Respiratory Physiology: Central Neural Control

State-dependent changes in breathing are caused by nonrespiratory (tonic) inputs to the brainstem systems that control ventilation. In wakefulness, tonic excitatory inputs include those from the reticular formation, brainstem aminergic systems, and hypothalamic orexin-containing neurons. In non-rapid eye movement (NREM) sleep, decrements in these excitatory inputs can explain the features of breathing characteristic of this state. In patients with obstructive sleep apneas, these decrements facilitate upper airway obstructions because upper airway muscles fail to properly compensate for airway collapsing effects of the negative pressure generated by respiratory pump muscles. In rapid eye movement (REM) sleep, there are tonic excitatory inputs to the respiratory system that cause the irregularities and rapidity of breathing, as well as tonic and phasic inhibitory inputs that may cause periods of ineffective ventilation. The loss of excitation mediated by serotonin and norepinephrine contributes to the REM sleep-related hypotonia of motor neurons that innervate the genioglossus and possibly also other upper airway muscles. The respiratory muscles are controlled by central neural systems that are influenced by feedback from chemical and mechanical sensors and by the sleep-waking state of the nervous system. This chapter deals with the mechanisms by which state of consciousness affects the respiratory system. We begin with non-rapid eye movement (NREM) sleep, a state in which the respiratory system seems to be in its most elemental configuration, and we then consider rapid eye movement (REM) sleep, in which there are both excitatory and inhibitory effects on the respiratory system. We will show that state-dependent effects are the result of either the presence or the absence of tonic inputs to the central respiratory controller. RESPIRATORY ACTIVITY IN NON-RAPID EYE MOVEMENT SLEEP Characteristics of Breathing in NREM Sleep The frequency of breathing is lower and more regular in NREM sleep than in wakefulness [1,2]. Peak instantaneous airflow rate and the peak negative pressure developed against a narrowed airway decrease, whereas upper airway resistance increases. Tidal volume increases as the result of an increased duration of inspiration, but minute ventilation decreases and end-tidal carbon dioxide concentrations increase. Responses to carbon dioxide and low oxygen are intact (see Chapter 18). In humans, there may be periodic breathing, particularly at high altitude (see Chapter 20) [2]. Medullary Respiratory Neuronal Activity in NREM Sleep There is a decrease in medullary respiratory neuronal activity in NREM sleep [3-6] (Fig. 17-1). The neurons affected are those in the ventral and dorsal respiratory groups. Some cells are affected more than others. Quantitative analysis shows that the effect of sleep on a respiratory neuron is proportional to the amount of nonrespiratory (tonic) activity in the activity of that cell. That is, the drive to a cell may be rhythmic (i.e., respiratory related) or it may be nonrhythmic (i.e., tonic). Respiratory cells whose activity depends primarily on tonic inputs and only weakly on respiratory-modulated inputs are affected more by sleep than respiratory cells whose activity depends primarily on rhythmic, respiratory-related inputs. Some upper airway motor neurons are in the former category of cells, and their activity decreases accordingly. In contrast, neurons whose activity is primarily determined by rhythmic, respiratory inputs do not show dramatic changes in activity in NREM sleep compared to relaxed wakefulness (Fig. 17-2). This indicates that sleep affects primarily neurons that receive large amounts of nonrespiratory inputs. Consistent with this is the finding that iontophoresis of glutamate onto silent and sleep-sensitive respiratory neurons reveals their respiratory activity pattern during sleep [6] This indicates that respiratory-modulated inputs to these cells are not lost in sleep but become subthreshold because of a loss of state-dependent tonic excitatory inputs. Pontine, Mesencephalic, and Telencephalic Respiratory Neuronal Activities There are changes in the activities of pontine parabrachial respiratory-modulated neurons in NREM sleep. Increases and decreases are observed, but on average, decreases are minimal [7-10]. One study has found that respiratory activity in the pons is weak even in wakefulness and that with sleep there are few consistent changes [11]. The functional significance of these and other results showing state-related changes in the activity of respiratory cells in the amygdala, anterior cingulate gyrus, orbital frontal cortex, and mesencephalic central gray[12-14] is not known. Figure 17-1. The activity of a sleep-sensitive respiratory neuron, and the locations of it and others like it. AMB, Nucleus ambiguus; CN, cochlear nucleus; FTL, lateral tegmental field; IO, inferior olive; P, pyramidal tract; RB, restiform body; SOL, solitary tract; 5 SP, nucleus of the spinal tract of V; 5 ST, spinal tract of V; 7, facial nucleus; 12 N, hypoglossal nerve. (From Orem J, Montplaisir J, Dement W: Changes in the activity of respiratory neurones during sleep. Brain Res 1974;82:309-315.) Figure 17-2. The activity of an inspiratory cell during wakefulness and non-rapid eye movement (NREM) sleep. Action potentials and respiration (downward deflection signals inspiration) during wakefulness (A-1) and during NREM sleep (A-2). B, Cycle-triggered histogram of the activity. The omega-squared statistic (with a value of 0.913) expresses the high relationship of the activity to breathing. C and D, Although the numbers of discharges per breath were equivalent in wakefulness (C-1) and NREM sleep (C-2), the frequency of discharge (plotted as slope in D) was slightly, but significantly, lower in NREM sleep. This effect seemed related to the duration of inspiration and was observed for breaths of different durations within wakefulness, as well as between wakefulness and NREM sleep. D/N, Drowsiness/NREM sleep; E, expiration; I, inspiration; n.s., not significant; W/D, wakefulness/drowsiness. (Data from Orem J, Osorio I, Brooks E, et al: Activity of respiratory neurons during NREM sleep. J Neurophysiol 1985;54:1144-1156.) State-Dependent Excitatory Inputs to Respiratory System That Decrease during NREM Sleep Recordings from medullary respiratory neurons show that tonic inputs to them decrease in NREM sleep. Although all sources of these inputs are not known, the following have been implicated by various studies: (1) the brainstem reticular formation, (2) the collection of higher structures that exert behavioral control on the respiratory system, (3) the aminergic brainstem nuclei, and (4) the hypothalamic orexin-containing neurons. These systems all excite the respiratory system and may collectively constitute the wakefulness stimulus for breathing. Another major excitatory drive to 2 breathe originates in central neurons sensitive to pH/CO . Although breathing remains under chemical control during NREM sleep, the contribution of some of these neurons may vary with the sleep-wake cycle. Reticular Formation Stimulation of the reticular formation excites the respiratory system [15-24] Midbrain reticular stimulation causes a reduction in the duration of expiration and an increased rate of rise and amplitude of phrenic nerve activity. It also causes an increase in laryngeal abductor activity [23], converting it from patterns characteristic of NREM sleep to those of wakefulness, and, like wakefulness, reticular stimulation preferentially facilitates the activity of the muscles of the upper airway rather than the muscles of the diaphragm [23]. Respiratory activation declines slowly after the cessation of reticular stimulation. These results imply that, during the transition from wakefulness to NREM sleep, the muscles of the upper airway may lose their tonic excitatory inputs to a greater extent than the diaphragm. This could lead to occlusive collapse of the airway during sleep. Other studies indicate that the neural systems driving upper airway motor neurons are more sensitive than those of phrenic motor neurons to the depressive effects of ethanol, diazepam, pentobarbital, halothane, hypocapnia, chemical stimuli, and thermal depression of neuronal activity near the ventral medullary surface [25-30]. Similarly, the systems controlling the upper airway muscles are more sensitive than those of the diaphragm to the excitatory effects of protriptyline, strychnine, cyanide, and doxapram. There is a preferential activation of upper airway muscles on arousal to wakefulness in response to occlusion [31]. In cats, tracheal occlusions instituted during NREM sleep cause progressive augmentation of both laryngeal abductor and diaphragmatic activity, but increases in laryngeal activity exceed the increases in diaphragmatic activity. The greatest augmentations between one breath and the next were seen when the first occluded breath occurred in sleep and the next in wakefulness. This increase in activity at the transition from sleep to wakefulness was greater for the laryngeal abductors than for the diaphragm. Similarly, the progressive response of the genioglossus muscle to occlusion, as well as the response to hypoxia and hypercapnia, is quantitatively greater than the diaphragmatic response [32,33]. Other studies confirm the powerful effect of arousal on upper airway dilating activity and demonstrate that airway-dilating responses to occlusions and negative pressure during sleep are weak compared to those in wakefulness. It has been suggested that the weaker response in sleep may contribute to pharyngeal collapse in patients with obstructive sleep apnea.34,35 According to this idea, occlusion is the result of failed compensation in sleep. The idea is supported by demonstrations of greater genioglossal muscle activity (compensatory activity) in wakefulness in patients with obstructive sleep apnea than in normal subjects [35]. Behavioral Control Behavioral control of breathing may be reflexive, as occurs in sneezing, coughing, vomiting, and eructation, or voluntary, as during speaking, breath holding, and playing a wind instrument. These behavioral acts require the integration of nonrespiratory inputs into circuits of the respiratory oscillator. Behavioral respiratory acts generally occur only in wakefulness. For example, mechanical and chemical stimulation of the larynx [36] and bronchopulmonary stimulation [37] cause coughing in wakefulness but not in sleep [38]. It is not known why these responses can occur in wakefulness but not sleep, but it may be that the readiness of behavioral control in wakefulness constitutes a stimulus for the respiratory system. The potential for behavioral controllers to affect breathing directly is clear in REM sleep, when they may act in association with dreams (see "Increased Respiratory Neuronal Activity in REM Sleep: Endogenous Excitatory Drives," later). In contrast, in NREM sleep, effects on breathing may be the result of the absence of behavioral control. This may be relevant to obstructive sleep apnea if what is lost in sleep is a wakefulness-dependent behavioral compensation for a high upper airway resistance. The list of structures that can contribute to behavioral control of brainstem and spinal respiratory neurons includes structures from all levels of the neuraxis. The controls exerted by telencephalic structures, amygdala, and the central gray may occur in relation to emotional and volitional acts [39-41]. Many of these higher structures, such as the central nucleus of the amygdala, the anterior cingulate gyrus, the orbital frontal cortex, and the central gray, contain cells that have state-dependent respiratory activity [12-14]. Stimulation or inactivation of limbic [13,14], subcortical [42] and cerebellar [43] structures can influence the respiratory system. The site of behavioral control within the respiratory neuraxis varies depending on the behavioral act. It may be exerted directly on respiratory motor neurons, thus bypassing the central respiratory generator, or on medullary premotor and higher-order central respiratory neurons. Aminergic Systems Serotonin (5-hydroxytryptamine [5-HT])-containing and norepinephrine-containing neurons of the brainstem are an important source of sleep-related changes in breathing. Their activity decreases during sleep and they have extensive axonal projections to respiratory regions. Both central respiratory neurons and respiratory motor neurons have receptors for 5-HT and norepinephrine. The activity of the neurons belonging to these two aminergic systems is highest during active wakefulness, declines during NREM sleep, and is minimal or absent during REM sleep [44-46] Serotonin has excitatory effects on motor neurons, including those innervating the upper airway and respiratory pump muscles [47-51]. Antagonists of the excitatory effects of 5-HT reduce the spontaneous activity of XII motor neurons, thus showing the presence of an endogenous serotonergic excitatory drive [52,53]. Data from pharmacologic models of REM sleep (see "The Atonia of REM Sleep and the Carbachol Models," later) reveal that the suppression of XII motor neuron activity produced during the carbachol-induced REM sleep-like state is associated with silencing of medullary serotonergic cells and decrements in extracellular levels of 5-HT in the XII nucleus region [54,55] (Fig. 17-3). Likewise, norepinephrine levels are reduced in the XII motor nucleus region during the motor atonia elicited by electrical stimulation of the pontine REM sleep-triggering region [56]. Figure 17-3. Extracellular level of 5-hydroxytryptamine (5-HT) is reduced in the region of the hypoglossal (XII) motor nucleus during the rapid eye movement (REM) sleep-like atonia produced by pontine injection of carbachol. A, 5-HT level in microdialysis samples collected in successive 20-minute intervals from the XII nucleus in a decerebrate, paralyzed, and artificially ventilated cat. At the end of collection of sample 14, carbachol was injected into the pons and produced a suppression of (fictive) postural and respiratory activity. One hour and three samples later, pontine microinjection of atropine was made to terminate the atonia. The level of 5-HT decreased in association with the onset of the atonia and then increased when the atonia was terminated. The inset shows the location of the dialysis probe in this experiment. NTS, Nucleus tractus solitarius; XII, hypoglossal motor nucleus. B, Moving averages of the activities recorded from the XII nerve (Hypo) and a cervical nerve branch innervating dorsal neck muscles (C4) at the times of transition into and out of the carbachol-induced atonia. The bars attached to the marker arrows in A indicate the position of the records in B relative to the changes in 5-HT level shown in A. (Modified from Kubin L, Reignier C, Tojima H, et al: Changes in serotonin level in the hypoglossal nucleus region during the carbachol-induced atonia. Brain Res 1994;645:291-302, with