo and situ. Persistence of rhythmicity in the presence of synaptic antagonism of GABAA and glycine receptor mediated fast inhibitory neurotransmission indicates pacemaker generating mechanisms sufficiently capable of independently generating this behavior in vivo and transected intact preparations maintaining the preBötzC as the most rostrally preserved structure. The role of GABAB receptor mediated neuromodulation in respiratory rhythm generation and pattern formation is comparatively significantly less investigated. GABABergic activation of postsynaptic and presynaptic membrane receptors variably upregulates potassium conductances and downregulates calcium conductances. Respiratory rhythm and pattern are powerfully modulated in vivo, in situ, and in vitro by superfusion or localized microinjections of GABABergic agonists and antagonists, though are typically not abolished by these experimental interventions. Directionality and magnitude of these effects exhibit maturational changes. The relative depolarization of chloride reversal potentials during the early neonatal period, with gradual shifts towards normal hyperpolarizing values during development, suggests GABABergic signaling may mediate the inhibitory neurotransmission necessary to generate triphasic eupnea. We review and discuss the role of spontaneously bursting oscillators and network mechanisms predicating upon fast inhibitory synaptic neurotransmission in contributing to respiratory rhythmogenesis and pattern formation.

Cardiorespiratory collapse following a seizure is a suspected cause of sudden unexpected death in epilepsy (SUDEP), the leading cause of epilepsy-related mortality. In the commonly used Kcna1 gene knockout (Kcna1−/−) mouse model of SUDEP, cardiorespiratory profiling reveals an array of aberrant breathing patterns that could contribute to risk of seizure-related mortality. However, the brain structures mediating these respiratory abnormalities remain unknown. We hypothesize that Kv1.1 deficiency in respiratory control centers of the brain contribute to respiratory dysfunction in Kcna1−/− mice leading to increased SUDEP risk. Thus, in this study, we first used immunohistochemistry to map expression of Kv1.1 protein in cardiorespiratory brain regions of wild-type Kcna1+/+ (WT) mice. Next, GFAP and Iba1 immunostaining was used to test for the presence of astrogliosis and microgliosis, respectively, in the cardiorespiratory centers of Kcna1−/− mice, which could be indicative of seizure-related brain injury that could impair breathing. In WT type mice, we detected Kv1.1 protein in all cardiorespiratory centers examined, including the basolateral amygdala, dorsal respiratory group, dorsal motor nucleus of vagus, nucleus ambiguus, ventral respiratory column, and pontine respiratory group, as well as chemosensory centers including the retrotrapezoid and median raphae nuclei. Extensive gliosis was observed in the same areas in Kcna1−/− mice suggesting that seizure-associated brain injury could contribute to respiratory abnormalities.

The respiratory pattern generator of mammals is anatomically organized in lateral respiratory columns (LRCs) within the brainstem. LRC compartments serve specific functions in respiratory pattern and rhythm generation. While the caudal medullary reticular formation (cMRF) has respiratory functions reportedly related to the mediation of expulsive respiratory reflexes, it remains unclear whether neurons of the cMRF functionally belong to the LRC. In the present study we specifically investigated the respiratory functions of the cMRF. Tract tracing shows that the cMRF has substantial connectivity with key compartments of the LRC, particularly the parafacial respiratory group and the Kölliker-Fuse nuclei. These neurons have a loose topography and are located in the ventral and dorsal cMRF. Systematic mapping of the cMRF with glutamate stimulation revealed potent respiratory modulation of the respiratory motor pattern from both dorsal and ventral injection sites. Pharmacological inhibition of the cMRF with the GABA-receptor agonist isoguvacine produced significant and robust changes to the baseline respiratory motor pattern (decreased laryngeal post-inspiratory and abdominal expiratory motor activity, delayed inspiratory off-switch and increased respiratory frequency) after dorsal cMRF injection, while ventral injections had no effect. The present data indicate that the ventral cMRF is not an integral part of the respiratory pattern generator and merely serves as a relay for sensory and/or higher command-related modulation of respiration. On the contrary, the dorsal aspect of the cMRF clearly has a functional role in respiratory pattern formation. These findings revive the largely abandoned concept of a dorsal respiratory group that contributes to the generation of the respiratory motor pattern.

Expression of the transcription factor FOXP2 is linked to brain circuits that control motor function and speech. Investigation of FOXP2 protein expression in respiratory areas of the ponto-medullary brainstem of adult rat revealed distinct rostro-caudal expression gradients. A high density of FOXP2 immunoreactive nuclei was observed within the rostral pontine Kolliker-Fuse nucleus, compared to low densities in caudal pontine and rostral medullary respiratory nuclei, including the: (i) noradrenergic A5 and parafacial respiratory groups; (ii) Botzinger and pre-Botzinger complex and; (iii) rostral ventral respiratory group. Moderate densities of FOXP2 immunoreactive nuclei were observed in the caudal ventral respiratory group and the nucleus retroambiguus, with significant density levels found in the caudal half of the dorsal respiratory group and the hypoglossal pre-motor area lateral around calamus scriptorius. FOXP2 immunoreactivity was absent in all cranial nerve motor nuclei. We conclude that FOXP2 expression in respiratory brainstem areas selectively delineates laryngeal and hypoglossal pre-motor neuron populations essential for the generation of sound and voice.

Previous investigations from our laboratory have documented that the neuropil of the phrenic nucleus contains a dense accumulation of punctate nicotinamide adenine dinucleotide phosphate diaphorase staining. In this study we investigated the occurrence and origin of punctate nitric oxide synthase immunoreactivity in the neuropil of the phrenic nucleus in C3–C5 segments, supposed to be the terminal field of the premotor bulbospinal respiratory nitric oxide synthase-immunoreactive pathway in the dog. As the first step, nitric oxide synthase immunohistochemistry was used to characterize nitric oxide synthase-immunoreactive staining of the phrenic nucleus and nitric oxide synthase-containing neurons in the dorsal and rostral ventral respiratory group and in the Botzinger complex of the medulla. Dense punctate nitric oxide synthase immunoreactivity was found on control sections in the neuropil of the phrenic nucleus. Several thin bundles of nitric oxide synthase-immunoreactive fibers were found to enter the phrenic nucleus from the lateral and ventral column. Nitric oxide synthase-containing neurons were revealed in the dorsal respiratory group of medulla corresponding to the ventrolateral nucleus of the solitary tract and in the rostral ventral respiratory group beginning approximately 1 mm caudal to the obex and reaching to 650 μm rostral to the obex. Axotomy-induced retrograde changes, consisting in a strong upregulation of nitric oxide synthase-containing neurons, were found in the dorsal and rostral ventral respiratory group contralateral to the hemisection performed at the C2–C3 level. Concurrently, a strong depletion of the punctate nitric oxide synthase immunopositivity in the neuropil of the phrenic nucleus ipsilaterally with the hemisection was detected, thus revealing that a crossed premotor bulbospinal respiratory pathway contains a fairly high number of nitric oxide synthase-immunopositive fibers terminating in the phrenic nucleus. The use of the retrograde fluorescent tracer Fluorogold injected into the phrenic nucleus and an analysis of sections cut through the dorsal and rostral ventral respiratory group and Botzinger complex of medulla and processed for nitric oxide synthase immunocytochemistry revealed that approximately 73.8% of crossed premotor bulbospinal respiratory nitric oxide synthase-immunoreactive axons originate in the rostral ventral respiratory group and 26.2% is given by nitric oxide synthase-containing neurons of the dorsal respiratory group. A few premotor nitric oxide synthase-immunoreactive axons originating from the Botzinger complex were found. In summary, the present study provides evidence for a hitherto unknown premotor bulbospinal respiratory nitric oxide synthase-immunoreactive pathway connecting the bulbar respiratory centers with the motor neurons of the phrenic nucleus in the dog.

1. Long-term facilitation is a respiratory memory expressed as an increase in motor output lasting more than an hour. This change is induced by repeated hypoxia, stimulation of carotid chemoreceptors, or electrical stimulation of the carotid sinus nerve or brainstem mid-line. The present work addressed the hypothesis that persistent changes in medullary respiratory neural networks contribute to long-term facilitation. 2. Carotid chemoreceptors were stimulated by close arterial injection of CO(2)-saturated saline solution. Phrenic nerve efferent activity and up to 30 single medullary neurones were recorded simultaneously in nucleus tractus solitarii (NTS) including the dorsal respiratory group (DRG), Botzinger-ventral respiratory group (Böt-VRG), and nucleus raphe obscurus of nine adult cats, anaesthetized, injected with a neuromuscular blocking agent, vagotomized and artificially ventilated. 3. The firing rates of 87 of 105 neurones (83 %) changed following induction of long-term facilitation. Nine of eleven DRG and Böt-VRG putative premotor inspiratory neurones had increased firing rates with long-term facilitation. Fourteen of twenty-one raphe obscurus neurones with control firing rates less than 4 Hz had significant long-term increases in activity. 4. Cross-correlogram analysis suggested that there were changes in effective connectivity of neuron pairs with long-term facilitation. Joint peristimulus time histograms and pattern detection methods used with 'gravity' analysis also detected changes in short time scale correlations associated with long-term facilitation. 5. The results suggest that changes in firing rates and synchrony of VRG and DRG premotor neurones and altered effective connectivity among other functionally antecedent elements of the medullary respiratory network contribute to the expression of long-term facilitation.

It has been well established that inspiratory neurons of the dorsal respiratory group (DRG) are classified into two types based on whether they receive inputs from slowly-adapting pulmonary stretch receptors (SARs) or not. Inspiratory neurons with SAR inputs are called Iβ and the others are called Iα neurons. In this study, we identified a novel group of inspiratory neurons in Nembutal-anesthetized, paralyzed, and artificially-ventilated rats. (1) These DRG inspiratory neurons were activated characteristically by lung deflation. (2) They were orthodromically activated by electrical stimulation of the vagus nerve at a low intensity just above the threshold for afferents from SARs or rapidly-adapting pulmonary stretch receptors (RARs). (3) The orthodromic latencies (ranged from 1.9 to 2.5 ms) indicated that they receive direct inputs from low-threshold vagal afferents. (4) Unlike Iβ neurons, they hardly responded to lung inflation and never exhibited tonic firing in response to maintained lung inflation. (5) The majority (92%) of them were antidromically activated by electrical stimulation of the cervical spinal cord. These deflation-sensitive inspiratory neurons clearly form a distinct group, and their firing pattern is consistent with the hypothesis that they receive inputs from RAR afferents as well as the central inspiratory drive. The results indicating that DRG inspiratory neurons are classified into at least three groups provide new insights into the organization and role of the DRG.

Respiratory and cardiovascular responses to microinjections (10 nl) of l -glutamate (10 mM) into the dorsal medulla were studied in spontaneously breathing urethane-anesthetized, adult male Wistar rats. A total of 10 patterns of respiratory and cardiovascular responses were observed: (1) hypotension alone; (2) hypotension and bradycardia; (3) hypotension and apnea; (4) hypotension, bradycardia, and apnea; (5) apnea alone; (6) hypotension and fast and shallow breathing; (7) hypotension, bradycardia, and fast and shallow breathing; (8) fast and shallow breathing alone; (9) sighs; and (10) increase in BP and HR accompanied with fast and shallow breathing. The sites from which a combination of hypotension, bradycardia, and apnea was elicited, occupied a region in the medial subnucleus of nucleus tractus solitarius (nTS), the reticular formation just ventral to it, and the dorsal motor nucleus of vagus. The sites from which hypotension alone or a combination of hypotension and apnea were elicited occupied the margins of the medial subnucleus of nTS. The sites from which apnea alone was elicited were located in the ventrolateral part of nTS and the reticular formation just ventral to it. In the commissural subnucleus of nTS, the responses comparable to those elicited by peripheral chemoreceptor stimulation (i.e., increase in BP, HR, and respiratory rate) were located in a midline region just caudal to the calamus scriptorius, the sites from which sighs were elicited were located slightly lateral and deeper, the sites from which fast and shallow breathing were elicited were located in the dorsal portion, slightly lateral to the midline. These results are expected to prove useful in studies in which microinjection technique is used to identify transmitters/receptors involved in mediating respiratory and cardiovascular reflex responses.

In order to characterize the brain stem circuitry that produces vocalization, the activities of brain stem respiratory neurons were recorded extracellularly during vocalization induced by electrical stimulation of the periaqueductal gray in decerebrate cats. After the onset of stimulation, the respiratory rhythm ceases, and a preparatory inspiration is induced. Following this initial inspiration, vocalization characterized by increased activities of the intrinsic laryngeal adductor and the major expiratory muscles is induced. During vocalization, most of the dorsal respiratory group inspiratory neurons increase their firing rates in phase with an increase of diaphragm activity. Inspiratory neurons with a continuous discharge pattern in the rostral ventral respiratory group increase their firing rates to augment intrinsic laryngeal abductor motoneurons and bulbospinal inspiratory neurons in the dorsal respiratory group. On the other hand, most of the bulbospinal augmenting expiratory neurons in the Bötzinger complex cease firing just after the onset of periaqueductal gray stimulation for the remainder of the stimulation period. These results indicate that at least some part of the coordinated activations of intrinsic laryngeal and respiratory muscles during vocalization are mediated via the central respiratory neurons that produce breathing.

To clarify the difference between the firing patterns of the dorsal respiratory group (DRG) and ventral respiratory group (VRG) neurons during spontaneous augmented breaths, extracellular single unit recording of 139 respiratory-related neurons (inspiratory: 98/139, expiratory: 41/139) was performed in pentobarbital anesthetized rats. Both the I and E neurons were further classified into six groups: (1) I-augmenting, (2) I-decrementing, (3) I-other cells, (4) E-augmenting, (5) E-decrementing and (6) E-other cells. During the augmented breaths, most inspiratory neurons in the DRG (25/26) and VRG (67/72) show an increase in their discharge frequency irrespective of the cell type, but the discharges after an augmented breath were inhibited. Changes in these inspiratory neurons coincided with those of diaphragm electromyogram activity. With regard to relative changes in the mean firing rate during the inspiratory phase II of augmented breaths, there was a significant difference between the I-augmenting DRG and VRG neurons (353.5 +/- 56.9% vs. 237.5 +/- 17.1%, P0.01), but not in the I-decrementing and I-other neurons. On the other hand, the activities of the expiratory DRG (9/9) and VRG (27/32) neurons decreased during the augmented breath. A significant difference in the relative mean firing rate during the expiratory phase of augmented breaths was observed between the E-decrementing DRG and VRG neurons (27.3 +/- 5.2% vs. 58.0 +/- 6.3%, P0.05), but not between the E-augmenting and E-other neurons. These results suggested that during spontaneous augmented breaths the firing patterns of the DRG neurons were not qualitatively different from those of the VRG neurons.

Lateral lemniscus, a relay nucleus of auditory sensation, is involved in the control of phonatory movements such as human speech and vocalization of animals. The present study was designed to test whether neurons in the lateral lemniscus contributed to the control of swallowing, one of non-phonic oro-pharyngolaryngeal movements. In acutely decerebrated cats (n=15), swallowing was induced by electrical stimulation (20-80μA at 10Hz for 20s with rectangular pulses of 0.2ms duration) delivered to the superior laryngeal nerve (SLN). Repetitive electrical stimulation (30-50μA at 50Hz for 10-20s) applied to the dorsal nucleus of the lateral lemniscus (LLD) increased the number and reduced the latency to the onset of the SLN-induced swallowing. On the other hand, stimulation of the ventral nucleus of the lateral lemniscus and the paralemniscal area, corresponding to the ventrolateral part of the parabrachial nucleus and the Kolliker-Fuse nucleus, often suppressed the SLN-induced swallowing. Microinjection of NMDA (0.1-0.15μl, 5.0-10mM) into the LLD through a stereotaxically placed glass micropipette facilitated the SLN-induced swallowing, i.e., the number was increased and the latency of swallowing was reduced. We also injected muscimol (a gamma amino-butyric acid (GABA)A receptor agonist), bicuculline (a GABAA receptor antagonist) and baclofen (a GABAB receptor agonist) into the LLD (0.1-0.15μl and 5.0mM for each substance). It was observed that an injection of muscimol suppressed the SLN-induced swallowing. However, an injection of bicuculline facilitated the swallowing. An injection of baclofen did not alter the swallowing. These results suggest the presence of functional topography in the lateral lemniscus and the paralemniscal area in relation to the control of swallowing. The facilitatory LLD-effects on swallowing are modulated by glutamatergic and GABAergic receptors on neurons in the LLD.

It was hypothesized that, because rats appear to lack a prominent disynaptic projection from the dorsal respiratory group to phrenic motoneurons (Phr), they would lack the short-latency excitation of Phr output seen in cats in response to stimulation of some cranial nerve afferents. Single-pulse superior laryngeal nerve (SLN) stimulation elicited a short-latency bilateral excitation of glossopharyngeal (IX) and hypoglossal (XII) nerves and an ipsilateral excitation of pharyngeal branch of vagus (PhX) in 67% of rats, but no excitation of Phr. Vagus (X) stimulation elicited a bilateral excitation of Phr and a predominantly ipsilateral excitation of IX and PhX. Single-pulse stimulation of SLN or X also elicited longer-latency, bilateral decreases in activity of all recorded nerves. Repetitive stimulation (50 Hz) of SLN or X suppressed inspiratory activity and prolonged expiration. Lung inflation (7.5 cmH2O) inhibited Phr and PhX activity; X stimulation inhibited Phr but prolonged PhX activity. In conclusion, rats predictably lack the SLN-induced short latency Phr excitation but exhibit other short latency reflexes for which the underlying circuitry is not clear.

Prostaglandin (PG) E 2 acts on the brain stem to modulate breathing activity in the ovine fetus. The source of this PGE 2 is unknown and we hypothesized that it is produced locally in the developing brain and functions in a paracrine and/or autocrine manner. The purpose of the present study was to establish whether prostaglandin-H synthase-1 (PGHS-1), a crucial enzyme in de novo prostaglandin synthesis, is present and its gene expressed in the ovine fetal brain. Immunohistochemical and molecular hybridization techniques were used to identify sites of PGHS-1 immunoreactivity and PGHS-1 mRNA expression respectively in the brain of the ovine fetus in late gestation (approximately 126 days gestation, term 145 days). PGHS-1 immunoreactivity was localized to specific regions of the fetal brain, including the cortex, hypothalamus, hippocampal formation, superior colliculus of the midbrain, parabrachial nucleus of the pons, and the reticular formation, raphe, nucleus of the solitary tract, and gracile and cuneate nuclei of the medulla. The relative abundance of PGHS-1 mRNA in selected brain regions, as determined by Northern blot analysis, correlated qualitatively with the number of PGHS-1 immunoreactive neurons identified in each region. In situ hybridization demonstrated PGHS-1 mRNA to be localized in the same neurons or nuclei as PGHS-1 immunoreactivity. These results indicate that PGHS-1 is synthesized de novo in many brain regions including two that are important in respiratory control: the pneumotaxic center (parabrachial nucleus) and the dorsal respiratory group (nucleus tractus solitarius) suggesting that prostaglandins that modulate fetal respiratory activity are synthesized endogenously.

This study tested the hypothesis that the short-latency excitation of the phrenic motor output elicited by superior laryngeal nerve (SLN) stimulation requires non-NMDA receptor-mediated neurotransmission in the region of the dorsal respiratory group (DRG) of the adult cat. Injection of the non-NMDA receptor antagonist NBQX into the DRG severely attenuated or abolished the short-latency excitation, indicating that the short-latency excitation requires non-NMDA receptor-mediated neurotransmission within the DRG.

1. The vestibular system responds to accelerations of the head and produces reflex responses that serve a variety of compensatory functions. The neuronal circuitry that mediates vestibulo-respiratory reflexes is largely unknown. The purpose of the present study was to investigate the possible role of bulbospinal inspiratory neurons located in the para-ambigual region of the ventral respiratory group (VRG) in mediating these reflexes. Experiments were carried out in cats that were decerebrated, paralyzed, and artificially ventilated. 2. Activation of the vestibular nerve by electrical stimulation produced prominent bilateral reflex responses recorded from the phrenic nerve, which supplies the diaphragm. The responses could be complex and consisted of a decrease and/or increase in nerve discharge. 3. Extracellular recordings were made from 35 VRG inspiratory neurons that were antidromically activated from the upper cervical spinal cord. Almost one-half of these neurons (15/35, 43%) responded to vestibular stimulation. The neuronal response patterns were consistent with VRG inspiratory neurons contributing to the vestibular reflex response simultaneously recorded from the phrenic nerve. 4. The present results indicate that approximately one-half of VRG bulbospinal inspiratory neurons contribute to vestibulo-respiratory reflexes. These findings are in contrast to our recent neuroanatomic and electrophysiological studies which revealed a paucity of vestibular inputs to the dorsal respiratory group (DRG) located in the ventrolateral nucleus of the solitary tract. Thus there appears to be a difference between inspiratory neurons in the DRG and VRG in regard to participating in vestibulo-respiratory reflexes.

The interaction between afferent inputs from carotid body chemoreceptors (CCRs) and from slowly adapting pulmonary stretch receptors (PSRs) on the discharge patterns of medullary inspiratory (I) and expiratory (E) neurons was characterized in thiopental sodium-anesthetized, paralyzed, ventilated dogs. A cycle-triggered ventilator was used to produce control and test pulmonary afferent input patterns. The CCRs were stimulated by phase-synchronized bolus injections of CO2-saturated saline into the common carotid arteries. Only those neurons whose discharge time course was altered by both inflation and CCR activation were studied. The dorsal respiratory group (DRG) I inflation-insensitive neurons were also included. Cycle-triggered histograms of unit activity were obtained for the neuronal responses to inflation, CO2 bolus, and their combination, as well as for the spontaneous control condition. Linearity of the interaction was tested by comparing the sum of the net individual responses to the net response of the combined afferent inputs. The results suggest that a linear (additive) interaction between CCR and PSR inputs exists for the DRG I inflation-sensitive neurons, the ventral respiratory group (VRG) I decrementing, and caudal VRG E augmenting neurons, while a nonadditive interaction exists for caudal VRG E decrementing bulbospinal neurons. The implications of these findings are discussed.

There has been controversy whether the dorsal respiratory group (DRG), identified in the cat and several other species as a concentration of mainly inspiratory neurons located in the ventrolateral subnucleus of the solitary tract, also exists in the rat. The aim of this study was to re-examine this question by systematically exploring this region with extracellular microelectrodes, in anesthetized and artificially ventilated rats. One-hundred and forty-two units were recorded which fired in phase with central respiratory cycles (determined by recording from the phrenic nerve) and/or lung inflations. One-hundred and nineteen recordings were thought to be from neuronal cell bodies (confirmed in some cases by excitatory responses to microelectrophoretic administration of dl -homocysteic acid), while the remaining 23 were from lung vagal afferents. Most neurons in the former group (87/119) were inspiratory. Out of 96 neurons tested for spinal projections only 14 (12 inspiratory, 2 expiratory) responded antidromically following stimulation at C3 segment. These results confirm the existence of the DRG in the rat and demonstrate that neurons located in this region have firing patterns generally similar to those previously described in the cat. The main difference is the relative paucity in the rat of neurons projecting spinally below the C2 level, which indicates that most DRG neurons in this species do not project directly to phrenic and intercostal motoneurons, but to other, as yet unidentified, neuronal groups within the brainstem or upper cervical segments.

Selective neuronal vulnerability of the motor cortex, basal ganglia, brainstem, medulla, cerebellum, C6 and L6 segments of the spinal cord were studied after 15 min of cardiac arrest followed by 1 h of normoxic or hyperoxic resuscitation using the suppressive Nauta method in dogs. Hyperoxic resuscitation causes characteristic somatodendritic argyrophilia of the interneuronal pool in the spinal cord and lower medulla. Cuneate, lateral reticular, supraspinal, and caudal trigeminal nuclei as well as the dorsal and ventral respiratory neuronal groups were heavily involved. Similarly, the Purkinje cells, neurons in the middle and deep portions of the mesencephalic tectum, perirubral, pretectal, posterior commissure, middle-sized striatal and giant pyramidal (Betz's) neurons in the motor cortex became argyrophilic. Hyperoxic resuscitation versus normoxic resuscitation causes statistically significant somatodendritic argyrophilia of the dorsal respiratory group, cuneate, dorsal lateral geniculate and thalamic reticular nuclei.

The location and axonal projection of a type of respiratory neurones (termed bIE neurones), which show burst firing at the time of phase transition from inspiration to expiration, were studied in Nembutal-anaesthetized, paralysed and artificially ventilated cats. The bIE neurones showed maximum firing at the sharp decline of inspiratory activity. All of the bIE neurones tested projected to medullary respiration-related areas of the ventral respiratory group (VRG), the dorsal respiratory group (DRG), and the Bötzinger complex (BOT). The bIE neurones were distributed in the dorso-medial border of the main assembly of respiratory neurones of the VRG and BOT. The possibility that these bIE neurones participate in inspiratory termination is discussed.

A great number of severed central nervous system (CNS) neurons of the adult rat have the capability to regrow axons into peripheral nerve autografts. In the present experiment, autologous segments of the peroneal nerve were inserted perpendicularly to the dorsal surface of the medulla oblongata more or less laterally within either the ventral respiratory group or the so-called dorsal respiratory group (ventrolateral grafts, n = 5; dorsolateral grafts, n = 5). From 2 to 4.5 months after the graft implantation, spontaneous unitary activities ( n = 197) were recorded within all the grafted nerves: they were found to arise from both central respiratory (R, n = 60) and non-respiratory (NR, n = 137) neurons which were giving off regenerated axons along the nerve grafts. The graft reinnervation by respiratory axons was found to be significantly more abundant within the medullary ventrolateral grafts than within the dorsolateral ones. The low rate of axonal regeneration from respiratory neurons observed within the dorsolateral grafts provides further evidence that the number of the respiratory neurons in the dorsal respiratory group, if present at all, is much smaller than that of the ventral respiratory group in the rat.

We investigated connections between the rostral ventrolateral medulla (rVLM) and the two main inspiration-related medullary areas, i.e., the dorsal respiratory group (DRG) and the rostral ventral respiratory group (rVRG) in the cat. Non respiration-related tonically firing units encountered in the rVLM displayed either antidromic or orthodromic responses to DRG or rVRG microstimulation. Some units responded to the stimulation of both regions. We suggest that at least part of rVLM neurons are components of medullary loops operating in the control of breathing.

Axonal projections of second order neurons activated by vagal afferent fibers originating from pulmonary rapidly adapting receptors (RARs) were studied electrophysiologically in Nembutal-anesthetized, paralyzed and artificially ventilated cats. Extracellular recordings from these neurons (referred to as 'RAR-cells') were made in the commissural subnucleus (COM) of the nucleus of the solitary tract (NTS). They were identified by a combination of stimuli including stimulation of the vagus nerve(s), hyperinflation and deflation of the lungs, and a brief period of ammonia vapor inhalation. A total of 80 RAR-cells were tested for axonal projections to a respiration-related area in the brain-stem, either the dorsolateral rostral pons or the dorsal respiratory group (DRG) or the ventral respiratory group (VRG) or the Bötzinger complex (BOT) and/or the spinal cord. Twenty-two of the 47 (47%) RAR-cells tested for ipsilateral pontine projection could be antidromically activated, and in 8 cases evidence for axonal arborization was obtained. Only 1 of the 11 RAR-cells tested for DRG projection, and 1 of the 10 RAR-cells tested for VRG projection, were antidromically activated. No RAR-cells were activated from the BOT (n = 8) or from the C3-C4 segments of the spinal cord (n = 11). Bilateral lesions of the COM abolished the reflex responses induced by ammonia inhalation or hyperinflation of the lungs, but not the Hering-Breuer reflex. These results indicate that a pathway from the COM to the rostral pons forms part of the reflex arc originating from RARs.

To determine distribution, discharge patterns and the spinal projections of medullary respiratory neurons (RNs), a systematic mapping of 806 RNs was made in the medulla of anesthetized rabbits. In disagreement with previous reports that there are no discrete medullary respiratory neuronal groups in rabbits, two neuronal groups were identified: (1) dorsal respiratory group (DRG), associated with the nucleus tractus solitarius; and (2) ventral respiratory group (VRG), associated with the nucleus ambiguus compact formation. The density of RNs in the DRG was much lower than that in the VRG. In the VRG, 3 subdivisions of RN populations were found: predominantly expiratory neurons in the caudal and the rostral parts, and mainly inspiratory neurons in the intermediate region. Nine distint types of RNs were classified on the basis of firing patterns. Nearly all types were found in both the DRG and each VRG subdivision. Antidromic mapping of 64 VRG neurons revealed that 67% projected to the spinal cord. Expiratory bulbospinal neurons in the rostral subdivision of the VRG projected only to the cervical cord (mainly ipsilaterally). Most neurons of the intermediate and caudal subdivisions of the VRG (74%) appeared to project either contralaterally or ipsilaterally below T1. The axonal conduction velocity was 40–50 m/s by two-point determinations. We conclude that respiratory neuronal groups in the medulla of the rabbit are generally similar to those of the cat. Nearly equal proportions of bulbospinal RNs projected to the ipsilateral vs contralateral spinal cord.

To determine whether bulbar respiratory neurons are active in a manner similar to respiratory motor nerves, in which diverse activity patterns have been observed in dogs, the discharge of 112 bulbospinal and 155 propriobulbar respiratory neurons was observed during retching and vomiting evoked by stimulation of the vagal nerve in decerebrate, paralyzed dogs. During retching, the respiratory activities of all respiratory neurons changed. The discharge patterns during retching were classified into seven types in accordance with the discharge phase in the retching cycle and the discharge frequency. Inspiratory bullbospinal neurons (39) in the ventral respiratory group (VRG) exhibited discharge synchronous with retches at higher (14/39) or lower (4/39) frequencies than those of their respiratory discharge, discharge between retches at higher (3/39) or lower (14/39) frequencies and discharge depressed throughout the retching period (4/39). Inspiratory bulbospinal neurons (30) in the dorsal respiratory group (DRG) and expiratory bulbospinal neurons (23) in the VRG also fired with some of the seven types of discharge. No significant difference was recognized between the ratio of the numbers of inspiratory bulbospinal neurons firing with the seven types of discharge and that of expiratory bulbospinal neurons. On the other hand, the respiratory propriobulbar neurons (73) in the VRG and those (45) in the DRG seemed to be differently activated; that is, 44% of the propriobulbar neurons in the VRG fired synchronously with retches at a higher frequency than their frequency during breathing, while only 22% of those in the DRG fired at such a high frequency with retches. However, the ratio of whole inspiratory neurons (174) exhibiting the seven types of discharge was almost the same as that of whole expiratory neurons (93). During vomiting (fictive expulsion), three of five expiratory neurons fired, but most (14/16) of inspiratory neurons did not. In conclusion, all of the activity patterns which have been observed during retching in motor fibers innervating respiratory muscles are found in both bulbospinal and propriobulbar respiratory neurons in dogs. Furthermore, since the firing patterns of inspiratory bulbospinal neurons during retching in the dogs in this study are apparently different from those in cats (Miller et al., 1990), inspiratory neurons and muscles seem to be differently organized during retching in cats and dogs.

The purpose of this study was to examine whether treadmill running exercise after phrenicotomy resulted in change in neuronal expression of c-Fos protein in the medulla oblongata in Wistar rats. All rats were subjected to left-sided unilateral phrenicotomy. In test experiments, rats were trained with treadmill running for 4 weeks, while in control experiments rats were not provided with the opportunity to exercise. c-Fos protein, a transient transcription factor, was immunohistochemically stained by the ABC method, and surveyed in the medulla oblongata. Numbers of c-Fos-immunoreactive (c-Fos-ir) neurons were counted microscopically, and their distribution in the medulla oblongata was mapped. In the dorsal respiratory group (DRG), numbers of c-Fos-ir neurons were found to be significantly increased in the test rats compared with control rats, while in the ventral respiratory group (VRG) no significant differences were found in numbers of c-Fos-ir neurons between test and control rats. These findings suggest that DRG neurons may play more important roles than VRG neurons in recovery of respiration in treadmill exercise after unilateral phrenicotomy in Wistar rats.

The effect of enflurane on the firing activity (spikes/sec) of the inspiratory neurons of the dorsal respiratory group (DRG) of the medulla oblongata was studied in decerebrate, paralyzed, mechanically ventilated cats before and after bilateral cervical vagotomy. Inspiratory neuronal activity, phrenic neurogram, arterial blood pressure, tracheal pressure, and end tidal CO2 concentration were recorded. Cells whose firing activity was in phase with that of the phrenic nerve were considered inspiratory neurons. Administration of 1 and 2% enflurane in oxygen produced gradual, significant, and dose-dependent depression of the cell activity with cervical vagi either intact or severed. Recovery of the cell activity occurred after termination of enflurane administration. In cats with intact vagi, 10 min after introduction of 1 and 2% enflurane, the cell activity (mean +/- SE) expressed as percentage of the control was 70 +/- 6% (P less than 0.05) and 48 +/- 5% (P less than 0.01), respectively. Bilateral cervical vagotomy did not affect the degree of cell depression due to enflurane. Hypercarbia induced by inhalation of 5% CO2 increased cell activity, but it did not block enflurane-induced cell depression, although it reduced it. It may be concluded that enflurane depresses the activity of the inspiratory neurons of the DRG. The results also suggest that the respiratory depressant effect of enflurane has a central component and that the DRG region may serve as a site to mediate the enflurane-induced respiratory depression.

Brainstem projections to the phrenic nucleus were studied in rabbits using horseradish peroxidase conjugated with wheat germ agglutinin (WGA-HRP) as a retrograde and anterograde neuronal tracer. Injections of 1% WGA-HRP were centered in the phrenic nucleus in the C4-C5 ventral horn in 4 rabbits to identify pontomedullary nuclear groups that contain neurons projecting to the midcervical spinal cord. Regions of the rabbit brainstem that are homologous to the ventral respiratory group (VRG), dorsal respiratory group (DRG), Bötzinger Complex (BötC) and Kölliker-Fuse nucleus in the cat and rat were shown to provide the major pontomedullary projections to the phrenic nucleus. Injections of WGA-HRP into physiologically identified locations within DRG, VRG and BötC anterogradely labelled bulbospinal axons of these groups. These injections produced presumptive terminal labelling in the C4-C5 ventral horn in the region containing the phrenic cell column and the transverse phrenic motoneuron dendrite bundles as defined by WGA-HRP labelling of phrenic motoneurons. These results indicate: 1) The presumptive excitatory (DRG, VRG) and inhibitory (BötC) bulbospinal control of phrenic motoneurons arise from the same medullary respiratory groups in the rabbit as in the cat and rat. 2) The bulbospinal control of phrenic motoneurons is primarily via direct projections to the phrenic motor nucleus, and not through segmental propriospinal interneurons. 3) As in the rat, the bulbospinal contribution of the DRG is less pronounced in the rabbit than in the cat. 4) The rabbit and rat have a slight ipsilateral predominance in their bulbospinal projections to phrenic nucleus; whereas these projections have a contralateral predominance in the cat.

In decerebrate, paralyzed and ventilated cats, we monitored the intracellular responses of 30 inspiratory neurons of the dorsal respiratory group (DRG) to stimulation of vagal and expiratory muscle (internal intercostal and abdominal) afferents. We hypothesized that the inhibitory effects of stimulation of expiratory muscle afferents, previously reported, would block the excitatory responses of inspiratory neurons of the DRG to vagal stimulation. Although prolonged stimulus trains to expiratory muscle afferents caused respiratory phase-switching, single shocks or short trains elicited no responses in 17 bulbospinal neurons, excitatory responses in 6, and inhibitory responses in 2. Of the 4 propriobulbar neurons tested, 2 had inhibitory responses and 2 did not respond. In only 2 neurons, both bulbospinal, did conditioning stimuli to expiratory muscle afferents block or reduce the excitatory effects of vagal stimulation. These results suggest that interaction of vagal and expiratory muscle afferents, which might account for the absence of a change in inspiratory duration despite increased vagal afferent feedback at elevated end-expiratory lung volumes, does not occur within the DRG.

1. We visualized respiratory-related neuron network activity in the dorsal part of the pons and medulla of an in vitro preparation from newborn rats by optical recordings using a voltage-sensitive dye. We measured optical signals from several seconds before to several seconds after the inspiratory phase using the inspiratory motor nerve discharge as the trigger signal and we averaged the optical signals of 20-50 respiratory cycles to obtain an optical image correlating specifically to inspiratory activity. 2. Four areas that were excited or inhibited corresponding to the respiratory cycles were detected. (i) The most rostral activity was in the rostral and lateral parts of the pons, with activity mainly in the inspiratory phase, corresponding to the pontine-respiratory group. (ii) In the midpontine level, inspiratory activity followed by long-lasting hyperpolarization appeared in the midlateral parts. This part was presumed to reflect activity in the locus coeruleus. The hyperpolarization became almost negligible after treatment with the alpha-adrenergic antagonist, phentolamine. (iii) In the dorsal medulla, the predominantly inspiratory activity was detected at the rostral level of the area postrema. This part was considered to reflect activity mainly of the hypoglossal nucleus. (iv) At a similar level, we also detected weak and disperse inspiratory activity extending more laterally and caudally than that of the hypoglossal nucleus activity. This might reflect activity of the dorsal respiratory group. 3. In conclusion, the present optical recording study revealed that the dorsal part of the lower brainstem in the in vitro preparation is noticeably active as well as the ventral part shown in the previous study. This method is very useful for analysis of pharmacological properties, as well as the spatio-temporal pattern of respiratory-related network activity in the brainstem.

We used spectral analysis and event-triggered averaging to determine the effects of chemical inactivation of the medullary lateral tegmental field (LTF) on 1) the relationship of intratracheal pressure (ITP, an index of vagal lung inflation afferent activity) to sympathetic nerve discharge (SND) and phrenic nerve activity (PNA) and 2) central respiratory rate in paralyzed, artificially ventilated dial-urethane-anesthetized cats. ITP-SND coherence value at the frequency of artificial ventilation was significantly ( P < 0.05; n = 18) reduced from 0.73 ± 0.04 (mean ± SE) to 0.24 ± 0.04 after bilateral microinjection of muscimol into the LTF. Central respiratory rate was unexpectedly increased in 12 of these experiments (0.28 ± 0.03 vs. 0.95 ± 0.25 Hz). The ITP-PNA coherence value was variably affected by chemical inactivation of the LTF. It was unchanged when central respiratory rate was also not altered, decreased when respiratory rate was increased above the rate of artificial ventilation, and increased when respiratory rate was raised from a value below the rate of artificial ventilation to the same frequency as the ventilator. Chemical inactivation of the LTF increased central respiratory rate in four of six vagotomized cats but did not significantly affect the PNA-SND coherence value. These data demonstrate that the LTF 1) plays a critical role in mediating the effects of vagal lung inflation afferents on SND but not PNA, 2) helps maintain central respiratory rate in the physiological range, but 3) is not involved in the coupling of central respiratory and sympathetic circuits.

Previous transneuronal tracing studies conducted in the ferret revealed that a large population of neurons that provides inputs to diaphragm and abdominal motoneurons is located in the ventral magnocellular portion of the medial medullary reticular formation. These observations raise the possibility that the neural substrate underlying respiratory rhythmogenesis may be different in the ferret than in other species in which this circuitry has been explored. In the present study, systematic tracking was conducted through the ferret medulla to map the locations of neurons with activity related to the contractions of respiratory muscles. As in the cat, rat, and rabbit, neurons with respiratory-related discharges were distributed either lateral or ventrolateral to the solitary nucleus (dorsal respiratory group) or in the vicinity of nucleus retroambigualis, nucleus ambiguus and the retrofacial nucleus (ventral respiratory group). Although the general organization of respiratory group neurons appeared to be similar in the ferret to that in other mammals, a difference was that few expiratory neurons were located rostrally in the ventral respiratory group. These data suggest that the ventral magnocellular medullary reticular formation is not essential for respiratory rhythm generation, at least during quiet breathing, but may participate in regulating the excitability of respiratory motoneurons or in coordinating the contractions of respiratory muscles during nonrespiratory responses (e.g. coughing or emesis).

We examined the role of dorsal respiratory group (DRG) inspiratory neurons as transmitters of respiratory drive to phrenic and intercostal motoneurons and as relays of afferent information to ventral respiratory group (VRG) bulbospinal, inspiratory neurons. Attempts to antidromically activate 76 DRG neurons from the spinal cord at the C7 segment resulted in only 4 (5.3%) successes (3 contralateral, 1 ipsilateral). Cross-correlating DRG neuron discharge with that of the ipsilateral (56) and contralateral (20) phrenic nerve detected common activation peaks in 2 and 3 cases respectively, with no evidence for monosynaptic connections. Cross-correlating DRG neuron discharge with that of bulbospinal, inspiratory VRG neurons found some evidence for interaction. Peaks in 7 of 73 (10%) cross-correlation histograms were attributed to a monosynaptic excitation of DRG neurons by VRG neurons, although a common activation cannot be ruled out; troughs, some with an accompanying peak, in 9 (12.3%) histograms were interpreted as a combined excitation of the DRG neuron and delayed inhibition of the VRG neuron. In addition, 2 cross-correlation histograms showed peaks with latencies and half-amplitude widths consistent with a disynaptic excitation of a DRG neuron by a bulbospinal inspiratory VRG neuron. Cross-correlating the discharge of 57 pairs of DRG inspiratory neurons (6 contralateral) detected common activation peaks in 7 (12.3%) cases (none contralateral) and one case interpreted as evidence for a disynaptic excitation. These findings suggest that the role of the DRG inspiratory neurons in rats differs from that in cats, primarily because they do not act to transmit respiratory rhythmic drive directly to phrenic and intercostal motoneurons. The results offer some support for an excitation of DRG neurons by VRG inspiratory neurons, but no support for a role of DRG inspiratory neurons as mediators of afferent information transfer to VRG bulbospinal inspiratory neurons.

Extracellular recordings were made in the dorsal respiratory group (DRG) and adjacent reticular formation following single-shock stimulation of the anterior ethmoidal nerve (AEN) and during sneeze evoked by repetitive stimulation of the AEN in nembutal-anaesthetized, curarized and ventilated cats. These neurones were characterised according to (i) their activity during the respiratory cycle (as inspiratory augmenting or decrementing (I Aug or I Dec), expiratory augmenting or decrementing (E Aug or E Dec), silent or tonic), and (ii) their axonal projection (bulbospinal or non-bulbospinal-non-vagal (BS or NBS-NV)). Following single-shock stimulation of the AEN, most of the inspiratory neurones were transiently inhibited, whereas E Aug neurones were activated and E Dec neurones were activated and then inhibited. Silent neurones responded with a multispike or a paucispike pattern. Following repetitive stimulation of the AEN and during the resulting sneeze reflex, I Aug neurones increased their activity in parallel with the phrenic activity, I Dec neurones fired at the onset and at the end of the inspiration, E Dec and some silent neurones fired either during the compressive phase or after the expulsive phase, whereas E Aug and some silent neurones fired during the expulsive phase. We conclude that sneeze involves a reconfiguration of the central respiratory drive which uses, at least partly, the respiratory network to trigger a non-ventilatory defensive motor act.

Membrane potential changes and/or discharges from 36 inspiratory neurons were recorded intracellularly in the dorsal respiratory group (DRG; i.e., the ventrolateral subdivision of the nucleus tractus solitarii) in decerebrate, paralyzed, and ventilated cats. Electrical activities were recorded from both somata (n = 10) and axons (n = 26). Activities during quiet breathing were compared with those observed during fictive coughing and swallowing evoked by repetitive electrical stimulation of afferent fibers of the superior laryngeal nerve (SLN). These nonrespiratory behaviors were evident in paralyzed animals as characteristic discharge patterns of the phrenic, abdominal, and hypoglossal nerves. Twenty-six neurons exhibiting antidromic action potentials in response to electrical stimuli applied to the cervical (C3-5) spinal cord were classified as inspiratory bulbospinal neurons (IBSNs). These neurons were considered as premotoneurons. The remaining 10 inspiratory neurons (I-NAA) were not antidromically activated by electrical stimuli applied to either cervical spinal cord or ipsilateral cervical vagus. These neurons are thought to be propriobulbar neurons. We recorded the activity of 31 DRG inspiratory neurons (24 IBSNs and 7 I-NAA) during coughing. All but one (a late-recruited IBSN) discharged a burst of action potentials during the coughing-related phrenic nerve activity. Typically, ramp-like membrane depolarization trajectories and discharge frequencies during coughing were similar to those observed during inspiration. We recorded the activity of 33 DRG inspiratory neurons (23 IBSNs and 10 I-NAA) during swallowing. Most (28/33) neurons were briefly activated, i.e., discharged a burst of action potentials during swallowing, but peak discharge frequency decreased compared with that measured during inspiration. The membrane potentials of nine somata exhibited a brief bell-shaped depolarization during swallowing, the amplitude of which was similar to that observed during inspiration. These results suggest that some inspiratory premotoneurons and propriobulbar neurons of the DRG might be involved in nonrespiratory motor activities, even if clearly antagonistic to breathing (e.g., swallowing). We postulate the existence in the medulla oblongata of adult mammals of neurons exhibiting a "functional flexibility".

1. This study addressed the hypothesis that there is a parallel processing of input from carotid chemoreceptors to brainstem neurones involved in inspiratory phase timing and control of inspiratory motor output amplitude. Data were from fifteen anaesthetized, bilaterally vagotomized, paralysed, artificially ventilated cats. Carotid chemoreceptors were stimulated by close arterial injection of 200 microliters of CO2-saturated saline solution. 2. Planar arrays of tungsten microelectrodes were used to monitor simultaneously up to twenty-two neurones in the nucleus tractus solitarii (NTS) and ventral respiratory group (VRG). Spike trains were analysed with two statistical tests of respiratory modulation, cycle-triggered histograms, peristimulus-time histograms, cumulative sum histograms and cross-correlograms. 3. In NTS, 16 of 26 neurones with respiratory and 12 of 27 without respiratory modulation changed firing rate during carotid chemoreceptor stimulation. In the VRG 72 of 112 respiratory and 14 of 48 non-respiratory neurones changed firing rate during stimulation. 4. The spike trains of 85 of 1276 pairs (6.7%) of cells exhibited short time scale correlations indicative of paucisynaptic interactions. Ten pairs of neurones were each composed of a rostral VRG phasic inspiratory neurone that responded to carotid chemoreceptor stimulation with a decline in firing rate and a caudal VRG phasic inspiratory neurone that increased its firing rate. Cross-correlograms from two of the pairs had features consistent with excitation of the caudal neurones by the rostral cells. A decrease in the duration of activity of the rostral VRG neurones was paralleled by the decrease in inspiratory time of phrenic nerve activity. Caudal VRG inspiratory neurones increased their activity as phrenic amplitude increased. Spike-triggered averages of all four neurones indicated post-spike facilitation of phrenic motoneurones. 5. The results support the hypothesis that unilateral stimulation of carotid chemoreceptors results in parallel actions. (a) Inhibition of rostral VRG I-Driver neurones decreases inspiratory duration. (b) Concurrent excitation of premotor VRG and dorsal respiratory group inspiratory neurones increases inspiratory drive to phrenic motoneurones. Other data suggest that responsive ipsilateral neurones act to regulate contralateral neurones.

Exposure of anesthetized paralyzed vagotomized peripherally chemodenervated cats to hypoxia results in initial depression and subsequent loss of the phrenic neurogram. To determine whether hypoxic respiratory depression results from the inhibition of respiratory premotor neurons by bulbospinal neurons of the Botzinger complex (Bot-E neurons), extracellular recordings were made of dorsal and ventral respiratory group bulbospinal inspiratory neurons and Bot-E neurons during acute hypoxic hypoxia. All neurons recorded decreased firing rate during hypoxia. Bot-E neurons became silent before the loss of phasic phrenic activity during hypoxia and commenced firing before or coincident with the return of the phrenic neurogram during reoxygenation. Inspiratory neurons ceased firing coincident with phrenic silence. Dorsal respiratory group and ventral respiratory group neurons that had a late onset of firing with respect to the phrenic neurogram during normoxia fired progressively earlier in inspiration during hypoxia, an effect that was reversed during reoxygenation. These data are consistent with inhibition and/or disfacilitation as the mechanism of hypoxic respiratory depression but suggest that Bot-E neurons are not the source of this inhibition.

Ipsilateral and contralateral pairs of augmenting expiratory neurons were recorded simultaneously from the Bötzinger complex using glass-coated tungsten microelectrodes in pentobarbitone-anaesthetized cats. The neurons were identified both by firing pattern and by antidromic activation from the contralateral site of the dorsal respiratory group. Cross-correlation histograms of the extracellularly recorded action potentials were calculated in order to detect short time-scale synchronizations of firing indicative of synaptic connections between the neurons. The cross-correlation histograms for 40 ipsilateral pairs of neurons less than 1 mm apart showed eight (20%) narrow troughs (mean half-amplitude width +/- SD, 1.1 +/- 0.37 ms) at short latencies (mean latency +/- SD, 1.0 +/- 0.35 ms) suggestive of monosynaptic inhibition. These included two cross-correlation histograms which showed troughs on both sides of time zero, indicating a mutual inhibition. For another four pairs of neurons (10%), a central broad peak suggestive of common activation due to either excitation or release from inhibition was evident. Contralateral pairs of expiratory neurons of the Bötzinger complex were examined in a similar manner. The cross-correlation histograms for 43 pairs of neurons showed five (12%) narrow troughs (mean half-amplitude width +/- SD, 1.2 +/- 0.67 ms) at short latencies (mean latency +/- SD, 2.7 +/- 1.47 ms) suggestive of monosynaptic inhibition. These included one cross-correlation histogram which showed troughs (one not statistically significant) on both sides of time zero, indicating a mutual inhibition. For another two pairs of neurons (4.6%) a central, broad peak suggestive of common activation due to either excitation or release from inhibition was evident. We conclude that inhibitory interconnections exist between augmenting expiratory neurons of the Bötzinger complex ipsilaterally and contralaterally. These connections may synchronize the expiratory burst of activity within this population and assist in the patterning of the burst.

The parabrachial complex, also known as the pneumotaxic center or pontine respiratory group, has long been recognized as an important participant in respiratory control. One line of evidence supporting this idea is the demonstration of changes in breathing pattern following injection of neuroactive substances into or near the parabrachial complex. However, it is not yet known exactly which cell groups and projections mediate those responses. In order to address this issue, we explored the topographic organization of respiratory responses to chemical stimulation of the parabrachial complex of the rat and examined the descending projections of the most sensitive sites. Injection of glutamate (5–100 pmol) at specific sites in or near the parabrachial nucleus produced three distinct site-specific response patterns. First, hyperpnea followed glutamate injection into far rostral and midcaudal areas of the Kolliker-Fuse nucleus and most of the lateral parabrachial nucleus, including the external lateral, central lateral, dorsal lateral, and superior lateral subnuclei. Threshold hyperpneic effects were manifested as single, deepened breaths of premature onset. Suprathreshold doses of glutamate at these locations produced tachypnea. Neurons in these sites projected to the ventral respiratory group in the ventrolateral medulla. Second, the most intense inspiratory facilitatory responses were seen at mid to rostral levels of the Kolliker-Fuse nucleus, near the ventrolateral tip of the superior cerebellar peduncle. Even at threshold doses of glutamate, exhalation was incomplete, resulting in a breathing pattern that resembled apneusis (an inspiratory cramp). This site contained an especially dense cluster of neurons that projected either to the ventrolateral medulla or to the dorsal respiratory group in the nucleus of the solitary tract, but not to both areas. The third type of response, decreases in respiratory rate, occurred following glutamate injection at the most lateral and ventral boundaries of the Kolliker- Fuse nucleus. The most sensitive apneic sites were not found in the parabrachial nucleus but along the dorsal and medial edge of the principal sensory trigeminal nucleus and extending ventrally between the sensory and motor trigeminal nuclei. Scattered neurons in these sites were retrogradely labeled from the ventral but not the dorsal respiratory group. These results indicate that there are anatomically and functionally distinct cell populations in and near the parabrachial complex that, when chemically stimulated, can produce specific and sometimes opposing effects on respiration. The predominant effect of lateral parabrachial stimulation is respiratory facilitation, while inhibitory effects are elicited by trigeminal injections of glutamate.

The rostral ventrolateral medulla (rVLM) is known to play an important role in cardiorespiratory control. In the rVLM an 'apnoea region', in which unilateral focal blocks induce strong depressant effects on inspiratory activity up to complete apnoea, has been described. This study was designed to systematically investigate the effects provoked by unilateral micro-injections (10-30 nl) of D,L-homocysteic acid 160 mM into this region on respiratory activity and arterial blood pressure in pentobarbitone anaesthetized, vagotomized, paralyzed and artificially ventilated cats. Micro-injections into the rostral portion of this area caused depressant respiratory responses up to complete apnoea, while micro-injections into more caudally located sites induced excitatory respiratory responses. Similar effects were observed in the activity of phrenic nerves and inspiration-related medullary neurons of both the dorsal and ventral respiratory group. The respiratory responses could be accompanied by marked increases in blood pressure (> or = 30 mmHg), especially at locations ventral to the retrofacial and facial nucleus; however, they could also occur in the absence of appreciable changes or even in association with slight decreases in blood pressure. Similar respiratory and pressor effects were observed in carotid sinus denervated cats. The results indicate that two distinct rVLM neuronal populations, one located more rostrally and the other more caudally, may have an important role in the genesis and/or maintenance of respiratory rhythm by exerting respectively inhibitory and excitatory influences on inspiratory activity. Furthermore, they support the hypothesis that different neural substrates of the rVLM are involved in the regulation of respiratory and cardiovascular functions.

The location and firing patterns of medullary respiratory neurons have been described in a small number of species. The cat has been the most widely studied species, but some potentially important differences have recently been noted in others. A more complete survey of species is required to determine the significance of these differences. We describe the location and firing patterns of respiratory neurons in the medulla of anesthetized, paralyzed and mechanically ventilated adult guinea pigs. Extracellular single-unit recordings were made from the medulla, their phase relationship with phrenic nerve activity used to define them as respiratory and their location marked with fast green. Respiratory units were concentrated ventrolateral to the nucleus tractus solitarius (NTS) and within and surrounding the nucleus ambiguus (NA), corresponding to the dorsal respiratory group (DRG) and ventral respiratory group (VRG) of the cat, respectively. Most DRG respiratory units were inspiratory, while the VRG contained equal numbers of inspiratory and expiratory units. The DRG and VRG both contained early, late and constant-frequency inspiratory and expiratory units. In general, these findings are similar to those in other mammalian species examined, consistent with these basic aspects of the respiratory network being highly conserved.

Synaptic inputs from afferents in the superior laryngeal nerve (SLN) to medullary respiratory neurons (n=154) in the dorsal respiratory group (DRG), ventral respiratory group (VRG) and the region of the Bo¨tzinger complex (BOT) were studied in anesthetized cats. Single pulse stimulation of the SLN-evoked monosynaptic EPSPs in most inspiratory bulbospinal (I-BS) neurons in the DRG, and disynaptic or oligosynaptic chloride-dependent IPSPs in other I-BS neurons in the DRG and VRG. Stimulation of laryngeal afferents also inhibited oligosynaptically expiratory bulbospinal neurons in the VRG, and all types of respiratory neurons recorded in the BOT region. Oligosynaptic potentials (usually EPSPs) were recorded in inspiratory and expiratory laryngeal motoneurons. These results provide evidence of a processing of SLN-evoked synaptic responses by all tested groups of medullary respiratory neurons. The pathways mediating these synaptic responses are discussed.

Experiments were performed on anaesthetized, vagotomized rabbits. Respiratory movement and phrenic rhythmical discharge were reversibly abolished by the symmetrical injection of 1% procaine into the medial area of the nucleus retrofacialis (mNRF). Blocking other areas of the medulla had no obvious effect on respiratory rhythm, with the exception of the rostral portion of the ventral respiratory group (VRG), which overlaps with the mNRF. When the mNRF was blocked, most inspiratory and expiratory neurons recorded in the VRG and DRG (dorsal respiratory group) gradually started to fire continuously, and no longer exhibited respiratory rhythm. A minority of respiratory neurons was inactivated during apnea. Stimulation of the caudal portion of the DRG and VRG evoked only a short cluster of phrenic discharges instead of rhythmical firing, indicating that the respiratory neurons situated in these areas cannot generate rhythmic activity by themselves. This suggests that the mNRF plays an important role in the genesis and maintenance of basic respiratory rhythm.

We analyzed the membrane potential of 161 respiratory neurons in the medulla of decerebrate rats which were paralyzed and ventilated. Three types of inspiratory (I) neurons were observed: those displaying progressive depolarization in inspiration (augmenting I neurons), those which gradually repolarized after maximal depolarization at the onset of inspiration (decrementing I neurons) and those exhibiting a plateau or bell-shaped membrane potential trajectory throughout inspiration (I-all neurons). Three types of expiratory (E) neurons were also encountered: those in which the membrane potential progressively depolarized (augmenting E neurons), those in which the membrane potential repolarized during the interval between phrenic bursts (decrementing E or post-I neurons) and those exhibiting a plateau or bell-shaped membrane potential trajectory throughout expiration (E-all neurons). Axonal projections of these medullary neurons were identified in the cranial nerves (n = 34), or in the spinal cord (n = 19) as revealed by antidromic stimulation and/or by reconstruction following horseradish peroxidase (HRP) labeling. The other 108 neurons were not antidromically activated (NAA) by the stimulations tested, or had their axons terminating inside the medulla as revealed by HRP labeling. All these respiratory neurons, except for 3 which were hypoglossal motoneurons, had their somata within the ventrolateral medulla, in the region of the nucleus ambiguus, homologous to the ventral respiratory group (VRG) of the cat. No dorsal respiratory group (DRG) was detected within the medulla of the rats. Due to this absence of a DRG, it is concluded that the neural organization of respiratory centers is quite different in cats and rats.

The involvement of the dorsal part of the medulla (the so-called dorsal respiratory group: DRG) in the networks participating in respiratory function was investigated in newborn (in vitro) and adult (in vivo) rats. In the dorsal part of the medulla of the isolated brainstem of newborn rats, no respiratory neurons were found and stimulations or lesions neither modified nor suppressed the respiratory output. On the contrary, similar experiments suggest that sites in the ventral medulla have a fundamental importance for respiration. In adult rats, lesion of the DRG areas by electrocoagulation induced transient changes in respiratory timing, and resulted in a significant decrease in the amplitude of the contralateral phrenic output. These results suggest that the dorsal part of the medulla is not involved in controlling respiratory activity in the newborn rat. In adults, no definite conclusion can be reached, but the functional role of the DRG, if any, is probably restricted.

Unilateral focal cold blocks in the region of the nucleus tractus solitarius and the dorsal respiratory group of neurons, DRG, of anaesthetized cats consistently caused apneustic-type breathing. There was no concomitant change in the initial rate of rise of inspiratory activity. The apneustic prolongation of inspiratory duration, TI, was most pronounced in, but was not confined to, the DRG. The apneustic effects were more marked after vagotomy. In cats with intact vagus nerves being given artificial ventilation, focal cooling at certain sites of the DRG region could produce 'unlocking' of the respiratory rhythm from that of the respiratory pump. At other sites in this region, focal cooling could selectively block the effects of the inspiration-facilitating reflex induced by deflation without blocking the inspiration-inhibiting Hering-Breuer reflex. Unilateral focal cold blocks in the region of the intermediate part of the ventral respiratory group of neurons, VRG, generally caused depression of the rate of rise of inspiratory activity, but almost never apneustic effects. All effects of unilateral focal cooling both in the DRG and VRG were bilaterally symmetrical. No systematic differences between the effects on phrenic and external intercostal inspiratory activity were found in response to focal cooling either of the DRG or VRG suggesting that differential control of phrenic and external intercostal motoneurons is not exerted mainly at the level of these medullary structures. The results suggest that the DRG and VRG areas exert somewhat different effects on the respiratory pattern: DRG appears to be more concerned with integration of vagal and other inputs contributing to the inspiratory off-switch mechanisms which, however, are not confined only to the DRG. The VRG inspiratory mechanisms, on the other hand, appear to be more involved in the gain control of the inspiratory output intensity.

Brainstem neurones which project to the immediate vicinity of the spinal motoneurones which supply the intercostal and abdominal respiratory muscles were identified by means of the retrograde transport of horseradish peroxidase (HRP). A combined electrophysiological and histological technique was used in which recording of phasic inspiratory or expiratory motoneurone activity within upper (T3-T4) or lower (T8-T9) thoracic segments was followed by the iontophoretic injection of HRP at these recording sites. HRP labelled cells were concentrated in those brainstem regions known to contain phasic respiratory neurones, namely the ventrolateral nucleus of the solitary tract (vl-NTS) or dorsal respiratory group (DRG), the ambiguus complex or ventral respiratory group (VRG) and the parabrachial pontine (PB) nuclei. In 18 cats, 248 cells were labelled in these three respiratory regions of the brainstem while 668 were much more diffusely distributed in other regions of the medulla and pons. The ipsilateral and contralateral contributions within the respiratory regions were respectively; 23%:77% (DRG), 33%:67% (VRG), 95%:5% (PB). These results are considered in the general context of previous electrophysiological and histological findings, but also with particular reference to a related study of the projections from brainstem neurones to the phrenic nucleus [32].

The effects of intravenous administration of fentanyl (50 and 100 micrograms/kg) on the discharge activity of the medullary inspiratory neurons and of the phrenic nerve were studied following vagotomy in nine decerebrate, paralyzed mechanically ventilated cats. In six cats, the inspiratory neurons explored were in the dorsal respiratory group (DRG) associated with the nucleus of the tractus solitarius (NTS), while in the remaining three, they were in the ventral respiratory group (VRG). In the former group, the rhythmic discharge of the inspiratory neurons was disrupted by fentanyl and replaced by a continuous discharge superimposed with irregularly occurring bursts. These changes were also reflected by the phrenic nerve discharge. Inspiratory neuronal activity increased significantly (P less than 0.05) at 1 and 5 min after completion of fentanyl injection. Disruption of the rhythmic activity of the inspiratory neurons and its replacement by a continuous and irregular discharge may lead to sustained contraction of inspiratory muscles and cessation of respiration. In the VRG, the activity of the inspiratory neurons was totally abolished by fentanyl. Thus, it appears that different groups of medullary inspiratory neurons have differential sensitivity to fentanyl. Nalbuphine, an opiate agonist-antagonist, restored the normal pattern and magnitude of the activity of the inspiratory neurons.