A novel machine learning system for identifying sleep-wake states in mice

Research into sleep-wake behaviors relies on scoring sleep states, normally done by manual inspection of electroencephalogram (EEG) and electromyogram (EMG) recordings. This is a highly time-consuming process prone to inter-rater variability. When studying relationships between sleep and motor function, analyzing arousal states under a four-state system of active wake (AW), quiet wake (QW), nonrapid-eye-movement (NREM) sleep, and rapid-eye-movement (REM) sleep provides greater precision in behavioral analysis but is a more complex model for classification than the traditional three-state identification (wake, NREM, and REM sleep) usually used in rodent models. Characteristic features between sleep-wake states provide potential for the use of machine learning to automate classification. Here, we devised SleepEns, which uses a novel ensemble architecture, the time-series ensemble. SleepEns achieved 90% accuracy to the source expert, which was statistically similar to the performance of two other human experts. Considering the capacity for classification disagreements that are still physiologically reasonable, SleepEns had an acceptable performance of 99% accuracy, as determined blindly by the source expert. Classifications given by SleepEns also maintained similar sleep-wake characteristics compared to expert classifications, some of which were essential for sleep-wake identification. Hence, our approach achieves results comparable to human ability in a fraction of the time. This new machine-learning ensemble will significantly impact the ability of sleep researcher to detect and study sleep-wake behaviors in mice and potentially in humans.

Immortal orexin cell transplants restore motor-arousal synchrony during cataplexy

Waking behaviors such as sitting or standing require suitable levels of muscle tone. But it is unclear how arousal and motor circuits communicate with one another so that appropriate motor tone occurs during wakefulness. Cataplexy is a peculiar condition in which muscle tone is involuntarily lost during normal periods of wakefulness. Cataplexy therefore provides a unique opportunity for identifying the signaling mechanisms that synchronize motor and arousal behaviors. Cataplexy occurs when hypothalamic orexin neurons are lost in narcolepsy; however, it is unclear if motor-arousal decoupling in cataplexy is directly or indirectly caused by orexin cell loss. Here, we used genomic, proteomic, chemogenetic, electrophysiological, and behavioral assays to determine if grafting orexin cells into the brain of cataplectic (i.e., orexin^-/-) mice restores normal motor-arousal behaviors by preventing cataplexy. First, we engineered immortalized orexin cells and found that they not only produce and release orexin but also exhibit a gene profile that mimics native orexin neurons. Second, we show that engineered orexin cells thrive and integrate into host tissue when transplanted into the brain of mice. Next, we found that grafting only 200-300 orexin cells into the dorsal raphe nucleus-a region densely innervated by native orexin neurons-reduces cataplexy. Last, we show that real-time chemogenetic activation of orexin cells restores motor-arousal synchrony by preventing cataplexy. We suggest that orexin signaling is critical for arousal-motor synchrony during wakefulness and that the dorsal raphe plays a pivotal role in coupling arousal and motor behaviors.

Neurobiological and immunogenetic aspects of narcolepsy: Implications for pharmacotherapy

Excessive daytime sleepiness (EDS) and cataplexy are common symptoms of narcolepsy, a sleep disorder associated with the loss of hypocretin/orexin (Hcrt) neurons. Although only a few drugs have received regulatory approval for narcolepsy to date, treatment involves diverse medications that affect multiple biochemical targets and neural circuits. Clinical trials have demonstrated efficacy for the following classes of drugs as narcolepsy treatments: alerting medications (amphetamine, methylphenidate, modafinil/armodafinil, solriamfetol [JZP-110]), antidepressants (tricyclic antidepressants, selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors), sodium oxybate, and the H₃-receptor inverse agonist/antagonist pitolisant. Enhanced catecholamine availability and regulation of locus coeruleus (LC) norepinephrine (NE) neuron activity is likely central to the therapeutic activity of most of these compounds. LC NE neurons are integral to sleep/wake regulation and muscle tone; reduced excitatory input to the LC due to compromise of Hcrt/orexin neurons (likely due to autoimmune factors) results in LC NE dysregulation and contributes to narcolepsy/cataplexy symptoms. Agents that increase catecholamines and/or LC activity may mitigate EDS and cataplexy by elevating NE regulation of GABAergic inputs from the amygdala. Consequently, novel medications and treatment strategies aimed at preserving and/or modulating Hcrt/orexin-LC circuit integrity are warranted in narcolepsy/cataplexy.

Brain Circuitry Controlling Sleep and Wakefulness

PURPOSE OF REVIEW

This article outlines the fundamental brain mechanisms that control sleep-wake patterns and reviews how pathologic changes in these control mechanisms contribute to common sleep disorders.

RECENT FINDINGS

Discrete but interconnected clusters of cells located within the brainstem and hypothalamus comprise the circuits that generate wakefulness, non-rapid eye movement (non-REM) sleep, and REM sleep. These clusters of cells use specific neurotransmitters, or collections of neurotransmitters, to inhibit or excite their respective sleep- and wake-promoting target sites. These excitatory and inhibitory connections modulate not only the presence of wakefulness or sleep, but also the levels of arousal within those states, including the depth of sleep, degree of vigilance, and motor activity. Dysfunction or degeneration of wake- and sleep-promoting circuits is associated with narcolepsy, REM sleep behavior disorder, and age-related sleep disturbances.

SUMMARY

Research has made significant headway in identifying the brain circuits that control wakefulness, non-REM, and REM sleep and has led to a deeper understanding of common sleep disorders and disturbances.

New Neuroscience Tools That Are Identifying the Sleep-Wake Circuit

The complexity of the brain is yielding to technology. In the area of sleep neurobiology, conventional neuroscience tools such as lesions, cell recordings, c-Fos, and axon-tracing methodologies have been instrumental in identifying the complex and intermingled populations of sleep- and arousal-promoting neurons that orchestrate and generate wakefulness, NREM, and REM sleep. In the last decade, new technologies such as optogenetics, chemogenetics, and the CRISPR-Cas system have begun to transform how biologists understand the finer details associated with sleep-wake regulation. These additions to the neuroscience toolkit are helping to identify how discrete populations of brain cells function to trigger and shape the timing and transition into and out of different sleep-wake states, and how glia partner with neurons to regulate sleep. Here, we detail how some of the newest technologies are being applied to understand the neural circuits underlying sleep and wake.

Monoamine Release during Unihemispheric Sleep and Unihemispheric Waking in the Fur Seal

STUDY OBJECTIVES

Our understanding of the role of neurotransmitters in the control of the electroencephalogram (EEG) has been entirely based on studies of animals with bilateral sleep. The study of animals with unihemispheric sleep presents the opportunity of separating the neurochemical substrates of waking and sleep EEG from the systemic, bilateral correlates of sleep and waking states.

METHODS

The release of histamine (HI), norepinephrine (NE), and serotonin (5HT) in cortical and subcortical areas (hypothalamus, thalamus and caudate nucleus) was measured in unrestrained northern fur seals (Callorhinus ursinus) using in vivo microdialysis, in combination with, polygraphic recording of EEG, electrooculogram, and neck electromyogram.

RESULTS

The pattern of cortical and subcortical HI, NE, and 5HT release in fur seals is similar during bilaterally symmetrical states: highest in active waking, reduced in quiet waking and bilateral slow wave sleep, and lowest in rapid eye movement (REM) sleep. Cortical and subcortical HI, NE, and 5HT release in seals is highly elevated during certain waking stimuli and behaviors, such as being sprayed with water and feeding. However, in contrast to acetylcholine (ACh), which we have previously studied, the release of HI, NE, and 5HT during unihemispheric sleep is not lateralized in the fur seal.

CONCLUSIONS

Among the studied neurotransmitters most strongly implicated in waking control, only ACh release is asymmetric in unihemispheric sleep and waking, being greatly increased on the activated side of the brain.

COMMENTARY

A commentary on this article appears in this issue on page 491.

REM Sleep at its Core - Circuits, Neurotransmitters, and Pathophysiology

Rapid eye movement (REM) sleep is generated and maintained by the interaction of a variety of neurotransmitter systems in the brainstem, forebrain, and hypothalamus. Within these circuits lies a core region that is active during REM sleep, known as the subcoeruleus nucleus (SubC) or sublaterodorsal nucleus. It is hypothesized that glutamatergic SubC neurons regulate REM sleep and its defining features such as muscle paralysis and cortical activation. REM sleep paralysis is initiated when glutamatergic SubC cells activate neurons in the ventral medial medulla, which causes release of GABA and glycine onto skeletal motoneurons. REM sleep timing is controlled by activity of GABAergic neurons in the ventrolateral periaqueductal gray and dorsal paragigantocellular reticular nucleus as well as melanin-concentrating hormone neurons in the hypothalamus and cholinergic cells in the laterodorsal and pedunculo-pontine tegmentum in the brainstem. Determining how these circuits interact with the SubC is important because breakdown in their communication is hypothesized to underlie narcolepsy/cataplexy and REM sleep behavior disorder (RBD). This review synthesizes our current understanding of mechanisms generating healthy REM sleep and how dysfunction of these circuits contributes to common REM sleep disorders such as cataplexy/narcolepsy and RBD.

Noradrenergic modulation of masseter muscle activity during natural rapid eye movement sleep requires glutamatergic signalling at the trigeminal motor nucleus

Noradrenergic neurotransmission in the brainstem is closely coupled to changes in muscle activity across the sleep-wake cycle, and noradrenaline is considered to be a key excitatory neuromodulator that reinforces the arousal-related stimulus on motoneurons to drive movement. However, it is unknown if α-1 noradrenoceptor activation increases motoneuron responsiveness to excitatory glutamate (AMPA) receptor-mediated inputs during natural behaviour. We studied the effects of noradrenaline on AMPA receptor-mediated motor activity at the motoneuron level in freely behaving rats, particularly during rapid eye movement (REM) sleep, a period during which both AMPA receptor-triggered muscle twitches and periods of muscle quiescence in which AMPA drive is silent are exhibited. Male rats were subjected to electromyography and electroencephalography recording to monitor sleep and waking behaviour. The implantation of a cannula into the trigeminal motor nucleus of the brainstem allowed us to perfuse noradrenergic and glutamatergic drugs by reverse microdialysis, and thus to use masseter muscle activity as an index of motoneuronal output. We found that endogenous excitation of both α-1 noradrenoceptor and AMPA receptors during waking are coupled to motor activity; however, REM sleep exhibits an absence of endogenous α-1 noradrenoceptor activity. Importantly, exogenous α-1 noradrenoceptor stimulation cannot reverse the muscle twitch suppression induced by AMPA receptor blockade and nor can it elevate muscle activity during quiet REM, a phase when endogenous AMPA receptor activity is subthreshold. We conclude that the presence of an endogenous glutamatergic drive is necessary for noradrenaline to trigger muscle activity at the level of the motoneuron in an animal behaving naturally.

Brain biology: jerked around by sleep

During rapid eye movement sleep, the forelimb muscles of newborn rats jerk and twitch in an organized pattern, the fidelity of which improves with time. The coordinated nature of such sleep movements may instruct the developing brain how to more effectively execute movements during wakefulness.

Dopamine triggers skeletal muscle tone by activating D1-like receptors on somatic motoneurons

The dopamine system plays an integral role in motor physiology. Dopamine controls movement by modulation of higher-order motor centers (e.g., basal ganglia) but may also regulate movement by directly controlling motoneuron function. Even though dopamine cells synapse onto motoneurons, which themselves express dopamine receptors, it is unknown whether dopamine modulates skeletal muscle activity. Therefore, we aimed to determine whether changes in dopaminergic neurotransmission at a somatic motor pool affect motor outflow to skeletal muscles. We used microinjection, neuropharmacology, electrophysiology, and histology to determine whether manipulation of D(1)- and D(2)-like receptors on trigeminal motoneurons affects masseter and/or tensor palatini muscle tone in anesthetized rats. We found that apomorphine (a dopamine analog) activated trigeminal motoneurons and triggered a potent increase in both masseter and tensor palatini tone. This excitatory effect is mediated by D(1)-like receptors because specific D(1)-like receptor activation strengthened muscle tone and blockade of these receptors prevented dopamine-driven activation of motoneurons. Blockade of D(1)-like receptors alone had no detectable effect on basal masseter/tensor palatini tone, indicating the absence of a functional dopamine drive onto trigeminal motoneurons, at least during isoflurane anesthesia. Finally, we showed that D(2)-like receptors do not affect either trigeminal motoneuron function or masseter/tensor palatini muscle tone. Our results provide the first demonstration that dopamine can directly control movement by manipulating somatic motoneuron behavior and skeletal muscle tone.

Dopaminergic regulation of sleep and cataplexy in a murine model of narcolepsy

STUDY OBJECTIVES

To determine if the dopaminergic system modulates cataplexy, sleep attacks and sleep-wake behavior in narcoleptic mice.

DESIGN

Hypocretin/orexin knockout (i.e., narcoleptic) and wild-type mice were administered amphetamine and specific dopamine receptor modulators to determine their effects on sleep, cataplexy and sleep attacks.

PATIENTS OR PARTICIPANTS

Hypocretin knockout (n = 17) and wild-type mice (n = 21).

INTERVENTIONS

Cataplexy, sleep attacks and sleep-wake behavior were identified using electroencephalogram, electromyogram and videography. These behaviors were monitored for 4 hours after an i.p. injection of saline, amphetamine and specific dopamine receptor modulators (D1- and D2-like receptor modulators).

MEASUREMENTS AND RESULTS

Amphetamine (2 mg/kg), which increases brain dopamine levels, decreased sleep attacks and cataplexy by 61% and 67%, suggesting that dopamine transmission modulates such behaviors. Dopamine receptor modulation also had powerful effects on sleep attacks and cataplexy. Activation (SKF 38393; 20 mg/kg) and blockade (SCH 23390; 1 mg/kg) of D1-like receptors decreased and increased sleep attacks by 77% and 88%, without affecting cataplexy. Pharmacological activation of D2-like receptors (quinpirole; 0.5 mg/kg) increased cataplectic attacks by 172% and blockade of these receptors (eticlopride; 1 mg/kg) potently suppressed them by 97%. Manipulation of D2-like receptors did not affect sleep attacks.

CONCLUSIONS

We show that the dopaminergic system plays a role in regulating both cataplexy and sleep attacks in narcoleptic mice. We found that cataplexy is modulated by a D2-like receptor mechanism, whereas dopamine modulates sleep attacks by a D1-like receptor mechanism. These results support a role for the dopamine system in regulating sleep attacks and cataplexy in a murine model of narcolepsy.

Noradrenaline triggers muscle tone by amplifying glutamate-driven excitation of somatic motoneurones in anaesthetized rats

Postural muscle tone is potently suppressed during sleep and cataplexy. Since brainstem noradrenergic cell discharge activity is tightly coupled with state-dependent changes in muscle activity, it is assumed that noradrenergic drive on to somatic motoneurones modulates basal muscle tone. However, it has never been determined whether noradrenergic neurotransmission acts to directly regulate motoneurone activity or whether it functions to modulate prevailing synaptic activity. This is an important distinction because noradrenaline regulates cell excitability by both directly depolarizing neurones and by indirectly potentiating glutamate-mediated excitation. We used reverse-microdialysis, electrophysiology, neuro-pharmacological and histological techniques in anaesthetized rats to determine whether strengthening noradrenergic drive (via exogenous noradrenaline application) on to trigeminal motoneurones affects masseter muscle tone by increasing spontaneous motoneurone activity or whether it acts to amplify prevailing glutamate-driven excitation. Although noradrenaline is hypothesized to modulate motor activity, we found that direct stimulation of trigeminal motoneurones by alpha(1)-adrenoceptor activation had no direct effect on basal masseter tone. However, when glutamate-driven excitation was increased at the trigeminal motor pool by either endogenous glutamate release (induced by the monosynaptic masseteric reflex) or exogenous AMPA application, noradrenaline triggered a potent increase in basal masseter tone. The stimulatory effects of noradrenaline were unmasked and rapidly switched on only in the presence of glutamatergic transmission. Blockade of AMPA receptors abolished this excitatory effect, indicating that noradrenergic drive requires ongoing glutamatergic activity. Our data indicate that exogenous noradrenergic drive does not directly affect spontaneous motoneurone discharge activity in anaesthetized rats; rather, it triggers postural muscle tone by amplifying prevailing glutamate-driven excitation.

Connections between respiratory neurones in the neonatal rat transverse medullary slice studied with cross-correlation

In the transverse medullary slice prepared from neonatal rats the hypoglossal nerve rootlets exhibit a bursting 'respiratory' rhythm as do neurones in the pre-Bötzinger complex (PBC). We used cross-correlation analysis of the rhythmic multiunit discharges recorded from hypoglossal nerve rootlets, hypoglossal nucleus neurones and PBC neurones to investigate the connections between these groups. All cross-correlograms computed between left and right hypoglossal nerves, and between hypoglossal neurones and contralateral hypoglossal nerves, displayed central peaks with broad half-amplitude widths (mean +/- S.D. of 29.6 +/- 10.4 and 37.3 +/- 6.0 ms, respectively), which we interpreted as evidence for activation from a common source. Five of the 18 cross-correlograms computed between left and right PBC neurones displayed peaks either side of time zero with narrower half-amplitude widths (mean +/- S.D. of 9.3 +/- 1.9 ms) superimposed on broader central peaks, which we interpreted as evidence for mutual excitation and common activation, respectively. Cross-correlograms computed between PBC neurones and contralateral hypoglossal neurones or nerves did not display consistent features, but some of those computed between PBC and ipsilateral hypoglossal neurones (two of eight) or nerves (two of five) displayed peaks with broad half-amplitude widths (mean +/- S.D. of 36.8 +/- 6.9 ms), offset from time zero by 6 ms (except for one at 18 ms), which we interpreted as evidence for excitation of hypoglossal neurones and motoneurones by PBC neurones. We concluded that rhythm is synchronised between left and right sides by mutual excitatory connections between left and right PBC neurones. The rhythm is transmitted to ipsilateral hypoglossal neurones by a paucisynaptic pathway. Both hypoglossal neurones and PBC neurones receive a common activation from as yet unidentified sources.

Excitatory effects of hypocretin-1 (orexin-A) in the trigeminal motor nucleus are reversed by NMDA antagonism

Hypocretin-1 and -2 (Hcrt-1 and -2, also called orexin-A and -B) are newly identified neuropeptides synthesized by hypothalamic neurons. Defects in the Hcrt system underlie the sleep disorder narcolepsy, which is characterized by sleep fragmentation and the involuntary loss of muscle tone called cataplexy. Hcrt neurons project to multiple brain regions including cranial and spinal motor nuclei. In vitro studies suggest that Hcrt application can modulate presynaptic glutamate release. Together these observations suggest that Hcrt can affect motor output and that glutamatergic processes may be involved. We addressed these issues in decerebrate cats by applying Hcrt-1 and -2 into the trigeminal motor nucleus to determine whether these ligands alter masseter muscle activity and by pretreating the trigeminal motor nucleus with a N-methyl-d-aspartate (NMDA) antagonist to determine if glutamatergic pathways are involved in the transduction of the Hcrt signal. We found that Hcrt-1 and -2 microinjections into the trigeminal motor nucleus increased ipsilateral masseter muscle tone in a dose-dependent manner. We also found that Hcrt application into the hypoglossal motor nucleus increases genioglossus muscle activity. Pretreatment with a NMDA antagonist (d-(-)-2-amino-phosphonovaleric acid) abolished the excitatory response of the masseter muscle to Hcrt-1 application; however, pretreatment with methysergide, a serotonin antagonist had no effect. These studies are the first to demonstrate that Hcrt causes the excitation of motoneurons and that functional NMDA receptors are required for this response. We suggest that Hcrt regulates motor control processes and that this regulation is mediated by glutamate release in the trigeminal motor nucleus.

Excitatory effects of hypocretin-1 (orexin-A) in the trigeminal motor nucleus are reversed by NMDA antagonism

Hypocretin-1 and -2 (Hcrt-1 and -2, also called orexin-A and -B) are newly identified neuropeptides synthesized by hypothalamic neurons. Defects in the Hcrt system underlie the sleep disorder narcolepsy, which is characterized by sleep fragmentation and the involuntary loss of muscle tone called cataplexy. Hcrt neurons project to multiple brain regions including cranial and spinal motor nuclei. In vitro studies suggest that Hcrt application can modulate presynaptic glutamate release. Together these observations suggest that Hcrt can affect motor output and that glutamatergic processes may be involved. We addressed these issues in decerebrate cats by applying Hcrt-1 and -2 into the trigeminal motor nucleus to determine whether these ligands alter masseter muscle activity and by pretreating the trigeminal motor nucleus with a N-methyl-d-aspartate (NMDA) antagonist to determine if glutamatergic pathways are involved in the transduction of the Hcrt signal. We found that Hcrt-1 and -2 microinjections into the trigeminal motor nucleus increased ipsilateral masseter muscle tone in a dose-dependent manner. We also found that Hcrt application into the hypoglossal motor nucleus increases genioglossus muscle activity. Pretreatment with a NMDA antagonist (d-(-)-2-amino-phosphonovaleric acid) abolished the excitatory response of the masseter muscle to Hcrt-1 application; however, pretreatment with methysergide, a serotonin antagonist had no effect. These studies are the first to demonstrate that Hcrt causes the excitation of motoneurons and that functional NMDA receptors are required for this response. We suggest that Hcrt regulates motor control processes and that this regulation is mediated by glutamate release in the trigeminal motor nucleus.

Acetazolamide and respiratory chemosensitivity to CO(2) in the neonatal rat transverse medullary slice

Hypoglossal nerve rootlets in the transverse medullary slice prepared from neonatal rats exhibit a bursting 'respiratory' rhythm that increases in frequency with CO(2), presumably due to activation of chemosensitive cells such as the central chemoreceptors. Carbonic anhydrase is associated with areas of central chemoreception and we propose a hypothesis for its involvement in the chemoreception process. We tested this hypothesis by blocking its activity with acetazolamide in six slice preparations. However, the addition of 1 mM acetazolamide dissolved in dimethyl sulphoxide to the superfusing bathing solution produced no alteration in the bursting frequency response of the slice to CO(2). We concluded that the chemoreception process producing the CO(2) response of the superfused, transverse medullary slice does not involve carbonic anhydrase.

Respiratory pre-motor control of hypoglossal motoneurons in the rat

The goal of this study was to determine the origin and transmission pathway of respiratory drive to hypoglossal motoneurons. First we recorded intracellularly from 28 antidromically activated inspiratory hypoglossal motoneurons (resting membrane potential, -50+/-3 mV), and found that injection of chloride ions had no discernible effect on the shape of their membrane potential trajectories. We concluded that the membrane potential trajectories of these hypoglossal motoneurons were determined primarily by inspiratory excitation. To determine the origin of this excitation we cross-correlated the extracellular discharge of medullary inspiratory neurons, including those in the hypoglossal motor nucleus, with the hypoglossal nerve discharge. We found 27 inspiratory neurons within the hypoglossal motor nucleus that were not antidromically activated from the ipsilateral hypoglossal nerve; their cross-correlograms featured either central peaks (1.7+/-0.2 ms) alone (n=14; 39%), or central peaks (1.3+/-0.2 ms) followed by troughs (1.3+/-0.1 ms) at short latencies (1.1+/-0.4 ms) (n=13; 36%), and suggest that these neurons are hypoglossal interneurons. We recorded from 238 inspiratory neurons throughout the rest of the medulla; the cross-correlograms of 19 neurons (8%), located mostly in the lateral tegmental field, displayed narrow half-amplitude peaks (1.0+/-0.1 ms) at short latencies (0.9+/-0.1 ms), which we interpreted as evidence for monosynaptic excitation of hypoglossal motoneurons.We conclude that the respiratory control of hypoglossal motoneurons originates from inspiratory premotor neurons scattered throughout the lateral tegmental field and interneurons within the hypoglossal motor nucleus.

Bilateral coordination of inspiratory neurones in the rat

Inspiratory activity on the left and right sides must be coordinated to be effective. We used cross-correlation to examine the hypothesis that the coordination of left and right medullary inspiratory neurones is produced by excitation from common sources and by midline-crossing excitatory connections among these neurones. In adult rats, a total of 185 contralateral pairs of inspiratory neurones ( n=370) were recorded extracellularly, and classified, according to their firing pattern, as augmenting ( n=262), constant ( n=82) or decrementing ( n=26). Of the 262 augmenting inspiratory neurones, 98 were classified as phrenic premotor neurones by cross-correlation with phrenic nerve discharge. The 185 cross-correlograms showed little evidence of common activation, or midline-crossing excitatory connections. Of the 45 cross-correlograms for pairs of augmenting neurones, only 4 (approximately equal to 9%) indicated a common activation, and only one a monosynaptic connection. Of the 45 for pairs of augmenting and phrenic premotor neurones, only 9 (20%) showed a common activation, and only 2 a monosynaptic excitatory connection. Of the 19 pairs of phrenic premotor neurones, 5 from the same rat showed high-frequency oscillations, and 1 a monosynaptic excitatory connection. Cross-correlograms for pair combinations of other types of neurones also exhibited few features. We suggest that, in the adult rat, although both common activation and excitatory cross-connections exist as a means for coordinating left and right ventral group inspiratory neurones to the same respiratory rhythm, they are insufficient to account for it.

Respiratory control of hypoglossal motoneurones in the rat

In this study of adult and neonatal rats, we used cross-correlation analysis to detect synchronous neuronal events in hypoglossal and phrenic nerves to infer synaptic connections. We found evidence for the common excitation of medial and lateral hypoglossal motoneurones in 12 anaesthetized adult rats but not in 6 in vitro brainstem-spinal cord preparations. We did not find evidence for the common activation of phrenic and hypoglossal motoneurones in 23 adult and 10 neonatal rat preparations. We confirmed this negative result by demonstrating that 26 medullary inspiratory neurones activating phrenic motoneurones did not activate hypoglossal motoneurones in 23 adult decerebrate rats (except in one case). We also found that 15 Bötzinger expiratory neurones inhibiting phrenic motoneurones did not inhibit hypoglossal motoneurones. We conclude that: (1) motoneurones of the medial and lateral hypoglossal nerve branches receive inspiratory drive from a common premotor population in adult rats, but in neonatal rats adjacent nerve rootlets do not; (2) in both adult and neonatal rats phrenic premotor neurones do not monosynaptically excite hypoglossal motoneurones; (3) Bötzinger expiratory neurones that inhibit phrenic motoneurones do not inhibit hypoglossal motoneurones. We therefore suggest that the respiratory control of hypoglossal motoneurones is separate from that of phrenic motoneurones.

Bilateral synchronisation of respiratory motor output in rats: adult versus neonatal in vitro preparations

The synchronisation of the discharges recorded from left and right phrenic nerves in the adult rat is produced in part by shared excitation from a common premotor neurone population. However, such synchronisation has not been examined for hypoglossal motoneurones in adult rats, or for phrenic and hypoglossal motoneurons in neonatal in vitro preparations. In adult rats, cross-correlograms computed between the inspiratory discharges of the left and right phrenic nerves, and the left and right hypoglossal nerves displayed central peaks with half-amplitude widths of 1.4+/-0.1 and 1.7+/-0.1 ms (mean+/-SE), respectively. We interpret these as evidence for common excitation. However, such central peaks were absent in the same cross-correlograms computed for neonatal in vitro preparations, although central peaks were observed in cross-correlograms computed between the discharges recorded from adjacent phrenic nerve rootlets. We conclude that, in the adult rat, left and right hypoglossal nerve discharges are synchronised by excitation from a common premotor neurone population, as for the phrenic nerves, but this type of synchronisation is undetectable in neonatal in vitro preparations. We speculate that the differences between the adult and neonatal preparations are due to developmental changes in respiratory drive transmission pathways.

Nucleus raphé obscurus modulates hypoglossal output of neonatal rat in vitro transverse brain stem slices

Nucleus raphé obscurus (NRo) modulates hypoglossal (XII) nerve motor output in the in vitro transverse brain stem slice of neonatal rats (1-5 days old); chemical ablation of NRo and its focal CO(2) acidification modulated the bursting rhythm of XII nerves. We microinjected a 4.5 mM solution of kainic acid into the NRo to disrupt cellular activity and observed that XII nerve activity was temporarily abolished (n = 10). We also microinjected CO(2)-acidified (pH = 6.00 +/- 0.01) artificial cerebrospinal fluid (aCSF) into the NRo (n = 6), the pre-Bötzinger complex (PBC) (n = 6), as well as a control region in the lateral tegmental field equidistant to NRo, PBC, and the XII motor nuclei (n = 12). CO(2) acidification of the control region had no effect on XII motor output. CO(2) acidification of the NRo significantly (P < 0.05) increased the burst discharge frequency of XII nerves by 77%; integrated burst amplitude and burst duration increased by 64% and 52%, respectively. CO(2) acidification of the PBC significantly (P < 0.05) increased the burst discharge frequency of XII nerves by 65%, but neither integrated burst amplitude nor burst duration changed. These results demonstrate that chemical ablation of the NRo can abolish XII nerve bursting rhythm and that stimulation of the NRo with CO(2)-acidified aCSF can excite XII nerve bursting activity. From these observations, we conclude that, in transverse brain stem slices, the NRo contains pH/CO(2)-sensitive cells that modulate XII motor output.

Functional synaptic connections among respiratory neurons

This presentation focuses on the application of methods to determine functional connections between neurons in the respiratory network of adult decerebrate rats. We employ a general network investigation paradigm that first examines the intracellular recordings of a respiratory neuron and then determines which neurons synapse with it to produce the observed membrane potential changes. It is used to pursue the source of respiratory excitation and inhibition from its arrival at phrenic motoneurons to respiratory neurons in the medulla, and then examine some of the interactions among these neurons that shape their patterns of activity. Findings include a demonstration that phrenic motoneuron activity is determined by excitation from medullary inspiratory premotor neurons and inhibition by Bötzinger complex expiratory neurons, and that the latter neurons inhibit both medullary inspiratory premotor neurons and themselves. We conclude that these functional interconnections explain the activity patterns of some respiratory neurons, but the connections between neurons thought to be involved in rhythm generation remain to be demonstrated in adult rats.

Temperature and pH affect respiratory rhythm of in-vitro preparations from neonatal rats

We examined the respiratory rhythm of two in-vitro preparations from neonatal rats, the brainstem-spinal cord and transverse brainstem slice, recording the bursting activity of phrenic and hypoglossal nerves, respectively at 1 degree C intervals from 25 to 35 degrees C at two pH's, 7.4 and 7.1. In both preparations at either pH, burst frequency increased with temperature, burst duration declined and burst amplitude reached a peak at 30 degrees C. The shapes of the bursts changed from a decrementing pattern at low temperatures to a bell-shaped pattern at high temperatures. At reduced pH, frequency increased for temperatures between 25 and 32 degrees C in the brainstem-spinal cord but not in the slice. Burst duration was increased at reduced pH for temperatures between 27 and 29 degrees C in the brainstem-spinal cord, but not in the slice. Burst amplitude only changed with pH at the lower temperatures, decreasing at the lower pH in the brainstem-spinal cord and increasing in the slice. With respect to the effects of temperature, we concluded that both preparations were similarly affected, and that an increase in temperature alters the in-vitro burst pattern towards that observed in-vivo. With respect to the effects of pH, we concluded that effects differ between these in-vitro preparations and from in-vivo preparations, and that the difference between in-vivo and in-vitro preparations in their response to decreasing pH is not due to differences in temperature.

Mutual inhibition between Bötzinger-complex bulbospinal expiratory neurons detected with cross-correlation in the decerebrate rat

We examined the synaptic connections between pairs of Bötzinger-complex, bulbospinal expiratory neurons in decerebrate rats. All were antidromically activated from the spinal cord at the C2-C3 border. Cross-correlation histograms of 18 ipsilateral pairs showed troughs on both sides of time zero (8) and to one side of time zero (4); most (12) were accompanied by peaks at time zero. Similarly, cross-correlation histograms of the contralateral pairs (12) showed troughs on both sides of time zero (3) and to one side of time zero (3); few (2) were accompanied by peaks at time zero. We considered the troughs in these cross-correlation histograms to be evidence of inhibition between the neurons and sought confirmation of the inhibitory connection. First, using the antidromic activation stimulus, we computed post-stimulus histograms of the extracellularly recorded discharge for six neurons and found that three showed troughs. Then, we continued this approach, computing post-stimulus averages of the membrane potentials recorded intracellularly from these neurons after iontophoresis of chloride to reverse inhibitory synaptic potentials. Depolarising potentials were observed in 15 of 16 of these averages. We interpreted these as reversed inhibitory post-synaptic potentials and concluded that Bötzinger-complex, bulbospinal expiratory neurons inhibit one another in rats as they do in cats.

Synaptic connections to phrenic motoneurons in the decerebrate rat

Phrenic motoneuron membrane potential trajectories in decerebrate rats exhibit three stages; depolarisation during inspiration, a decreased depolarisation during early expiration and hyperpolarization during late expiration. These trajectories are a result of excitation by ventral-group medullary inspiratory neurons and upper-cervical inspiratory neurons during inspiration and the early part of expiration, and inhibition from Bötzinger-complex expiratory neurons during the late part of expiration.

Bötzinger-complex, bulbospinal expiratory neurones monosynaptically inhibit ventral-group respiratory neurones in the decerebrate rat

Extracellularly recorded action potentials from 49 Bötzinger-complex, bulbospinal expiratory neurones were used as triggers to compute 162 spike-triggered averages (STAs) of intracellular potentials recorded from 167 respiratory neurones in the ventral respiratory group (VRG) near the obex in 15 vagotomized, paralysed, ventilated and decerebrated rats. All of the Bötzinger-complex expiratory neurones were antidromically activated from the ipsilateral border between the C2/C3 segments of the spinal cord and discharged only during the late part of expiration with an augmenting pattern. We found evidence for monosynaptic inhibitory post-synaptic potentials (IPSPs) in 74 (approximately 44%) of the STAs computed using 34 (approximately 69%) of the trigger neurones. For vagal motoneurones, IPSPs were found in 24 of the 53 STAs of expiratory motoneurones, but in none of the 12 STAs of inspiratory motoneurones. For inspiratory neurones, IPSPs were found in 23 of the 33 STAs of bulbospinal neurones and in 6 of the 26 STAs of not antidromically activated (NAA) neurones. For expiratory neurones, IPSPs were found in one of the two STAs of bulbospinal neurones and in 20 of the 36 STAs of NAA neurones. We conclude that Bötzinger-complex, bulbospinal expiratory neurones monosynaptically inhibit bulbospinal inspiratory neurones, expiratory vagal motoneurones and other unidentified inspiratory and expiratory neurones in the VRG of rats during the late part of expiration.

Bötzinger-complex expiratory neurons monosynaptically inhibit phrenic motoneurons in the decerebrate rat

We examined respiratory neurons in the Bötzinger complex of the medulla oblongata in 18 vagotomized, paralyzed, ventilated, and decerebrated rats and tested the hypothesis that bulbospinal expiratory neurons in this region monosynaptically inhibit phrenic motoneurons. First, we surveyed the types of respiratory neurons found in the Bötzinger complex; only 11 of the 98 (approximately 11%) examined were bulbospinal, and all discharged only during late expiration (E2), usually with an augmenting discharge frequency (AUG). Then, we examined the spinal projections of 34 E2-AUG neurons using antidromic activation and found that all projected as far as the C4 or C5 segments of the spinal cord but no further caudally. Most (30, approximately 88%) had only unilateral projections, the majority (25, approximately 83%) ipsilateral, but 4 neurons (approximately 12%) had bilateral projections. Their axons could be antidromically activated at low currents (less than 10 microA) in the dorsal-lateral part of the spinal cord at the C2-3 border; 0.5-1.2 mm (mean+/-SD 0.84+/-0.23 mm) below the dorsal surface and 0.7-1.5 mm (1.19+/-0.25 mm) lateral from the midline. We sought evidence for connections from bulbospinal E2-AUG neurons to 118 phrenic motoneurons by computing spike-triggered averages (STAs) of their intracellular potentials triggered by the action potentials of 38 unilaterally-projecting E2-AUG neurons. Resting phrenic motoneuron membrane potentials ranged from -40 to -75 mV (-56+/-8 mV) and fluctuations with the respiratory cycle from 7 to 20 mV (14+/-4 mV). Of the 118 STAs computed, hyperpolarizations were evident in 18 (approximately 15%) STAs, evoked by 11 of 38 (approximately 29%) E2-AUG neurons. Their amplitudes varied from 35 to 550 microV (105+/-113 microV), 10-90% fall times from 0.4 to 0.9 ms (0.63+/-0.17 ms), and half-amplitude widths from 1.3 to 3.2 ms (2.0+/-0.52 ms). Most (16/95, approximately 17%) of the STAs that displayed hyperpolarizations were associated with ipsilateral trigger neurons but some (2/23, approximately 9%) resulted from contralateral trigger neurons. We conclude that Bötzinger-complex, expiratory neurons project to the C4 and/or C5 segments of the cervical spinal cord but no further caudal. Their axons are located dorsolaterally in the upper cervical segments of the spinal cord, and they monosynaptically inhibit phrenic motoneurons during the late part of expiration.

Bilaterally independent respiratory rhythms in the decerebrate rat

In rats, respiratory neurons in the medulla oblongata are arranged in longitudinally distributed groups that are duplicated on each side of the neuraxis. Our aim was to determine whether respiratory rhythm is generated independently by each side. We made a complete mid-sagittal section of the medulla oblongata, 3.5 mm rostral and 3.5 mm caudal to the obex, in decerebrate, vagotomized, and paralysed adult rats. Respiratory rhythm, monitored by recording the activity of both left and right phrenic nerves, was maintained and became asynchronous between the left and right sides. We concluded that in the adult rat each half of the medulla oblongata is capable of generating respiratory rhythm independently.

Day-night differences in the respiratory response to hypercapnia in awake adult rats

We investigated whether resting ventilation and the hypercapnic ventilatory response vary with time of day in awake adult rats. Respiratory frequency (fR), tidal volume (VT), inspired ventilation (VI), inspiratory interval (tI), carbon dioxide production (VCO2) and abdominal temperature (Tb) were measured before and during a hypercapnic stimulus (3.5% CO2 in air) at 10:00 and 22:00 h. VCO2, Tb and mean inspiratory air flow (VT/tI) were significantly higher at 22:00 h in air. VI/VCO2 was similar at 10:00 and 22:00 h. VI was significantly elevated by hypercapnia and the response at 22:00 h was 2.3 times greater than that at 10:00 h. VT/tI was unchanged at 10:00 h but significantly increased by hypercapnia at 22:00 h. VCO2 was significantly depressed at 10:00 h but not at 22:00 h. Tb was unaffected by hypercapnia. We conclude that the metabolic and ventilatory responses to hypercapnia are dependent on the time of day.