We contextualize a discussion of autonomorespiratory (e.g., sympathorespiratory, parasympathorespiratory, sympathoparasympathetic) coupling (Guyenet et al., 2019; Molkov et al., 2014; Zoccal, 2015) by evaluating the mechanisms generating the individual activities proper (Ghali, 2018, 2017a, 2017b, 2017c, 2019a, 2019b, 2019c) (Figs. 1-14). Distinct, though overlapping, propriobulbar neuronal microcircuit oscillators disparately distributed throughout and specifically organized within, the brainstem and upper cervical spinal cord emergently constituting brainstem and myelic subnetworks and networks (Goodchild and Moon, 2009; Guyenet et al., 2018; Molkov et al., 2017) generate neural breathing (Figs. 1-3) (Anderson et al., 2016; Anderson and Ramirez, 2017; Ghali and Marchenko, 2013, 2015, 2016a, 2016b; Marchenko et al., 2016; Richter, 1982; Richter et al., 1986; Richter and Spyer, 1990) and basal and reflexive changes in sympathetic and cardiovagal tone (Figs. 4 and 5). Multiple distinct (Anderson and Ramirez, 2017) discretely (Molkov et al., 2017) organized metencephalic and myelencephalic nodes constituted by propriobulbar and bulbospinal units exhibiting respiratory-related modulation (Marchenko et al., 2016) RTN (Guyenet et al., 2019); postinspiratory complex (Anderson et al., 2016) coupled by excitatory and inhibitory interneuronal interactions emergently generate triphasic eupnea (defined by Richter (1982); see Richter et al. (1986)) characterizing neural breathing (Ramirez and Baertsch, 2018) and sympathetic and parasympathetic oscillations modulating arterial and venous tone (Dittmar, 1873; Ghali, 2018, 2017a; Gordon and McCann, 1988; Guyenet, 2006) and myocardial contraction force and frequency (Lindsey et al., 1998; Massari et al., 1998), effectively modulating dynamic arterial and venous pressure, resistance, and flow and sinoatrial rate, atrioventricular conduction, atrial and ventricular contractility, output, and function (Ghali, 2018, 2017a; Guyenet, 2006).

Two bilateral rostrocaudally organized columns of nuclei residing within the ventrolateral and dorsal medulla (Figs. 1-3) (Molkov et al., 2017; Ramirez and Baertsch, 2018; Smith et al., 2009), receiving excitatory and inhibitory tonic and phasic modulatory inputs from chemosensitive neurons residing in the retrotrapezoid nucleus (Figs. 1 and 5) (Guyenet et al., 2019) and propriobulbar modulatory synaptic drive from Kölliker-Fuse and medial parabrachial nuclei in the dorsolateral metencephalic tegmentum (Dutschmann and Herbert, 2006; Mörschel and Dutschmann, 2009) emergently generate breathing. Interactions of neurons within the preBötzinger (preBötzC) (Gourévitch and Mellen, 2014; Gray et al., 2010; Morgado-Valle et al., 2010) and Bötzinger (BötzC) (Bongianni et al., 2010; Marchenko et al., 2016) complexes generate a two-phase (i.e., inspiratory and expiratory) neural respiratory rhythm. BötzC decrementing post-inspiratory neurons (Ezure and Manabe, 1988) receiving propriobulbar modulatory excitatory synaptic drive from Kölliker-Fuse and medial parabrachial nuclei in the dorsolateral metencephalic tegmentum (Dutschmann and Herbert, 2006; Mörschel and Dutschmann, 2009), reciprocally interacting with BötzC augmenting late-expiratory neurons via inhibitory neurosynaptic interactions (Smith et al., 2009), segregate expiration into post-inspiratory and late-expiratory phases generating triphasic eupnea. The core rhythm distributes to premotoneurons residing within the rostral and caudal divisions of the ventral respiratory group (see Tian and Duffin (1996a, b)). The dorsal respiratory group (see Richardson and Mitchell (1982)), monosynaptically or polysynaptically projecting to and synaptically driving inspiratory-related motoneurons residing within the brainstem (Ghali, 2019a; van Brederode and Berger, 2011) and spinal cord (Christakos et al., 1988, 1991; Marchenko et al., 2012). Ventral respiratory group brainstem interneurons project to, and putatively modulate the discharge of, neurons residing within the ventrolateral division of the nucleus tractus solitarius (NTS), both zones of which are coupled across the midline by propriobulbar units (Iakunin et al., 1982) and are critically implicated in the genesis of high-frequency oscillations (Romaniuk and Bruce, 1991), evident in respiratory-related neuronal and neural efferent spectra (Christakos et al., 1991; Ghali and Marchenko, 2013; Marchenko et al., 2012). We suggest these synaptic connections modulate mutual sympathorespiratory coupling across sympathetic and breathing oscillators.

Fig. 1.

Conceptual model of brainstem neuronal micro-oscillator synaptic drive to the phrenic motor nucleus. A discrete embryonically homogenous cluster of Phox2b transcription factor-expressing chemosensitive neurons extant near the pontomedullary confluence constitutes the retrotrapezoid nucleus. Hydrogen ion concentration dynamically parallels dynamic retrotrapezoid nucleus glutamatergic neuronal spiking frequency. Hydrogen ions generated by the reaction of arterial CO${}_2$ with H${}_2$O generate the weak acid H${}_2$CO${}_3$, which exists in equilibrium with its constituent dissociation products H${}^{+}$ cation and HCO${}_3$${}^{-}$ cation. The HCO${}_3$${}^{-}$ anion dissociates into H${}^{+}$ and CO${}_3$${}^{2-}$, with an equilibrium preferentially favoring the formation of HCO${}_3$${}^{-}$. These equilibria and dynamics are governed by the Henderson-Hasselbach proportionality, pH = pK${}_a$ + log ([H${}^{+}$] [A${}^{-}$]/[HA]). Retrotrapezoid nucleus glutamatergic units convey prominent tonic excitatory drive and support to propriobulbar and bulbospinal microcircuit oscillators constituting the Bötzinger and pre-Bötzinger complexes, ventral respiratory column nuclei, dorsal respiratory group, and metencephalic elements constituting the brainstem neural respiratory network. Neurons exhibiting differential phase preference extant within the Bötzinger and pre-Bötzinger complexes dynamically interact to generate a core two-phase neural respiratory rhythm and modulate premotoneuronal spiking frequency in the rostral and caudal divisions of the ventral respiratory group. BötzC dec post-I and aug late-E units convey inhibitory modulation to preBötzC pre-I, pre-I/I, and dec early-I units and propriobulbar excitatory drive to cVRG expiratory premotoneurons. PreBötzC pre-I, pre-I/I, and dec early-I units reciprocally inhibit BötzC dec post-I and aug-E cells. PreBötzC excitatory pre-I and pre-I/I units monosynaptically and paucisynaptically drive rVRG aug-I units. PreBötzC inhibitory dec early-I units putatively shape inspiratory ramp by inhibiting the rVRG aug-I premotoneuronal driver population during the early phase of the inspiratory epoch. Rostral ventral respiratory group premotoneurons drive the phasic activity of phrenic motoneurons through projections conveyed through ipsilateral (i.e., pathways which do not decussate or decussate twice across the midline at medullary then upper C${}_1$-C${}_2$ cervical spinal cord or phrenic motoneuronal levels) and contralateral (pathways which decussate once at either brainstem, upper cervical spinal cord, or phrenic motor nucleus levels) ventromedial and lateral funiculi of the myelic substance relaying to phrenic motoneurons monosynaptically or through interposed pre-phrenic interneurons located in the upper cervical spinal cord or phrenic motor nucleus. Phrenic motoneuronal neurites crossing the midline into the contralateral hemicord may receive descending axodendritic and axosomatic inputs from rostral ventral respiratory group axonal terminals conveyed through the ventromedial and lateral funiculi of the spinal cord. Phrenic motoneurons with dendrites decussating across the midline constitute a significant fraction of these units during early neonatal age, and evidence rapid age-dependent decreases. Medullophrenic units (chiefly from BötzC or Kölliker-Fuse nucleus bulbospinal cells) may convey phasic inhibitory modulation of PhMNs. Local pre-phrenic interneurons may coordinate convey phasic inhibition and convey tonic inhibitory modulation of PhMNs. Color traces beneath phrenic neurograms indicate the phase of activity (i.e., inspiratory, expiratory [post-I and E2], or tonically discharging units) of indicated excitatory (+) and inhibitory (-) synapses. RTN, retrotrapezoid nucleus; BötzC, Bötzinger complex; preBötzC, preBötzinger complex; rVRG, rostral ventral respiratory group; cVRG, caudal ventral respiratory group; C${}_1$-C${}_2$ pre-PhINs, C${}_1$-C${}_2$ pre-phrenic interneurons; PhNucl, phrenic nucleus; PhL, left phrenic nerve; PhR, right phrenic nerve. Modified with permission from Fig. 10 of Ghali and Marchenko (2016a)

The central pattern generator originating the breathing rhythm (see Von Euler et al. (1973)) receives powerful modulatory synaptic drive by central and peripheral receptors sensitive to oxygen (Garcia et al., 2016) and/or carbon dioxide tension (Ghali and Marchenko, 2016b; Guyenet et al., 2019), alveolar stretch (Breuer, 1868; Dutschmann et al., 2014; Hering, 1868; Luck, 1969), laryngeal fibers (D’iachenko and Preobrazhenskiĭ, 1984a, b), and arterial stretch (Baekey et al., 2010; Bishop, 1974; Lindsey et al., 1998; Morrison et al., 1984). Afferent inputs to propriobulbar interneuronal microcircuit oscillators generating the sympathetic and respiratory neural activities accordingly derive from higher-order neurons of the nucleus tractus solitarius (Fig. 5) (Jean, 1992). These influences monomodally and multimodally integrate and filter oscillatory sensory inputs centrally conveyed by peripheral afferents innervating the carotid sinus and body and aortic arch baroreceptors and chemoreceptors and diffusely distributed central chemoreceptors (Michelini, 1994; Moreira et al., 2007; Rogers et al., 2000; Takakura et al., 2007) located within the brainstem and cerebellum (Mulkey et al., 2006). Drugs modulating sympathetic neural discharge and arterial pressure may coordinately modify the activity of the central sympathetic oscillators, autonomic ganglia, and peripheral effectors (Gillis, 1995). Propriobulbar interneuronal microcircuit oscillators constituting the respective central pattern generators critically gate incoming central and peripheral sensory influences conveyed by afferent neuronal synaptic inputs (Baev et al., 1981), a principle well demonstrated in brainstem vestibular networks (Marlinskiĭ et al., 1989; Preobrazhenskiĭ et al., 1989; Shumilina and Preobrazhenskiĭ, 1991) and spinal cord substantia gelatinosa. Evidence suggests propriospinal interneuronal microcircuit oscillators may be capable of generating rhythmic activity in respiratory-related neural and motor outputs (Ghali and Marchenko, 2016a).

The majority of neurophysiologists presume sympathetic activity to originate within the propriobulbar and presympathetic neuronal circuitry of the rostral ventrolateral medulla (Fig. 3) (Guyenet et al., 2018). These zones are constituted chiefly of glutamatergic (VGlut2 positive) and catecholaminergic (tyrosine hydroxylase immunoreactive) cells. Though the RVLM propriobulbar and presympathetic neurons putatively generate sympathetic activity, evidence more strongly favors principal genesis by medullary lateral tegmental field propriobulbar interneurons (Ghali, 2018), a wise evolutionary construct in the event of ischemic injury (personal communication, V. Marchenko). Accordingly, early experiments interrogating brainstem circuitry generating sympathetic activity provided initial evidence indicating a diffuse distribution of medullary units exhibiting sympathetic-correlated discharge (Preobrazhenskiĭ, 1966). A robust set of data validates genesis of sympathetic activity derives from synchronized activity among sympathoexcitatory and sympathoinhibitory propriobulbar interneuronal microcircuit oscillators residing within the medullary lateral tegmental field (Fig. 4) (Barman, 2020; Dampney, 1994; Gebber and Barman, 1988; Ghali, 2018; Marchenko and Sapru, 2003), thence conveyed to diffusely distributed presympathetic neurons residing within the rostral ventrolateral and ventromedial medulla (Babic and Ciriello, 2004), caudal raphé, and ventrolateral metencephalic tegmentum projecting to and driving neuronal spiking of, intermediolateral cell column preganglionic sympathetic neurons (Barman, 2020; Ghali, 2018, 2017a; Gilbey et al., 1995; Zhou and Gilbey, 1992). Medullary lateral tegmental field neurons oscillate with frequencies of 2 to 6 Hz and 10 Hz, correlated with analogous spectra in cardiac sympathetic nerve discharge precede and drive similar activities in rostral ventrolateral medullary units (Barman, 2020; Ghali, 2018). Rostral ventromedial medullary presympathetic bulbospinal neurons specifically convey synaptic drive to intermediolateral cell column preganglionic sympathetic neurons projecting to postganglionic sympathetic neurons modulating cutaneous vasoconstrictor tone and thus contributing prominently to thermoregulation (see Guyenet (2006)) for outstanding review). Rostral ventrolateral medullary presympathetic neurons receive glutamatergic axodendritic, and axosomatic synaptic inputs drive biophysically transduced by N-methyl-D-aspartate (NMDA), $\alpha$-amino-3-hydroxy-5-methyl-4-isoxazolepropioinic acid (AMPA), and kainate-type glutamate receptors. $\alpha$${}_2$ autoreceptors localizing to the presynaptic axonal membranes of propriobulbar units exhibiting oscillations correlating with cardiac sympathetic nerve discharge negatively modulate the release of excitatory neurotransmitters from axonal terminals, mitigating the release of excitatory amino acid-mediated electrochemical amplification of somatodendritic conductance and contributing to hypotensive effects generated by administration of clonidine and $\alpha$-methyldopa. Sympathetic-related propriobulbar units receive prominent tonic inhibitory GABAergic and glycinergic modulation from specific nuclearly-arranged neuronal clusters and disparately distributed units located in surrounding reticular zones (Gatti et al., 1987). Neurogenesis of sympathetic nerve activity and rhythmic discharge may alternately be generated by a network (Lipski et al., 1996a, b) and/or pacemaker (Sun and Guyenet, 1986; Sun et al., 1988) mechanisms, organized by the medullary lateral tegmental field (Barman et al., 2005; Barman, 2020; Dampney, 1994; Ghali, 2017a; Marchenko and Sapru, 2003), rostral ventrolateral medulla (Guyenet, 2006), and distributed regions within the brainstem (Figs. 5) (Ghali, 2018; Goodchild and Moon, 2009), neuroanatomically overlapping and interacting with the neuronal propriobulbar circuitry generating and organizing neural breathing (Figs. 4 and 5) (see Molkov et al., 2014; Anderson and Ramirez, 2017; Molkov et al., 2017).

Fig. 2.

Brainstem oscillators generating breathing rhythms. Cells distributed along the ventrolateral column of pontomedullary nuclei generate the neural breathing rhythm and pattern. The retrotrapezoid nucleus/parafacial respiratory group (RTN/pFRG), lies ventral with respect to the facial nucleus (VII). The postinspiratory complex (PiCo) may be found caudal to the facial nucleus, rostral to the preBötzinger complex, and dorsomedial with respect to the rostral pole of the nucleus ambiguus (NA). The Bötzinger (BötC) and preBötzinger complexes (preBötC) may be found ventromedial to the nucleus ambiguus and rostral with respect to the rostral (rVRG) and caudal (cVRG) divisions of the ventral respiratory group. The lateral reticular nucleus (LRt) conceals the preBötzC, rVRG, and cVRG from the ventral surface of the medulla. We presume lateral reticular nucleus neuronal somata constitute interposed relay phrenicomedullary inputs from the phrenic nucleus to the bulbar ventral respiratory column. Analogously, the serpiginous dorsal margin of the inferior olive separates the ventral respiratory group from the ventral surface of the brainstem in humans. Compartmental borders containing the ventral respiratory column nuclei are not markedly distinct. Electrophysiological properties and differential spatiotemporal firing dynamics have characteristically distinguished these regions, though investigators continue to strive to identify specific cell surface proteins and transcription factors distinguishing these zones. PreBötzinger complex neurons frequently express SST1 and/or the transcription factor Dbx1. Modified with permission from (Fig. 1) of Anderson et al. (2016)

Diffusely distributed nuclearly arranged neuronal clusters residing within functionally distinct subnuclei of the nucleus tractus solitarius (Fig. 5) (Michelini, 1994; Moreira et al., 2007; Takakura et al., 2007; Zhao et al., 1997) play modulatory roles in, and constitute the chief nodes mediating and contributing to respiratory, sympathetic, and parasympathetic crossmodal modulation and coupling (Guyenet, 2006; Molkov et al., 2014). Nucleus tractus solitarius units monomodally filter and/or multimodally integrate a wide array of propriobulbar interneuronal inputs and axon terminals from peripheral afferent fibers, coordinately conveying monosynaptic and polysynaptic inputs to higher-order neurons (Fig. 5) (Von Euler et al., 1973), coordinately negatively modulating propriobulbar interneuronal sympathetic- and inspiratory-related microcircuit oscillators and positively modulating propriobulbar interneuronal cardiovagal- and expiratory-related microcircuit oscillators (Guyenet, 2006; Rogers et al., 2000; Moreira et al., 2007; Takakura et al., 2007). Dorsal respiratory group bulbospinal premotoneurons convey axodendritic and axosomatic synaptic drive to the brainstem and spinal cord respiratory-related motoneurons (Lipski et al., 1990). Nucleus ambiguus (i.e., N.A.) and dorsal motor nucleus of the vagus (DMV) convey cardiovagal preganglionic neurons en route to cardiac plexus ganglia postganglionic parasympathetic neurons powerfully modulating sinoatrial chronotropy, atrioventricular dromotropy, and atrial inotropy (Dyavanapalli et al., 2013; Eckberg, 2003; Taylor et al., 2009), though a few fibers coordinately supply ventricular myocardium (Coote, 2013). Postganglionic sympathetic and parasympathetic cardiovagal axons modulate sinoatrial action potential frequency by influencing the slope of spontaneous depolarization mediated by persistent sodium channel current in sinoatrial nodal cells, atrioventricular conductivity by regulating the slope of voltage-gated calcium current rise in atrioventricular nodal cells, and atrial and ventricular contractility by determining magnitude and rate of rising of cytosolic calcium flux deriving from the extracellular space and sarcoplasmic reticulum stores (Ghali, 2018, 2017a). The data collectively substantiates monomodal and multimodal sensory influences effectively converge upon, diverge through, and modulate the discharge of, brainstem sympathetic oscillators and cardiovagal premotoneurons (Ghali, 2018, 2017a). Visceromotor (Duda et al., 1970a, b; Kostiuk and Preobrazhenskiĭ, 1966; Preobrazhenskiĭ and Tamarova, 1966; Preobrazhenskiĭ et al., 1972), somatosympathetic (Ghali, 2019a), and somatorespiratory (Ghali, 2019a) reflexes exhibit sympathosomatic and respirosomatic coupling and extensive interconnectivity amongst propriobulbar interneuronal networks generating and mediating nociceptive responses, sympathetic discharge, and the breathing rhythm.

Prefacing a discussion of autonomorespirophasic modulation (Figs. 5-14), we briefly discuss the neural network and pacemaker mechanisms emergently generating the breathing rhythm and pattern (Molkov et al., 2017), sympathetic oscillations (Barman, 2020; Ghali, 2018, 2017a; Molkov et al., 2014), and cardiovagal premotoneuronal discharge (Bishop, 1974). We discuss respirophasic modulation of brainstem presympathetic neuronal oscillations (Guyenet et al., 1990; Haselton and Guyenet, 1989; Lipski et al., 1996a, b; McAllen, 1987; Miyawaki et al., 1996; Numao et al., 1987), sympathetic preganglionic and postganglionic discharge (Guyenet et al., 1990; Mandel and Schreihofer, 2006; Zhou and Gilbey, 1992) and cardiovagal premotoneuronal spiking (Gilbey et al., 1984; Lindsey et al., 1998). We reciprocally discuss oscillatory baroreceptor modulation of the breathing pattern (Baekey et al., 2010; Bishop, 1974; Brunner et al., 1982; Dove and Katona, 1985; Lindsey et al., 1998) and respiratory gating of the baroreflex (Baekey et al., 2010).

Fig. 3.

Interactions amongst the breathing generator, and sympathetic oscillators in the brainstem. PreBötzinger complex preinspiratory inspiratory phase-spanning and decrementing early inspiratory units receive phasic inhibitory modulation from Bötzinger complex decrementing post-inspiratory and augmenting expiratory units and tonic excitatory drive from retrotrapezoid nucleus tonic units and parafacial respiratory group pre-inspiratory units, which receive the tonic excitatory drive from CO${}_2$-driven chemosensitive tonic units extant within the same nucleus. PreBötzinger complex preinspiratory inspiratory phase spanning units drive augmenting inspiratory units in the rostral division of the ventral respiratory group and decrementing early inspiratory units in the preBötzinger complex. recurrent Excitatory and inhibitory synaptic coupling synchronizes preBötzinger complex spontaneously bursting preinspiratory-inspiratory neuronal spiking. Pre-Bötzinger complex decrementing early inspiratory units convey inhibitory drive to Bötzinger complex augmenting late-expiratory units. Bötzinger complex postinspiratory and augmenting late-expiratory units exhibit reciprocal inhibition and convey inhibitory axodendritic and axosomatic drive to the caudal division of premotoneurons in the caudal ventral respiratory group. Bötzinger complex excitatory postinspiratory units drive metencephalic inspiratory-expiratory phase-spanning units. The firing of metencephalic inspiratory-expiratory phase-spanning units, tonic excitatory glutamatergic retrotrapezoid nucleus units, parafacial respiratory group late-expiratory unit drives and shapes rostral ventrolateral medullary presymapthetic bulbospinal neuronal spiking. Rostral ventrolateral medullary presympathetic bulbospinal neurons receive the inhibitory drive from Bötzinger complex postinspiratory units and rostral ventral respiratory group early inspiratory units and convey descending excitatory inputs to intermediolateral cell column preganglionic sympathetic neurons. Propriobulbar units distributed within the vast expanse of the ventral respiratory column nuclei provide the tonic excitatory drive to rostral ventrolateral medullary presymapthetic units and caudal ventrolateral medullary GABAergic units. Caudal ventrolateral medullary GABAergic propriobulbar interneurons convey a prominent and physiologically critical inhibitory modulation upon rostral ventrolateral medullary presympathetic units, the pharmacological elimination of which generates rapid and massive rises of dynamic sympathogenic neural efferent activity and arterial pressure magnitude. Rostral ventral respiratory group early inspiratory units convey inhibitory modulation upon co-extant rostral ventral respiratory group augmenting inspiratory premotoneurons, caudal ventral respiratory group expiratory premotoneurons, and rostral ventrolateral medullary presympathetic units. The metencephalon, raphé, retrotrapezoid nucleus, and ventrolateral myelencephalon constitute sources of tonic excitatory drive to respiratory-related and sympathetic units. Metencephalic inspiratory expiratory phase-spanning units receive the tonic excitatory drive from metencephalic propriobulbar units and Bötzinger complex excitatory postinspiratory units, dynamic spiking of which confers respirophasic modulation upon the rostral ventrolateral medullary presympathetic units. The phrenic motoneuronal and sympathetic axons constituting the phrenic nerve drive the principal mammalian inspiratory neuromotor discharge. The set of abdominal nerves drive expiratory neuromotor output. The thoracic sympathetic nerve represents a common, empirically accessible measure of sympathetic activity. The model contains two populations of nucleus tractus solitarius units receiving oscillatory baroreceptor inputs. One population of barosensitive nucleus tractus solitarius neurons synaptically excites caudal ventrolateral medullary GABAergic interneurons, and the complement population of barosensitive nucleus tractus solitarius units synaptically excites Bötzinger complex inhibitory decrementing postinspiratory units. Barosensitive nucleus tractus solitarius cells, in turn, receive prominent inhibitory modulation from metencephalic inspiratory units and rostral ventral respiratory group early inspiratory neurons. Excitatory (dark red) and inhibitory (purple) populations of units and interactions are indicated. RVLM, rostral ventrolateral medulla; CVLM, caudal ventrolateral medulla; VLM, ventrolateral medulla; RTN, retrotrapezoid nucleus; pFRG, parafacial respiratory group; late-E, late expiration; BötC, Bötzinger complex; pre-BötC, preBötzinger complex; rVRG, rostral ventral respiratory group; cVRG, caudal ventral respiratory group; bs-E, bulbospinal expiratory; PN, phrenic nerve; AbN, abdominal; tSN, thoracic sympathetic nerve; VRC, ventral respiratory column; I, inspiratory; IE, inspiratory expiratory phase spanning; late-E, expiratory; post-I, postinspiratory; pre-I/I, preinspiratory inspiratory phase spanning; early-I(1), preBötzinger early-inspiratory units; early-I(2), rVRG early inspiratory units. Modified with permission from Fig. 6 of Molkov et al. (2017).

Sympathorespiratory coupling was initially described by Adrian and Bronk in 1932 in the anesthetized rabbit (Adrian et al., 1932), demonstrating prominent differential modulation of the sympathetic neural efferent discharge varying according to the respiratory cycle, though perhaps previously intimated by studies conducted in the 1800s. Notably, cervical sympathetic burst amplitude increased markedly and dynamically toward the end of the inspiratory epoch (Adrian et al., 1932). Studies have continued to provide evidence indicating respirophasic modulation of sympathetic neural efferent activity, with variable accentuations and troughs of sympathetic bursting corresponding with precise intervals within the inspiratory and expiratory epochs (Molkov et al., 2014; Moraes et al., 2012a, b, c; Zoccal and Machado, 2011; Zoccal, 2015). Reciprocally, respiratory-related neuronal and neural outputs exhibit baromodulation and strong correlation with dynamic arterial pressure magnitude (Darnall and Guyenet, 1990; Gilbey et al., 1986; Lipski et al., 1996a, b; McAllen, 1987; Zhou and Gilbey, 1992). Sympathorespiratory coupling occurs during eucapnic normoxia (Adrian et al., 1932), though it becomes strengthened by hypoxia and hypercapnia (Moraes et al., 2012a, b, c; Zoccal and Machado, 2011; Zoccal, 2015).

Crossmodal modulation of neural breathing, sympathetic oscillations, and cardiovagal premotoneuronal spiking couples the activities (Gilbey et al., 1995; McAllen, 1987; Molkov et al., 2014; Zoccal and Machado, 2011; Zoccal, 2015). Propriospinal microcircuit oscillators constituting emergently dynamic networks generate the respiratory rhythm and pattern (Figs. 1-3) (Ghali and Ghali, 2020), sympathetic oscillations (Figs. 4 and 5) (Barman, 2020), and cardiovagal premotoneuronal activity (Figs. 4 and  5) (Frank et al., 2009). The breathing cycle modulates rostral ventrolateral medullary sympathetic propriobulbar and presympathetic units (Gilbey et al., 1995; Lipski et al., 1996a, b; Mandel and Schreihofer, 2006; McAllen, 1987; Miyawaki et al., 1996; Terui et al., 1986), and spinal cord preganglionic sympathetic neurons (Zhou and Gilbey, 1992) (Fig. 7) sympathetic neural efferent nerve discharge (Fig. 6). Reciprocally, sympathetic and baroreceptor oscillations modulate respiratory-related brainstem and spinal cord neuronal spiking pattern (Baekey et al., 2010; Bishop, 1974; Brunner et al., 1982; Dove and Katona, 1985; Eckberg and Orshan, 1977; Grunstein et al., 1975; Lindsey et al., 1998). A large fraction of respiratory-related units recorded in unanesthetized decerebrate vagotomized cats exhibits pulse-synchronous modulation (Lindsey et al., 1998). Accentuations or attenuations of neuronal or neural efferent spiking frequency or burst clustering occurring in response to synaptic inputs conveyed by another group of neurons constitutes modulation (Gilbey et al., 1995; Lindsey et al., 1998; Zhou and Gilbey, 1992). We designate a rostroventrolateral medullary presympathetic neuronal unit exhibiting peak spiking frequency during the inspiratory epoch of breathing to possess inspiratory modulation (Lipski et al., 1996a, b; Mandel and Schreihofer, 2006; Miyawaki et al., 1996).

Fig. 4.

Brainstem zones generating and modulating sympathetic oscillations and cardiovagal premotoneuronal spiking. Nissl stained parasagittal sections 1.8 mm lateral with respect to the midline. The chief medullary pre-sympathetic units project to intermediolateral cell column (IML) preganglionic sympathetic neurons reside within the rostral ventrolateral medulla (RVLM), rostral ventromedial medulla (RVMM), caudal raphé, ventrolateral metencephalic tegmentum (A5 catecholaminergic group of neurons), paraventricular nucleus of the hypothalamus, and the medullocervical pressor area (MCPA) spanning the caudal extent of the medulla and rostral extent of the cervical spinal cord. The MCPA extends at least as far caudal as C3/C4. The CVLM may be found caudally positioned with respect to the RVLM, constituted chiefly by GABAergic propriobulbar inhibitory interneurons (I-CVLM), which confer inhibitory modulation upon RVLM presympathetic units. Some I-CVLM units are modulated by the baroreflex arc and are consequently activated by higher-order barosensitive nucleus tratus solitarius (NTS) units, in turn, modulated by barosensitive afferents conveying oscillatory baroreceptor inputs from carotid sinus and aortic arch, relaying via cranial nerves IX and X, respectively. Another group of GABAergic I-CVLM units are baro-independent and tonically inhibit RVLM pre-sympathetic units. Some units in CVLM are glutamatergic (E-CVLM) and convey excitatory drive to RVLM presympathetic units. Chemoreceptor afferents terminate onto second-order neurons in the commissural nucleus of the NTS, which in turn project to and excite RVLM units directly. Chemosensitive commissural NTS units project to, excite, and sensitize RTN glutamatergic chemosensitive units. Prominent rises of arterial pressure were elicited by microinjecting glutamate into the commissural division of the nucleus tractus solitarius (unpublished observations). Cerebellectomy was not required. Careful dissection of the arachnoid bridging the interval between the calamus scriptorius and cerebellum failed to transgress parenchyma and successfully prepared the dorsal medullary surface for micropipette insertion. RTN units convey tonic excitatory drive to RVLM presympathetic units. Water undergoes a carbonic anhydrase-catalyzed chemical reaction to generate carbonic acid, which dissociates into bicarbonate anion and hydrogen cation. Hydrogen ions potently stimulate chemosensitive units within the retrotrapezoid nucleus, raphé, and nucleus tractus solitarius, among other central chemoreceptor sites. The caudal pressor area (CPA) may be found caudally positioned with respect to CVLM and mediates sympathoexcitation through the disinhibition of RVLM presympathetic units by inhibiting CVLM GABAergic inhibitory interneurons, and via facilitation, by activating a population of E-CVLM sympathoexcitatory glutamatergic units. Efferent interneuronal projections from the CPA to the commissural NTS may underlie the commissural NTS-dependence of pressor responses elicited by CPA stimulation. However, further studies are necessary to more fully elucidate the physiological significance of this specific pathway (Takakura et al., 2007). Yet to be determined, sources convey inhibitory modulation upon the CPA. RVLM-independence of sympathoexcitatory pressor responses elicited by chemical microstimulation effectively physiologically distinguishes the medullocervical pressor area from the caudal pressor area (Figs. 4) of Goodchild and Moon (2009).

During normoxic normocapnia, autonomorespiratory crossmodal modulation chiefly originates from the provision of coordinate inputs deriving from higher-order neural nets (Abdala et al., 2009; Moraes et al., 2012b) and barosensitive, chemosensitive, and pulmonary stretch mechanosensitive monomodal and multimodal nucleus tractus solitarius units (Moreira et al., 2007; Takakura et al., 2007) and direct interaction amongst intermingled propriobulbar interneuronal microcircuits constituting the respiratory central pattern generator, sympathetic oscillators, and dorsal medullary cardiovagal premotoneurons (Moraes et al., 2012c, 2014a, 2014b; Sun et al., 1997). Peripheral stimuli coordinately driving higher order nucleus tractus solitarius neurons or directly stimulated central chemoreceptors and interactions amongst propriobulbar interneuronal microcircuit oscillators emergently constituting neuronal ensembles generating breathing, sympathetic oscillations, and cardiovagal premotoneuronal spiking emergently generate sympathorespiratory coupling (Figs. 6-14) (Molkov et al., 2014).

Strengthening of autonomorespiratory coupling (sympathorespiratory [SRM] and respirosympathetic [RSM] modulation) by hypoxia and hypercapnia (Figs. 8-11) (Molkov et al., 2014; Moraes et al., 2012a, b, c; Zoccal and Machado, 2011; Zoccal, 2015) indicates peripheral effectors or central genesis mechanisms may alternately or coordinately conceivably mediate evident crossmodal modulation. Autonomorespiratory coupling persists despite decerebrative encephalotomy, vagotomy, and peripheral chemodenervation (Adrian et al., 1932; Marchenko et al., 2016; Zhou et al., 1996) supports and evidences the direct interaction of the central pattern generators (CPGs) in mediating crossmodal modulation of the respiratory, sympathetic, and parasympathetic activities. We use the term “autonomorespiratory” to efficiently indicate crossmodal modulation and coupling amongst and between the brainstem breathing generator, sympathetic oscillators, and dorsal medullary cardiovagal premotoneurons residing within the nucleus ambiguus and dorsal motor nucleus of the vagus. We thus operantly define autonomorespiratory to constitute the manifestation of, and the mechanisms emergently generating, sympathorespiratory, parasympathorespiratory, and sympathoparasympathetic crossmodal modulation and coupling across distinct though overlapping propriobulbar interneuronal microcircuit oscillators, nuclei, and reticular zones.

Fig. 5.

Medullary lateral tegmental field and rostral ventrolateral medulla. The nucleus reticularis parvocellularis and nucleus reticularis oralis constitute the medullary division of the lateral tegmental field and span a vast extent of the reticular formation from the caudolateral medulla, becoming more compact with a rostromedial ascent towards the tip of the rostral ventrolateral medulla, giving the architectonic appearance of the cometary tail of the rostral ventrolateral medulla. The medullary lateral tegmental field contains barosensitive units conveying sympathoexcitatory and sympathoinhibitory propriobulbar synaptic inputs to rostral ventrolateral medullary bulbospinal inputs. Most neurophysiologists believe the medullary lateral tegmental field originates the sympathetic activity and conveys this to presympathetic bulbospinal units in nuclei residing within the metencephalon and myelencephalon, including the rostral ventrolateral medulla, rostral ventromedial medulla, caudal raphé, and ventrolateral metencephalic tegmentum. Medullary lateral tegmental field neuronal spiking precedes unitary activity of rostral ventrolateral medullary units and correlates with presympathetic bulbospinal neurons and cardiac sympathetic nerve discharge. Medullary lateral tegmental field propriobulbar interneuronal microcircuit oscillators modulate basal sympathoexcitation and mediate baroreceptor, chemoreceptor, Bezold-Jarisch, somatoautonomic, and visceroautonomic reflexes via NMDA and/or non-NMDA dependent glutamatergic fast excitatory synaptic neurotransmission. NRP, nucleus reticularis paragigantocellularis; NRV, nucleus reticularis ventralis; NTS, nucleus tractus solitarius; RVLM, rostral ventrolateral medulla.

Myelencephalic nuclei chiefly constituted by propriobulbar interneurons and presympathetic bulbospinal neurons (Ghali, 2018, 2017a) exhibit respirophasic modulation, in rostral ventrolateral medullary (Fig. 5) (Baekey et al., 2010; Lipski et al., 1996a, b; McAllen, 1987; Miyawaki et al., 1995; Terui et al., 1986; Zoccal et al., 2008, 2009a, 2009b), caudal ventrolateral medullary (Mandel and Schreihofer, 2006), caudal raphé (Gilbey et al., 1995), and ventrolateral metencephalic A5 catecholaminergic cell group neuronal recordings (McAllen, 1987; Molkov et al., 2017; Zoccal et al., 2008; Zoccal, 2015). respirophasic modulationPulmonary mechanosensitive nucleus tractus solitarius units modulate the discharge of presympathetic units modulating sinoatrial bursting frequency (Brodie and Russell, 1900; Daly, 1986). Rostral ventrolateral medullary neurons receive the tonic excitatory and inhibitory synaptic drive from propriobulbar interneuronal microcircuit oscillators (Frank et al., 2009; Gilbey et al., 1984; Guyenet and Koshiya, 1995; Guyenet, 2000; Mandel and Schreihofer, 2006; McAllen, 1986; Mendelowitz, 1999) and phasic oscillatory pulmonary mechanoreceptor, chemoreceptor, and baroreceptor synaptic inputs (Rogers et al., 2000; Haselton and Guyenet, 1989; Lipski et al., 1996a, b). Ventrolateral metencephalic A5 catecholaminergic cell group presympathetic bulbospinal units exhibit inspiratory, postinspiratory, or expiratory preferential modulation in anesthetized vagotomized rats (Gilbey et al., 1995). Respiratory propriobulbar interneuronal microcircuit oscillators may modulate sympathetic activity through direct synaptic interactions of overlapping propriobular interneuronal networks or indirectly by powerfully gating baroreflex influences upon sympathetic oscillators and modifying baroreflex gain, sensitivity, and transfer kinetics (Eckberg and Orshan, 1977; Porta et al., 2015; Wallin and Eckberg, 1982). Respiratory gating of the baroreflex may confer respirophasic modulation upon sympathetic and parasympathetic oscillations by directly regulating dynamic arterial pressure magnitude and/or indirectly influencing sinoatrial action potential frequency (Figs. 13 and 14). Dynamic fluctuations of cardiac interval spectral bands may modulate arterial pressure spectral power (Julien, 2006). Persistence of sympathetic modulation of neural breathing evident in sinoaortic denervated preparations would seem to indicate sympathetic oscillations generated by medullary lateral tegmental field propriobulbar interneurons distributes and propagates to propriobulbar interneuronal microcircuit oscillators constituting the breathing pattern generator (Barman, 2020; Ghali, 2018, 2017a; Molkov et al., 2014). As an aside, acoustic tempo at 80 beats per minutes coordinately elicited increases in myocardial contraction frequency and spectral power of cardiac interval Traube-Hering and Mayer waves in human subjects (Watanabe et al., 2015), with cochleoautonomic coupling putatively organized by the medullary lateral tegmental field (Ghali, 2018).

Catecholaminergic and glutamatergic propriobulbar and spinally projecting bulbar presympathetic neurons constitute the rostral ventrolateral medulla, from which pressor responses may be elicited by electrical activation or glutamate microinjections (Figs. 4 and 5). Rostral ventrolateral medullary neurons exhibit respirophasic modulation of spiking frequency (Haselton and Guyenet, 1989; McAllen, 1987; Miyawaki et al., 1995; Terui et al., 1986), determined by extracellular or intracellular unitary recordings, and membrane voltage trajectories, determined exclusively by intracellular recordings (Lipski et al., 1996a, b). These patterns include early-inspiratory, inspiratory, postinspiratory, and expiratory accentuation and typically phase-correlate with sympathetic neural efferent activity (i.e., cervical, thoracic, splanchnic, lumbar, renal nerves). McAllen (1987) revealed rostral ventrolateral medullary neurons exhibit peak unitary spiking frequency during the inspiratory epoch, with reciprocal postinspiratory attenuation of action potential frequency, in chloralose-anesthetized vagotomized cats. Haselton and Guyenet (1989) demonstrated inspiratory accentuation of rostral ventrolateral medullary neuronal spiking in halothane anesthetized rats. In contrast, Miyawaki et al. (1996) demonstrated inspiratory-troughs of rostral ventrolateral medullary neuronal spiking in anesthetized rats. Miyawaki et al. (1995) demonstrated respirophasic modulation of monomodally barosensitive rostral ventrolateral medullary neurons in anesthetized vagotomized rats, and Terui et al. (1986) demonstrated respirophasic modulation of coordinately barosensitive chemosensitive rostral ventrolateral medullary neurons in anesthetized vagotomized rabbits. Oscillatory baroreceptor inputs inhibit rostral ventrolateral medullary neuronal spiking frequency to a greater extent the inspiratory, compared with the expiratory, epoch, evidencing respiratory gating of baroinhibition of rostral ventrolateral medullary neurons, in vagotomized anesthetized rats (Miyawaki et al., 1995).

Overlapping Bötzinger complex and rostral ventrolateral medullary neuronal microcircuit oscillators constitute a chief node imposing respirophasic modulation upon sympathetic output (Moraes et al., 2012c; Sun et al., 1997), through which proximally-related nuclei (i.e., NTS, RTN) exert common coupling (Ghali, 2017a; Molkov et al., 2014). We conceive the rostral ventrolateral medulla represents a compact specialized extension of the medullary lateral tegmental field supplying presympathetic bulbospinal drive to intermediolateral cell column preganglionic sympathetic neurons (Marchenko and Sapru, 2003; personal communication, V. Marchenko; see Ghali (2018). RVLM units receive the pre-presympathetic drive from the medullary division of the LTF. The coordinate drive of multiple sets of preganglionic sympathetic neurons residing in the intermediolateral cell column by RVLM and extra-RVLM supraspinal presympathetic units constitutes a mechanism emergently generating common coupling across heterologous and homologous pairs of sympathetic neural discharges. Most commonly, sympathetic activity exhibits postinspiratory or a phase-spanning late-inspiratory post-inspiratory accentuation of burst amplitude during normoxic or hyperoxic eucapnia. Mild, moderate, and severe degrees of hypercapnia, acute hypoxia, and chronic intermittent hypoxia incipiently shift this modulation to the late-expiratory phase and synchronously elicit late-expiratory abdominal nerve bursting (Figs. 8-11) (Molkov et al., 2017; Zoccal et al., 2008; Zoccal and Machado, 2011; Zoccal, 2015). Oscillatory synchrony amplifies the sympathetic activity. Interrogating the mechanistic underpinnings of this behavior informs emergent synchrony across disparate neural networks and fundamental strategies of neuro-computational processing.

Mandel and Schreihofer (2006) demonstrated varying patterns of respirophasic modulation imposed upon caudal ventrolateral medullary neurons in chloralose-anesthetized vagotomized rats (Mandel and Schreihofer, 2006). These included inspiratory augmentation, inspiratory inhibition, inspiratory augmentation with postinspiratory inhibition, and postinspiratory augmentation of respirophasic bursting (Mandel and Schreihofer, 2006). Baromodulated caudal ventrolateral medullary neurons (Mandel and Schreihofer, 2006) may constitute a chief mechanism conveying inhibitory modulation upon rostral ventrolateral medullary excitatory presympathetic neurons via GABAergic interneuronal projections respirophasic modulation (see Fig. 5). GABAergic caudal ventrolateral medullary units exhibiting postinspiratory preferential modulation may generate rostral ventrolateral medullary neuronal spiking troughs during the postinspiratory epoch. respirophasic modulation Electrical stimulation of aortic depressor nerve generates greater inhibitory effects upon rostral ventrolateral medullary neuronal spiking during the inspiratory, compared with the expiratory, epochs. Baro- and respiratory- modulated caudal ventrolateral medullary neurons exhibiting inspiratory preferential spiking analogously exhibit robust postinspiratory inhibition (Mandel and Schreihofer, 2006). Postinspiratory accentuations and troughs of rostral ventrolateral medullary neuronal spiking frequency may result from the influences of excitatory and inhibitory populations of Bötzinger complex decrementing postinspiratory cells (Sun et al., 1997) or caudal ventrolateral medullary postinspiratory modulated neurons (Mandel and Schreihofer, 2006), respectively. Occasional differential phase-specific respirophasic modulation of rostral ventrolateral medullary and caudal ventrolateral medullary neuronal spiking frequency suggests dynamic network reconfiguration of common propriobulbar interneuronal microcircuit oscillators coupling these zones (Mandel and Schreihofer, 2006).

Caudal raphé presympathetic neurons convey axodendritic and axosomatic drive to preganglionic sympathetic units residing in thoracolumbar intermediolateral cell column propriospinal interneuronal microcircuit oscillators driving preganglionic sympathetic neurons (Morrison and Gebber, 1982) and are modulated by respiratory-related propriobulbar interneuronal microcircuit oscillators residing within a few discrete zones. Respirophasic modulation of brainstem presympathetic units thence propagates throughout the sympathetic and parasympathetic neural networks (Gilbey et al., 1995; Lindsey et al., 1992). Gilbey et al. (1995) interrogated respirophasic modulation of caudal raphé presympathetic neurons in anesthetized vagotomized rats by conducting extracellular recordings of the cellular somata of raphé neurons. Antidromic activation conducted at upper thoracic spinal cord segments T1 to T3 electrophysiologically identified presympathetic units (Gilbey et al., 1995) in raphé pallidus, obscurus, and magnus (Gilbey et al., 1995), with a significant fraction of units exhibiting respirophasic modulation evidenced in a significant fraction of units (19 of 27 neurons). Neuronal modulatory patterns included inspiratory, expiratory, post-inspiratory, and biphasic early-inspiratory and post-inspiratory preferential discharge (Gilbey et al., 1995), recapitulating patterns of respirophasic modulation observed in rostral ventrolateral medullary presympathetic and caudal ventrolateral medullary GABAergic interneurons (Gilbey et al., 1995).

Fig. 6.

Contemporaneously-obtained abdominal, thoracic sympathetic, and phrenic neurograms recordings in the soi-disant “working heart” brainstem preparation during normoxia and chronic intermittent hypoxic conditioning. Raw and integrated ($\int$) activities of abdominal (Abd), thoracic sympathetic (tSNA), and phrenic nerve (PND) discharge during hyperoxic eucapnia (left panels) and chronic intermittent hypoxic-(CIH) conditioning (right panels). The shaded gray area in the neural recordings highlights the late expiratory phase of the respiratory cycle. Note obvious emergence and development of synchronous late-expiratory bursting in the neurograms of both the abdominal and thoracic sympathetic nerve activities (arrows) of rats conditioned with chronic intermittent hypoxia. Modified with permission from Fig. 2 of Zoccal et al. (2008).

Preganglionic and postganglionic sympathetic neuronal spiking variably manifests inspiratory, expiratory, and/or postinspiratory patterns of modulation (Johnson and Gilbey, 1994; Zhou and Gilbey, 1992). In pentobarbital-anesthetized vagotomized rats, upper thoracic preganglionic sympathetic neurons exhibit a higher fraction of cells exhibiting respirophasic modulation compared with lower thoracic and lumbar preganglionic sympathetic neurons. These units exhibit predominantly postinspiratory patterns of modulation (Gilbey et al., 1986; Zhou and Gilbey, 1992). A rostrocaudal gradient in the magnitude of inspiratory and expiratory related respirophasic modulation of sympathetic activity exists along the length of the spinal cord, with differential convergence of inputs to the intermediolateral cell column preganglionic sympathetic neurons varying according to spinal segmental level. Respirophasic modulation of intermediolateral cell column preganglionic sympathetic neurons exhibits homology with unitary discharge patterns exhibited by the intercostal motoneurons supplying thoracic wall respiratory musculature (Gilbey et al., 1986). Cervical sympathetic preganglionic neurons exhibiting spontaneous spiking or activity elicitable by iontophoretically applied glutamate variably exhibit inspiratory or expiratory accentuation of burst amplitude, though some units remain non-modulated (Gilbey et al., 1986).

Spinal sympathoexcitatory neurons receiving convergent inputs from a specific zone within the rostral ventrolateral medulla characteristically span only a few spinal segments, consistent with the somatotopic organization of presympathetic neurons located within this region. Respirophasic modulation conferred upon intermediolateral cell column preganglionic sympathetic neurons likely derives from the composite influences of axodendritic and axosomatic synaptic inputs conveyed by presympathetic neurons residing within the rostral ventrolateral medulla (Baekey et al., 2010; Guyenet et al., 1990; Haselton and Guyenet, 1989; Lipski et al., 1996a, b; McAllen, 1987; Miyawaki et al., 1995; Terui et al., 1986; Zoccal et al., 2008, 2009a, 2009b), caudal raphé (Gilbey et al., 1995), and ventrolateral metencephalic A5 catecholaminergic cell group, propriobulbar interneuronal networks, and somatic thoracic afferent fibers relaying proprioceptive information regarding chest wall muscle contraction. Thoracic wall non-respiratory-related afferents may confer inhibitory or excitatory modulatory synaptic drive upon intermediolateral cell column preganglionic sympathetic units.

Zhou and Gilbey (1992) conducted an elegant study investigating the respirophasic modulation of thoracolumbar myelic preganglionic sympathetic neurons in pentobarbital-anesthetized rats. Investigated perikarya were distributed throughout segments T13-L2 of the ipsilateral intermediolateral cell chain (Zhou and Gilbey, 1992). Antidromic stimulation of the lumbar chain interposed between the 4${}^{th}$ and 5${}^{th}$ lumbar paravertebral ganglia electrophysiologically identified preganglionic sympathetic neurons (Zhou and Gilbey, 1992). Patterns of discharge of thoracolumbar preganglionic sympathetic neurons included inspiratory, early expiratory, and non-segmented expiratory preferential modulation, though a significant fraction of units lacked specific respirophasic modulation (Zhou and Gilbey, 1992). Glutamate iontophoresis elicited the resurgence of a quiescent population of neurons into readily evident neuronal spiking, with a similar proportion and distribution of active units exhibiting specific respirophasic modulation (Zhou and Gilbey, 1992). Preganglionic sympathetic neurons exhibiting respirophasic spiking coordinately demonstrated modulation varying with respect to the cardiac cycle (Zhou and Gilbey, 1992). Preganglionic sympathetic units are located at the T2 spinal level innervated neurons constituting the cervical sympathetic nerve (Zhou and Gilbey, 1992).

Sympathetic neural efferent activity exhibits respirophasic modulation of rhythmic bursting (Fig. 6 and Figs. 8-14) (Koshiya and Guyenet, 1995; Mandel and Schreihofer, 2006; McAllen, 1987; Miyawaki et al., 1996; Seals et al., 1993; St. Croix et al., 1999). McAllen (1987) demonstrated inspiratory modulation of sympathetic neural activity in neurons of the subretrofacial nucleus (i.e., rostral ventrolateral medulla) and cervical sympathetic nerve in chloralose-anesthetized vagotomized cats (McAllen, 1987). Haselton and Guyenet (1989) later demonstrated various patterns of respirophasic modulation of lumbar sympathetic neural efferent activity, including inspiratory modulation, inspiratory inhibition, and early inspiratory inhibition, followed by postinspiratory modulation, in halothane anesthetized rats. Microinjections of the GABA${}_A$ receptor antagonist bicuculline into the rostral tip of the ventrolateral medulla, but not median raphe, coordinately enhanced phrenic nerve burst amplitude and frequency, lumbar sympathetic nerve burst amplitude and extent modulation by phrenic nerve activity, and dynamic mean arterial pressure magnitude and blunted the baroreflex in halothane-anesthetized vagotomized rats. However, microinjections of the glycine receptor antagonist strychnine and the GABA${}_B$ receptor antagonist phaclofen were without effect (Guyenet et al., 1990). Empirically demonstrated patterns of respirophasic modulation of sympathetic bursting have included exclusively postinspiratory accentuation of cervical and lumbar sympathetic nerve activity and preferentially inspiratory accentuation of cardiac, splanchnic, renal, adrenal, and muscle sympathetic nerve activity in vagotomized sinoaortic denervated halothane anesthetized rats (Numao et al., 1987; Seals et al., 1993; St. Croix et al., 1999), inspiratory accentuation of splanchnic sympathetic nerve discharge of vagotomized chloralose-anesthetized rats (Mandel and Schreihofer, 2006), inspiratory accentuation of inferior cardiac sympathetic nerve in chloralose-anesthetized vagotomized rats, and inspiratory and postinspiratory accentuation of splanchnic sympathetic nerve discharge in sinoaortic-denervated vagotomized urethane-anesthetized rats (Koshiya and Guyenet, 1995). These findings collectively indicate significant interstudy heterogeneity of patterns of respirophasic modulation of sympathetic oscillators.

Experimental interventions powerfully modifying respirophasic modulation of the sympathetic nerve discharge inform a thoughtful understanding of mechanisms emergently generating sympathorespiratory coupling (Miyawaki et al., 1996). Microinjections of specific N-methyl-D-aspartate (NMDA) (DL-2-amino-5-phosphonovaleric acid), a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) (6-cyano-7-nitroquinoxaline-2,3-dione), and kynurenic acid-type glutamate receptor antagonists into the rostral ventrolateral medulla abolish postinspiratory modulation of splanchnic sympathetic neural efferent activity in pentobarbital anesthetized rats, without modifying inspiratory modulation of efferent discharge persisted (Miyawaki et al., 1996). Postinspiratory modulation of lumbar sympathetic neural efferent activity may be augmented by Hypercapnia, and pharmacological antagonism of GABA, but not GABA${}_B$ or glycine, receptor modulated signaling in the rostral ventrolateral medulla abolished postinspiratory modulation of lumbar sympathetic neural efferent activity (Miyawaki et al., 1996). Postinspiratory modulation of lumbar sympathetic neural efferent activity decreased in parallel with dynamic increases of arterial pressure magnitude, indicating baroinhibition of respirophasic modulation of sympathetic neural efferent activity, and in response to microinjections of the nonselective glutamate receptor antagonist kynurenic acid, in halothane-anesthetized vagotomized rats. These findings corroborate fast excitatory glutamate receptor modulated signaling that contribute to sympathorespiratory modulation and contribute to tonic GABA${}_A$ and glycine-mediated fast inhibitory mechanisms in modulating coupling amongst the sympathetic and respiratory discharges (Miyawaki et al., 1996).

Respirophasic modulation of sympathetic efferent neural discharge exhibits heterogeneity across different preparation types (Koshiya and Guyenet, 1995; Mandel and Schreihofer, 2006; McAllen, 1987; Miyawaki et al., 1996; Numao et al., 1987; Seals et al., 1993; Haselton and Guyenet, 1989; St. Croix et al., 1999). Differences among patterns of respirophasic modulation across studies could putatively stem from differences in vagal continuity and whether animals are spontaneously-breathing versus mechanically-ventilated (Koshiya and Guyenet, 1995; Mandel and Schreihofer, 2006; McAllen, 1987; Numao et al., 1987; Seals et al., 1993; St. Croix et al., 1999), powerfully influencing discharge patterns and properties of slow and fast adapting pulmonary stretch receptors and imposition of these oscillatory inputs upon nucleus tractus solitarius units (Fenik, 1992; Preobrazhenskiĭ et al., 1989). Respirophasic modulation of sympathetic neural efferent activity may alternately arise from the overlap of, and interaction between, Bötzinger complex propriobulbar interneuronal decrementing postinspiratory and augmenting late-expiratory units with rostral ventrolateral medullary presympathetic cells (Moraes et al., 2014a) or pulmonary stretch receptor influences upon sympathetic oscillators. Bötzinger complex decrementing post-inspiratory units driving respirophasic modulation of sympathetic oscillations may natively entrain this coupling, or be driven by, pulmonary stretch receptor mechanosensitive (Preobrazhenskiĭ et al., 1989) or barosensitive nucleus tractus solitarius units (Fig. 5) (Rogers et al., 2000).

Fig. 7.

Intracellular recordings of Bötzinger complex augmenting late-expiratory neurons. A: Phrenic neurogram and intracellular recordings of Bötzinger complex augmenting expiratory units reveal oscillations of membrane voltage with augmenting action potential spiking frequency during the late expiratory phase in normotensive (Aa) and spontaneously hypertensive (Ab) rats, Superimposition of Bötzinger complex augmenting late-expiratory intracellular recordings in normotensive and spontaneously hypertensive rats demonstrates phase-advance of neuronal spiking earlier into expiration (Ac), putatively implying Botzinger complex postinspiratory units may reciprocally exhibit less amplitude in the hypertensive condition. B: Antagonism of fast inhibitory synaptic neurotransmission failed to differentially modify neuronal input resistance in normotensive (Ba) or spontaneously hypertensive (Bb) rats, graphically depicted in (Bc). Delivery of electrical pulses in current-clamp mode ranging from 25 to 100 picoamperes enhanced Bötzinger complex augmenting late-expiratory neuronal spiking frequency in normotensive (Ca) and spontaneously hypertensive (Cb) rats. However, the two groups failed to exhibit differences in neuronal spiking-current density slope (Cc). Visual inspection of panels Ba, Bb, Ca, and Cb appear to evidence somewhat regular oscillations of neuronal spiking frequency (alternating thick and thin spike clusters) ~ 1.7 Hz, perhaps representing slow sympathetic oscillations distributing to Bötzinger complex neurons. Spike height also exhibits prominent oscillations in panels Ba and Bb. Ia, current; M$\mathrm{\Omega }$, megaohms, mV, microvolts; pA, picoamperes; Ri, internal resistance; SH, spontaneously hypertensive. Modified with permission from Fig. 4 of Moraes et al. (2014a).

Sympathetic oscillators receive the powerful modulatory synaptic drive from barosensitive nucleus tractus solitarius units (Fig. 5) (Mendelowitz, 1999). This constitutes a negative feedback mechanism regulating dynamic arterial pressure magnitude, dumping of spurts of blood into the microcirculation, and tissue perfusion dynamics (Alexander and Cuir, 1963). Dynamic arterial pressure rises to elicit sinoatrial deceleration and corrective attenuation of arteriolar and venular tone via enhanced oscillatory baroreceptor spiking frequency, coordinated by reductions in the sympathetic drive to the heart and vessels and increases of ambiguual cardiovagal premotoneuronal drive to electroconductive circuitry of the heart and atrial myocardium (Rogers et al., 2000). Baroreceptors sensitive to dynamic arterial stretch are most densely located within the aortic arch and carotid sinus, though may be found sparsely distributed throughout middle internal carotid (Toorop et al., 2013), vertebral (Agadzhanian and Kupriianov, 2008; Kupriyanov, 2009), and subclavian arteries (Chevalier-Cholat and Friggi, 1976, 1977; Kidd et al., 1974). Baroreceptor and baroafferent discharge frequency covary with dynamic arterial pressure magnitude. Oscillatory baroreceptor inputs centrally-conveyed to medial and interstitial nucleus tractus solitarius units via the aortic depressor nerve, carotid sinus nerve, and distributed afferent fibers join cranial nerves en route to myelencephalic oscillators may coordinately receive inputs from Hering’s nerve and vagal afferents relaying fast oscillatory chemoreceptor inputs and vagal afferents relaying slow oscillatory alveolar stretch. Multimodal barosensitive, chemosensitive, and/or pulmonary stretch mechanosensitive units residing within the nucleus tractus solitarius may constitute a mechanism of respiratory gating and multiplicatively synergistic chemoreceptor augmentation of oscillatory baroreceptor inputs to higher-order neural networks. Nucleus tractus solitarius propriobulbar interneuron microcircuit oscillatory convey synaptic drive to caudal ventrolateral medullary GABAergic units mediating state-dependent respirophasic and “baro-phasic” inhibition of rostral ventrolateral medullary presymapthetic neurons, eliciting contemporaneous and parallel amplification of ambiguual cardiovagal premotoneuronal spiking (Ghali, 2018, 2017a; Guyenet, 2006). The magnitude of sinoatrial deceleration in response to a given incremental augmentation of dynamic arterial pressure magnitude constitutes the chief metric characterizing baroreflex gain (Rogers et al., 2000), powerfully crossmodally modulated by neural respiratory rhythmicity and dynamic changes in pulmonary stretch (Baekey et al., 2010), along with spectral peaks nestled within dynamic arterial pressure magnitude and cardiac interval. The magnitude of sinoatrial deceleration elicited by a fixed rise of dynamic arterial pressure varies with respect to respiratory cycle phase (Baekey et al., 2010), implying phase-dependent differential stimulability or inhibitability of sympathetic and cardiovagal preganglionic premotoneuronal spiking, with peak unitary discharge frequency occurring during the period which they are most excitable (Rogers et al., 2000). Sinoatrial decelerations elicited by the baroreflex are preferentially more prominent during the early expiratory epoch when afferents are most activated (Gilbey et al., 1984).

Expiratory-related neurons exhibiting arterial pulse modulation indicates baromodulation of the expiratory epoch and supports barostimulation facilitates expiratory-related neural activity. Lindsey et al. (1998) demonstrated the inspiratory phase contracts in response to chronically elevated dynamic arterial pressure magnitude. Metencephalic zones gate baroreceptor influences upon the propriobulbar interneuronal microcircuit oscillators generating breathing. Accordingly, barosensitive nucleus tractus solitarius units receive modulatory synaptic drive preferentially exhibiting higher spiking frequency during the inspiratory epoch from neurons constituting dorsolateral metencephalic nuclei (Molkov et al., 2014). Electrical stimulation of propriobulbar interneuronal microcircuit oscillators constituting the Kölliker-Fuse nucleus reduces the barosensitivity of monomodally or multimodally sensitive nucleus tractus solitarius cells in response to carotid sinus oscillatory inputs. Enhanced metencephalic neuronal spiking frequency specifically reduces baroreceptor gain during the inspiratory epoch (Brunner et al., 1982; Dove and Katona, 1985; Li et al., 1999a, b; Lindsey et al., 1998). These findings illustrate propriobulbar interneuronal microcircuit oscillator inputs to sympathetic oscillators, and ambiguual cardiovagal premotoneurons constitute critical mechanisms mediating sympathorespiratory modulation (Molkov et al., 2017).

Fig. 12.

Effects of pontomedullary transection on respirophasic modulation of thoracic sympathetic nerve activity. Mechanical transection between the metencephalon and myelencephalon abolishes respirophasic modulation of thoracic sympathetic (thSNA) nerve bursting in the in situ arterially-perfused preparation of the unanesthetized decerebrate juvenile rat. Upper panels: Thoracic sympathetic nerve bursting evidence late-inspiratory and postinspiratory accentuation during normoxic normocapnia in the presence of pontomedullary continuity. Prior to pontomedullary transection (intact pons; upper panels), integrated ($\int$) and raw thoracic sympathetic nerve activity exhibit a clear respirophasic modulation. Lower panels: Encephalotomy between the metencephalon and myelencephalon abolishes or significantly attenuates respirophasic modulation of raw and integrated ($\int$) thoracic sympathetic nerve activity. Top to bottom: integrated ($\int$) and raw thoracic sympathetic nerve activity (thSNA); integrated ($\int$) and raw phrenic nerve activity (PNA). Modified with permission from Fig. 1 of Molkov et al. (2014).

Respiratory-related unitary spiking subjected to baromodulation constitutes a mechanism emergently generating sympathorespiratory coupling (see Lindsey et al. (1998)). Extracellular recordings of brainstem respiratory-related units exhibit pulse-synchronous modulation in vagotomized unanesthetized decerebrate cats (Lindsey et al., 1998). Higher unitary discharge frequency during the expiratory epoch parallels augmentations of dynamic arterial pressure magnitude (Lindsey et al., 1998). Oscillatory baroreceptor inputs modulate respiratory-related neuronal spiking occurring in response to dynamic changes of arterial pressure magnitude, coordinately with evidence provided validating supraspinal mechanisms generate the described coupling (Morrison et al., 1984). Accordingly, in the unanesthetized decerebrate juvenile cat, vagotomy significantly reduces the pulse-synchronous modulation of pontine respiratory-related neurons by the arterial pressure waveform, contemporaneously amplifying the augmenting inspiratory pattern of respirophasic modulation of the units (Dick et al., 2009). Barostimulation elicits barosensitive medial and interstitial glutamatergic nucleus tractus solitarius neuronal spiking, which reduces sympathetic activity by activating a population of GABAergic interneurons residing within the caudal ventrolateral medulla conveying inhibitory axodendritic and axosomatic drive to rostral ventrolateral medullary propriobulbar interneuronal microcircuit oscillators and interneuronally-coupled presympathetic bulbospinal cells (Dampney, 1994; Ghali, 2018, 2017a; Goodchild and Moon, 2009; Guyenet et al., 1990). As a parallel pathway, barosensitive glutamatergic nucleus tractus solitarius units convey excitatory synaptic drive to Bötzinger complex glycinergic decrementing postinspiratory neurons, mediating coordinate enhancement of rostral ventrolateral medullary presympathetic and Bötzinger complex expiratory neuronal spiking, imposing baromodulation upon the respiratory rhythm and mediating sympatho-expiratory coupling (Baekey et al., 2010; Molkov et al., 2014).

Barostimuli elicits differential effects upon neural breathing pattern, whether applied during the inspiratory versus expiratory epochs, effects prominently demonstrated in the in situ arterially-perfused preparation of the unanesthetized decerebrate juvenile rat (Figs. 13 and 14) (Baekey et al., 2010). The behavior elucidates an understanding of interactions amongst oscillatory baroreceptor inputs with propriobulbar interneuronal microcircuit oscillators generating breathing and indicate the latter gate the former. Barostimuli delivered during neural inspiration often fail to modify phrenic burst pattern, though successfully eliminate respirophasic modulation of sympathetic neural activity (Baekey et al., 2008). In contrast, barostimuli delivered during neural postinspiration and late expiration augment expiratory neural activity and eliminate respirophasic modulation of sympathetic neural efferent discharge (Baekey et al., 2008), supporting a model predicated upon prominent inhibition conveyed to Bötzinger complex decrementing postinspiratory neurons by decrementing early inspiratory neurons residing within the and rostral ventral respiratory group during neural inspiration (Molkov et al., 2017; Smith et al., 2007), preventing barostimuli delivered during inspiration from eliciting modifications of phrenic nerve bursting.

Baekey et al. (2010) present a characterizing model mechanism underlying the baromodulation of propriobulbar interneuronal microcircuit oscillators generating neural breathing. According to this model, predicated upon reciprocal interaction of decrementing postinspiratory neurons with augmenting expiratory neurons, two classes of barosensitive nucleus tractus solitarius neurons included one group conveying synaptic drive to caudal ventrolateral medullary GABAergic interneurons and another group conveying synaptic drive to metencephalic-dependent post-inspiratory units. Dynamically incrementing postinspiratory neuronal spiking frequency parallels correlated reductions of augmenting late-expiratory neuronal spiking frequency. Barostimuli delivered during the expiratory epoch coordinately elicit resurgence of postinspiratory neuronal spiking frequency commensurate with dynamic reductions of discharge precluding waning activity and attenuating augmenting late-expiratory neuronal spiking, in essence resetting the expiratory epoch. An excitatory population of Bötzinger complex expiratory-related neurons, putatively glutamatergic, may convey direct axodendritic and axosomatic synaptic drive to rostral ventrolateral medullary neurons (Molkov et al., 2014). Bötzinger complex decrementing postinspiratory neurons are released from inhibitory modulation by preBötzinger complex decrementing early inspiratory neurons and rostral ventral respiratory group inspiratory neurons during early neural expiration (Smith et al., 2007), in effect coordinately amplifying neural expiratory activity and attenuating respirophasic modulation of sympathetic discharge (Baekey et al., 2010). Bötzinger complex decrementing postinspiratory neurons (see Ezure and Manabe (1988)) convey excitatory and/or inhibitory modulation to Bötzinger complex augmenting late expiratory neurons (Smith et al., 2009), preBötzinger complex pre-inspiratory, pre-inspiratory-inspiratory phase-spanning, and decrementing early-inspiratory units (Molkov et al., 2010), and rostral ventrolateral medullary presympathetic cells (Molkov et al., 2014; Moraes et al., 2012c; Sun et al., 1997).In contrast, Bötzinger complex augmenting late-expiratory neurons making direct synaptic contacts with adjacent rostral ventrolateral medullary cells (Moraes et al., 2012c; Sun et al., 1997) augment presympathetic bulbospinal neuronal spiking during late expiration, constituting a chief mechanism mediating sympathorespiratory coupling and strengthening the same in response to hypercapnia (Molkov et al., 2014) and acute and chronic intermittent hypoxia (Moraes et al., 2012b, 2014b; Zoccal et al., 2007, 2008, 2009a).

Several mechanisms conspire to emergently generate sympathorespiratory, parasympathorespiratory, and sympathoparasympathetic coupling at rest and during hypercapnic and hypoxic conditions (Molkov et al., 2017; Moraes et al., 2012a, b, c, 2014a, 2014b; Zoccal et al., 2007, 2008, 2009a, 2009b; Zoccal and Machado, 2011; Zoccal, 2015). Common inputs may generate Autonomorespiratory coupling conveyed to propriobulbar interneuronal microcircuit oscillators constituting the breathing pattern generator, sympathetic oscillators, and ambiguual cardiovagal premotoneurons, i.e., RTN/pFRG, see Abdala et al. (2009), coordinate and heterologous regulation by common ascending inputs relaying peripheral oscillatory baroreceptor, chemoreceptor, and pulmonary stretch receptor inputs (Takakura et al., 2007), and direct neurosynaptic interactions amongst overlapping propriobulbar interneuronal microcircuit oscillators of the constituent generators (i.e., Bötzinger complex and rostral ventrolateral medulla, see Moraes et al. (2012c) and Sun et al. (1997)). The presented set of conjectures and hypotheses carry equivalent empirical validity and naturally evolve the construction of an elegant model recapitulating and elucidating the mechanistic behavior of experimentally-evident autonomorespiratory coupling (Molkov et al., 2014).

Bötzinger complex decrementing post-inspiratory (Ezure and Manabe, 1988) and augmenting late-expiratory (Jean, 1992) neurons overlap with rostral ventrolateral medullary catecholaminergic, and glutamatergic presympathetic bulbospinal units (Ghali, 2018, 2017a; Guyenet, 2006; Guyenet et al., 2018) and pre-Bötzinger complex preinspiratory, preinspiratory-inspiratory phase spanning, and decrementing early inspiratory units (Molkov et al., 2017; Morgado-Valle and Beltran-Parrazal, 2017) overlap with caudal ventrolateral medullary GABAergic interneurons (Goodchild and Moon, 2009; Mandel and Schreihofer, 2006; Paxinos et al., 1985). Excitatory and inhibitory neurosynaptic interactions coupling these cells emergently generates and strengthens sympathorespiratory coupling (Marchenko et al., 2016; Molkov et al., 2014). Thus, the relative balance of excitatory and inhibitory drive conveyed by, and synaptic contacts with, intermingled Bötzinger complex glycinergic decrementing postinspiratory, glutamatergic decrementing post-inspiratory, and GABAergic augmenting late expiratory propriobulbar neuronal circuitry may chiefly determine the extent of respirophasic modulation of rostral ventrolateral medullary neuronal spiking (Dergacheva et al., 2010; Farmer et al., 2016). Dynamic changes of Bötzinger complex postinspiratory neuronal spiking frequency (see Ezure and Manabe (1988)) modifies rostral ventrolateral medullary neuronal spiking frequency during the postinspiratory and late expiratory epochs of breathing (see Moraes et al. (2012c) and Sun et al. (1997)). We suggest these direct neurosynaptic interactions and their proximally-related coordinately coupling contribute to emergently generating preferential late-expiratory bursting synchronized across abdominal and thoracic sympathetic neural efferent bursting during hypercapnia (Molkov et al., 2014) and acute and chronic intermittent hypoxia (Moraes et al., 2012b, 2014b; Zoccal et al., 2007, 2008, 2009a). During normocapnic normoxia, Bötzinger complex augmenting late-expiratory units restrict Bötzinger complex decrementing postinspiratory spiking (see classic models of Molkov et al. (2014, 2017) and Smith et al. (2007, 2009)), disinhibiting rostral ventrolateral medullary spiking during the postinspiratory epoch (Moraes et al., 2012c; Sun et al., 1997) and conferring postinspiratory preferential modulation of rostral ventrolateral medullary cells across the majority of studies (Mandel and Schreihofer, 2006; Numao et al., 1987; Seals et al., 1993; St. Croix et al., 1999; Koshiya and Guyenet, 1995). Excitatory and/or inhibitory populations of pre-Bötzinger complex post-inspiratory neurons (Ausborn et al., 2018) may contribute to sympathorespiratory coupling amongst rostral ventrolateral medullary presympathetic neurons through modulation of GABAergic and/or glutamatergic interneurons residing within the caudal ventrolateral medulla.

Fig. 13.

Dynamic changes of arterial pressure magnitude elicit coordinate modification of phrenic and sympathetic neural bursting. A: Barostimuli applied during the inspiratory, postinspiratory, or late expiratory epochs in the in situ arterially-perfused unanesthetized decerebrate juvenile rat elicit differential effects upon neural breathing pattern. Barostimuli delivered during the inspiratory epoch failed to modify breathing rhythm or burst parameters. However, barostimuli applied during the postinspiratory, and late-expiratory epochs elicited robust amplification of expiratory epoch duration. Following mechanical transection separating the metencephalon from the myelencephalon, a barostimulus applied during late expiration contracted apneustic inspiratory bursts in the phrenic nerve discharge. Barostimuli most powerfully elicits reductions of sympathetic neural efferent activity when applied during inspiration, putatively consequent to preferentially inspiratory accentuation of sympathetic neural bursting during resting eucapnic hyperoxia and reduced excitatory synaptic drive from the glutamatergic excitatory population of Bötzinger complex decrementing postinspiratory neurons. $\int$SNA, integrated sympathetic nerve activity (mV, millivolts); $\int$PNA, integrated phrenic nerve activity (V, Volts); PP, perfusion pressure (mmHg); Timescale bar is shown in the lower right-hand corner (5 seconds). B: Computational model simulation of baroreceptor stimulus applied during inspiratory and expiratory epochs in preparations with theoretically preserved pontomedullary continuity intact model and during the inspiratory epoch following removal of the metencephalic compartment. The effects recapitulate empirically-observed respirophasic-variant influences of oscillatory baroreceptor inputs upon the breathing pattern. A simulated barostimulus presented during the inspiratory epoch with preserved pontomedullary continuity negligibly modifies breathing rhythm parameters, though effectively reduces sympathetic neural bursting. A simulated barostimulus applied during the late expiratory epoch elicits modest enhancement of expiratory epoch duration with reflex augmentation of amplitude and duration of ensuing inspiratory bursts coordinately with attenuation of sympathetic neural activity. Reflex augmentation of inspiratory bursting succeeds expiratory facilitation elicited by a barostimulus applied during late expiration, putatively resulting from recovery and resetting of phrenic motoneuronal membrane voltage trajectories and replenishing of neuronal metabolic adenosine triphosphate stores, permitting successively conveyed bulbospinal axodendritic and axosomatic synaptic drive conveyed from rostral ventral respiratory group augmenting inspiratory neurons to elicits early-inspiratory high-frequency oscillations exhibiting more robust coherence and synchrony amongst the population of phrenic motoneurons. Removal of the metencephalic compartment abolishes respirophasic modulation of sympathetic neural activity in the model, with barostimuli delivered during inspiration contracting apneustic inspiratory epochs and reducing sympathetic neural activity. Barostimuli applied during inspiration, and late expiration elicits similar reductions in sympathetic neural activity in preparations with preserved pontomedullary continuity. SNA; sympathetic nerve activity (amplitude normalized from 0 to 1); PNA, phrenic nerve activity (amplitude normalized from 0 to 1). Modified with permission from Fig. 3 of Molkov et al. (2014).

Hypercapnic stimulation of Phox2b transcription factor-expressing chemosensitive glutamatergic retrotrapezoid nucleus neurons (Guyenet et al., 2019) provides tonic drive to Bötzinger complex augmenting late-expiratory units (Molkov et al., 2017), which convey reciprocal inhibitory axodendritic and axosomatic synaptic drive to Bötzinger complex decrementing post-inspiratory neurons, according to the majority of modelers (Molkov et al., 2017; Smith et al., 2009). Experimental excitatory stimulation of propriobulbar interneuronal microcircuits oscillators residing within the Kölliker-Fuse and medial parabrachial nuclei variably generate excitation or inhibition of rostral ventrolateral medullary presympathetic neuronal activity (Baekey et al., 2010; Molkov et al., 2011; Smith et al., 2009). Accentuation or attenuation of either of these sets of Bötzinger complex expiratory populations coordinately modulates the discharge of rostral ventrolateral medullary neuronal spiking directly and through coordinate synaptic inputs from the originate drivers (Moraes et al., 2012c; Sun et al., 1997). Bötzinger complex and rostral ventrolateral medullary propriobulbar interneuronal microcircuit oscillators receive coordinate synchronizing axodendritic and/or axosomatic synaptic drive deriving from higher-order structures, including retrotrapezoid nucleus neurons (Guyenet et al., 2019), parafacial respiratory group pre-inspiratory cells (Onimaru et al., 2018), Kölliker-Fuse, medial, and lateral parabrachial nuclei (Dutschmann and Herbert, 2006; Mörschel and Dutschmann, 2009), and periaqueductal gray matter (Martin and Booker, 1878; Subramanian and Holstege, 2014), among other structures distributed throughout the telencephalon, diencephalon, brainstem, and cerebellum (Horn and Waldrop, 1997). Oscillatory baroreceptor, chemoreceptor, and pulmonary mechanoreceptor stimuli relaying through nucleus tractus solitarius units (Marchenko and Sapru, 2000; Moreira et al., 2007; Rogers et al., 2000; Takakura et al., 2007) and ascending spinoreticular pathways may also coordinately modulate Bötzinger complex and rostral ventrolateral medullary propriobulbar interneuronal microcircuit oscillator discharge.

Caudal ventrolateral medullary GABAergic interneuronal spiking imposes respirophasic modulation upon rostral ventrolateral medullary presympathetic neurons (Mandel and Schreihofer, 2006). PreBötzinger complex inhibitory decrementing early inspiratory neurons (see Molkov et al. (2017) and Smith et al. (2009)) may generate inhibitory postsynaptic currents observed in rostral and caudal ventrolateral medullary cells (Molkov et al., 2014). Populations of preBötzinger complex units alternately expressing GABAergic (Ellenberger, 1999) and glycinergic (Morgado-Valle et al., 2010) neurochemical phenotypes. Analogously inhibitory axodendritic and axosomatic drive may contribute to generating fast synchronous oscillatory coupling and synchrony, clearly manifest in phrenic nerve discharge spectra of the in situ arterially-perfused preparation of the unanesthetized decerebrate juvenile rat and in vivo unanesthetized decerebrate preparations of the adult rat (Ghali and Marchenko, 2013; Marchenko and Rogers, 2006a, b, 2007; Marchenko et al., 2012). Marchenko and Rogers (2009) provide precedence indicating the critical contribution of fast inhibitory neurosynaptic transmission in modulating the properties of fast sympathetic rhythms. Phrenic nucleus microinjections of gabazine or strychnine elicit severe perturbations of fast oscillatory synchrony in population phrenic nerve motor output Marchenko and Rogers (2009). Strychnine microinjections in the phrenic nucleus robustly amplified phrenic nerve activity during the postinspiratory phase, suggesting local glycinergic mechanisms restrict post-inspiratory activity in phrenic neural efferent discharge. However, the effects of gabazine were less phase-restricted in amplifying phrenic nerve activity across the respiratory cycle. Interactions at the level of nucleus tractus solitarius units (Jean, 1992) inform novel mechanisms of sympathorespiratory and parasympathorespiratory coupling. Oscillatory baroreceptor inputs transduced by pressure-sensitive receptors exhibiting highest density within the carotid sinus and aortic arch, though diffusely distributed throughout the internal carotid (Toorop et al., 2013), vertebral (Agadzhanian and Kupriianov, 2008; Kupriyanov, 2009), and subclavian arteries (Chevalier-Cholat and Friggi, 1976, 1977; Kidd et al., 1974) relays centrally to neurons residing within the medial and interstitial divisions of the nucleus tractus solitarius covarying dynamically with changes of arterial pressure magnitude (Rogers et al., 2000). Spatiotemporal bursting dynamics of integrated baroreceptor discharge frequency assumes the form of a sine wave. Oscillatory baroreceptors inputs emergently generate sympathorespiratory coupling by modulating overlapping Botzinger complex and rostral ventrolateral medullary propriobulbar interneuronal microcircuit oscillators (Moraes et al., 2012c; Sun et al., 1997) through GABAergic caudal ventrolateral medullary interneurons (Guyenet, 2006; Guyenet et al., 2019; Mandel and Schreihofer, 2006).

Mechanosensitive pulmonary stretch receptors residing within the alveolar interstitium may be electrophysiologically segregated into slow (Fenik, 1992; Preobrazhenskiĭ and Fenik, 1989) and fast (Zhdanov, 1984) adapting types of stretch receptors, with discharge frequency nonlinearly covarying according to an absolute, or static or dynamic rate of change of, pulmonary stretch (Preobrazhenskiĭ and Fenik, 1989). Slow adapting pulmonary stretch receptors chiefly sensitive to the absolute magnitude of stretch generate complex spike trains during the phrenic burst (Chen et al., 2008, 2010, 2011; Marchenko and Rogers, 2007b), compared with fast adapting pulmonary stretch receptors, which possess rapid desensitization kinetics and are thus specifically tuned to detecting the onset and offset of the inspiratory epoch (Zhdanov, 1984). Spike trains of these receptors relay centrally via pulmonary stretch mechanosensitive vagal afferents to pump cells chiefly residing in the ventrolateral division of the nucleus tractus solitarius (Moreira et al., 2007; Takakura et al., 2007). Higher-order nucleus tractus solitarius units may alternately receive the monomodal or multimodal convergent somatodendritic synaptic drive from afferents relaying oscillatory inputs from baroreceptors (Rogers et al., 2000), chemoreceptors (Takakura et al., 2007), or pulmonary stretch mechanoreceptors (Moreira et al., 2007; Li et al., 1999a, b; Takakura et al., 2007). Rises of dynamic arterial pressure magnitude occurring during the inspiratory epoch elicit multiplicative amplification of neuronal spiking frequency of multimodal nucleus tractus solitarius units, contrasted with rises of dynamic arterial pressure magnitude during expiration which elicit comparatively more attenuated augmentations of NTS neuronal spiking frequency (Rogers et al., 2000).

These findings collectively mechanistically indicate oscillatory mechanosensory pulmonary stretch spike cell trains heterosynaptically gates oscillatory baroreceptor inputs to NTS units. Reciprocally, oscillatory baroreceptor spike cell trains heterosynaptically gates oscillatory pulmonary stretch inputs to NTS units (Baekey et al., 2010). More simply stated, neural inspiratory activity gates baroreceptor oscillations and dynamic arterial pressure magnitude gates pulmonary stretch receptor spike trains conveyed to nucleus tractus solitarius units (Baekey et al., 2010). Glutamatergic barosensitive nucleus tractus solitarius units convey excitatory axodendritic and axosomatic synaptic drive to caudal ventrolateral medullary GABAergic propriobulbar interneurons and may heterosynaptically blunt the discharge of glutamatergic chemosensitive retrotrapezoid nucleus units (Guyenet, 2006), conferring oscillatory baroinhibition of rostral ventrolateral medullary presympathetic unitary spiking (Mandel and Schreihofer, 2006). By this mechanism, baroreceptor oscillations modulate the sympathetic neural efferent output (Diedrich et al., 2009; Ghali, 2018, 2017a; Guyenet et al., 2019). Critically, barosensitive nucleus tractus solitarius unitary spiking coordinately influences cardiovagal premotoneuronal spiking within the dorsal motor nucleus of the vagus and nucleus ambiguus innervating dendrites and somata of postganglionic parasympathetic cholinergic neurons in the cardiac plexus (Eckberg, 2003), accordingly modulating sinoatrial action potential frequency, atrioventricular conductivity, and myocardial contractility. Oscillatory fluctuations of arterial pressure, thus confer an oscillatory fluctuation upon the cardiac interval (Rogers et al., 2000), influencing spectral variability of dynamic arterial pressure magnitude.

A host of mechanisms reflecting the overall state and condition of the breathing generator, sympathetic oscillators, and ambiguual cardiovagal premotoneurons impose respirophasic modulation upon cardiovagal premotoneuronal spiking (Dergacheva et al., 2010; Farmer et al., 2016). The Bötzinger complex contains decrementing postinspiratory neurons exhibiting a predominantly glycinergic neurochemical repertoire (Ezure and Manabe, 1988), though a few of these units also exhibit a glutamatergic phenotype (Molkov et al., 2017; Smith et al., 2007). Bötzinger complex glycinergic and glutamatergic decrementing postinspiratory modulate cardiovagal premotoneuronal membrane voltage trajectories during the early expiratory epoch. The relative balance and weighing of synaptic inputs conveyed by these cells determine whether afferent influences generate predominantly expiratory facilitation or inhibition of efferent cardiovagal premotoneurons residing within the nucleus ambiguus and dorsal motor vagal nucleus (Dergacheva et al., 2010; Farmer et al., 2016). Direct synaptic drive conveyed by Bötzinger complex GABAergic cells to ambiguual cardiovagal premotoneurons (Dergacheva et al., 2010; Farmer et al., 2016) may represent a parallel mechanism generating attenuation of cardiovagal premotoneuronal tone and reciprocal sinoatrial accelerations.

Bötzinger complex decrementing postinspiratory neurons coordinately receive excitatory axodendritic and axosomatic drive from Kölliker-Fuse and medial parabrachial nucleus propriobulbar units (Dutschmann and Herbert, 2006; Mörschel and Dutschmann, 2009) and monosynaptic direct and polysynaptic indirect modulation conferred by monomodal or multimodal nucleus tractus solitarius units (Molkov et al., 2017; Smith et al., 2007) (Fig. 5). Bötzinger complex decrementing postinspiratory neurons modulate normal triphasic eupnea by segmenting the expiratory epoch through reciprocal inhibitory neurosynaptic interactions with Bötzinger complex augmenting late-expiratory neurons and via the provision of prominent inhibitory modulatory influences upon preBötzinger complex preinspiratory, preinspiratory inspiratory phase spanning and decrementing early inspiratory neurons (Molkov et al., 2017; Rybak et al., 2014; Smith et al., 2007). The direct overlap of the Bötzinger complex with rostral ventrolateral medullary propriobulbar and bulbospinal circuitry provides a neuroanatomic substrate generating common sympathorespiratory coupling and crossmodal modulation amongst these oscillatory zones (Moraes et al., 2014a, b).

According to the described neuroanatomic organization and patterns of electrophysiological interactions, any peripheral stimulus or set of influences modulating nucleus tractus solitarius unitary spiking frequency (Chen et al., 2011; Moreira et al., 2007; Rogers et al., 2000; Takakura et al., 2007) may coordinately modulate cardiovagal premotoneuronal (Bishop, 1974; Lindsey et al., 1998) and rostral ventrolateral medullary presympathetic unitary spiking (Moraes et al., 2012c; Sun et al., 1997) through axodendritic and axosomatic synaptic drive conveyed to Bötzinger complex neurons (Molkov et al., 2014). However, the specific type of stimulus modulating nucleus tractus solitarius neuronal spiking differential influences pattern of modulation conveyed upon propriobulbar interneuronal microcircuit oscillators generating breathing, sympathetic activity, and cardiovagal premotoneuronal discharge (Baekey et al., 2010; Lindsey et al., 1998). The magnitude and dynamic rates of changes of alveolar stretch detected by slow and fast adapting pulmonary stretch receptors (Chen et al., 2011) conveyed centrally through vagal afferents to the ventrolateral and interstitial divisions of the nucleus tractus solitarius (Marchenko and Sapru, 2000) commensurately modulate Bötzinger complex decrementing postinspiratory and augmenting late expiratory neuronal spiking to disfacilitates inspiration and enhance expiration (Molkov et al., 2017) and rostral ventrolateral medullary presympathetic units to generate sympathoinhibitory effects via the soi-disant sympathetic arm of the Hering Breuer reflex (Guyenet, 2006) and sinoatrial acceleration through disfacilitation of cardiovagal premotoneurons, consequent to inhibitory axodendritic and axosomatic modulation by Bötzinger complex GABAergic propriobulbar interneuronal microcircuit oscillators (Fenik, 1992; Preobrazhenskiĭ and Fenik, 1989; Zhdanov, 1984; Zhdanov et al., 1989).

Fig. 14.

Mechanism mediating resetting of expiration elicited by transient barostimuli. A: Computational modeling prediction of Bötzinger complex postinspiratory neuronal and phrenic nerve discharge to a barostimulus applied during the late expiratory epoch. Bötzinger complex post-inspiratory and augmenting late-expiratory neuron membrane voltage trajectories (first and third traces) and action potential spiking among both neuronal populations (second and fourth traces) are depicted. Shaded intervals within segregable respiratory epochs highlight inspiratory, early expiratory (E1), and late expiratory (E2) epochs. Barostimulation resets expiration by activating populations of Bötzinger complex postinspiratory units neurosynaptically inhibiting Bötzinger complex augmenting late-expiratory cells. Delivery of a barostimulus amid late expiration resurrects augmenting expiratory neuronal spiking and elicits a second complete expiratory effort. B: Extracellular recordings (first among each pair of upper and middle traces) and unitary spiking histograms (second among each pair of upper and middle traces) of Bötzinger complex decrementing post-inspiratory (upper pair of traces) and augmenting late expiratory (middle pair of traces) neurons. Integrated phrenic nerve activity ($\int$PNA) and perfusion pressure (PP) are successively depicted in the lower pair of traces. Barostimuli delivered during late expiration generate expiratory resetting in the in situ arterially-perfused preparation of the unanesthetized decerebrate juvenile rat. Barostimuli delivered during the late-expiratory epoch in the in situ preparation of the unanesthetized decerebrate juvenile rat enhances Bötzinger complex post-inspiratory neuronal spiking frequency and reduce Bötzinger complex augmenting late-expiratory activity, recapitulating model predictions. Modified with permission from Fig. 4 of Molkov et al. (2014).

Chemosensory stimulation of commissural nucleus tractus solitarius units coordinately enhances Bötzinger complex expiratory neuronal discharge and rostral ventrolateral medullary presympathetic unitary activity (Michelini, 1994; Moreira et al., 2007; Rogers et al., 2000; Schwaber et al., 1993; Takakura et al., 2007). Accordingly, hypercapnia or acute or chronic intermittent hypoxia enhances glutamatergic excitatory and GABAergic inhibitory modulation conveyed to dorsal medullary cardiovagal premotoneurons (Dergacheva et al., 2010; Farmer et al., 2016) permitting sinoatrial acceleration elicited by sympathoexcitation (see Guyenet et al. (2018, 2019)) to support chemoreceptor-driven amplifications of myocardial output and dynamic arterial pressure magnitude. Factors influencing baroreflex gain and sensitivity (see Rogers et al. (2000)) may modify patterns of modulation of Bötzinger complex decrementing post-inspiratory and augmenting late-expiratory neurons by chemosensitive commissural nucleus tractus solitarius neuronal synaptic inputs. Chemosensitive glutamatergic retrotrapezoid nucleus units convey diffuse tonic excitatory drive to the Bötzinger complex, chiefly augmenting late-expiratory neurons, and receive prominent GABAergic modulatory synaptic drive deriving from nucleus tractus solitarius pump cells (Guyenet et al., 2019). These pathways constitute a network mechanism permitting pulmonary stretch receptor-mediated blunting of chemosensory amplification of neural respiratory output, relaying to glutamatergic retrotrapezoid nucleus units via nucleus tractus solitarius pump cells (Takakura et al., 2007) and conjecturally powerfully restricting the preinspiratory component of hypoglossal bursting in the presence of preserved vagal continuity (Ghali, 2015, 2019c; Ghali and Marchenko, 2016b).

Peripheral chemoreceptors constituted glomus cells exquisitely sensitive to hypercapnia, and moderate and severe levels of hypoxemia chiefly utilize purinergic signaling (Baby et al., 2006) and convey stimulus mediated modifications of neuronal spiking frequency centrally to chemosensitive glutamatergic commissural nucleus tractus solitarius neurons (Zhang et al., 2011). Glutamatergic chemoreceptor neurons distributed throughout the retrotrapezoid nucleus and parafacial respiratory group complex (Guyenet et al., 2019; Onimaru et al., 2018), nucleus tractus solitarius (Nichols et al., 2008), raphé group of nuclei (Sobrinho et al., 2014), preBötzinger complex, and fastigial nucleus (Hernandez et al., 2004) are chiefly sensitive to elevations of carbon dioxide tension or hydrogen ion concentration, with a few zones coordinately exhibiting hypoxic sensitivity (e.g., preBötzinger complex, fastigial nucleus). Hypercapnia and hypoxia elicit and amplify discharge of peripheral chemoreceptor glomus cells (Zoccal, 2015) and chemosensitive bulbar units (Guyenet et al., 2019) conveying coordinate synaptic drive to breathing generators (Marchenko et al., 2016; Molkov et al., 2014; Zoccal et al., 2008; Zoccal and Machado, 2011), sympathetic oscillators (Gilbey et al., 1995), and cardiovagal premotoneurons (Bishop, 1974; Lindsey et al., 1998), crossmodally sensitizing each other’s discharge (Blain et al., 2010) and conveying augmented coherent neuronal spiking through propriobulbar interneuronal paucisynaptic interactions (Molkov et al., 2014).

Intravenous sodium chloride administration coordinately enhanced phrenic nerve bursting frequency and thoracic sympathetic neural burst amplitude in decorticate Holtzman rats, both effects of which were abolished by precollicular transection, supporting paraventricular and supraoptic nuclei mediate the effects (da Silva et al., 2019). Carotid body removal prevented the enhancement of thoracic sympathetic nerve activity elicited by intravenous administration of sodium chloride, though failed to prevent tachypnea, supporting carotid body glomus cells may represent effective osmoreceptors (da Silva et al., 2019). Peripheral and central chemoreceptor driven augmentation of the respiratory, central pattern generator and sympathetic oscillators becomes synchronized and commonly coupled through coordinate and independent bulbobulbar, spinoreticular, and peripheral afferent inputs couple and synchronize central and peripheral chemoreceptor mediated augmentation of neuronal ensembles constituting the constituent generators, strengthened by direct interactions between overlapping propriobulbar interneuronal arrays constituting the respiratory rhythm and pattern generator, sympathetic oscillators, and cardiovagal premotoneurons(Molkov et al., 2017; Zoccal, 2015). These pathways constitute mechanisms generating autonomorespiratory coupling and crossmodal modulation amongst disparate oscillators and nuclei at rest during normoxic normocapnia, which becomes amplified by hypercapnia and acute and chronic hypoxia (Figs. 1-5) (Molkov et al., 2017; Zoccal, 2015). Sympathorespiratory, parasympathorespiratory, and sympathoparasympathetic coupling may further be subject to modulation by cardiovascular and respiratory muscle training, putatively mediated by amplifying the cardiovagal sinoatrial decelerator tone, evidenced in powerful modifications of the cardiovascular spectral variabilities (de Abreu et al., 2019).

The respiratory rhythm (Anderson et al., 2016; Anderson and Ramirez, 2017; Ghali, 2019c; Morgado-Valle and Beltran-Parrazal, 2017; Ramirez and Baertsch, 2018) and pattern (Ghali, 2018; Marchenko et al., 2016; Ramirez and Baertsch, 2018; Smith et al., 2009), sympathetic oscillations (Ghali, 2018, 2017a, 2019b), and cardiovagal premotoneuronal activity (Baekey et al., 2010; Bishop, 1974) are generated by distinct, though overlapping, networks residing chiefly within the metencephalon and myelencephalon (Ghali, 2017a, b, c, 2019a, 2019b; Marchenko et al., 2016; Molkov et al., 2017) and powerfully modulated by descending synaptic inputs from the cerebrum (Antal, 1985), hippocampus (Ruit and Neafsey, 1988), amygdala (Lacuey et al., 2019), thalamus (Ogundele et al., 2017), hypothalamus (Fukushi et al., 2019), midbrain (Ghali and Ghali, 2020), and cerebellum (Horn and Waldrop, 1997; Schmid et al., 1988; Subramanian and Holstege, 2014). The widespread presence of autonomorespiratory coupling, manifest as baroreceptor and sympathetic modulation of neural respiratory activity (Baekey et al., 2010; Bishop, 1974; Lindsey et al., 1998), respirophasic modulation of sympathetic discharge (Haselton and Guyenet, 1989; Mandel and Schreihofer, 2006; McAllen, 1987; Molkov et al., 2017), and ambiguual cardiovagal premotoneurons (Dergacheva et al., 2010), and crossmodal modulation between sympathetic and parasympathetic outflows evidence coordinate inputs to, and direct interaction amongst and between, propriobulbar interneuronal microcircuit oscillators constituting the pattern generators mediating the neurogenesis of these activities and heterologous peripheral modulation of breathing, sympathetic oscillations, and cardiovagal premotoneurons by vagal afferents conveying oscillatory baroreceptor, chemoreceptor, and pulmonary stretch receptor inputs (Molkov et al., 2017; Zoccal, 2015). Sympathorespiratory and cardiorespiratory coupling thus critically requires integrity and functionality of the core metencephalomyelencephalic generator circuitry and elements diffusely distributed within the ventrolateral and dorsolateral metencephalic tegmentum, brainstem reticular formation, vagal inputs, and postinspiratory neurons (Fig. 5) (Molkov et al., 2017), elucidating mechanisms contributing to emergent genesis of autonomorespiratory coupling through studies utilizing contemporaneous respiratory and sympathetic neuronal (Terui et al., 1986) and peripheral sympathetic neural efferent recordings (St. Croix et al., 1999), directed lesioning (Dick et al., 2009; Molkov et al., 2011), stereotaxic localized pressure microinjections of pharmacological agonists and antagonists of excitatory and inhibitory neurotransmission and neuromodulation (Abdala et al., 2009; Moraes et al., 2012b), and cross-correlation, spectral, and coherence analyses of the recorded electrophysiological activities. While the alternating rhythmic discharge of inspiratory and expiratory activities of breathing neurally segmented into inspiratory, postinspiratory, and late expiratory epochs effectively serves a clear and identifiable functional purpose (Marchenko et al., 2016), the significance of rhythmic discharge in, and crossmodal modulation of, sympathetic oscillations, sympathetic neural efferent discharge, and ambiguual cardiovagal premotoneurons is not immediately obvious (Zoccal, 2015). Some have posited respirophasic modulation of the sympathetic neural output may contribute to and mediate coordinate changes in ventilation and vascular tone to contemporaneously optimize tissue oxygenation and perfusion during resting conditions and amplify vasoconstrictor arteriolar tone and amplitude of the microcirculatory vasomotion (Zoccal et al., 2009b). Accordingly, modeling studies have indicated interactions among the respiratory, central pattern generator, sympathetic oscillators, and parasympathetic nuclei may serve to robustly improve myocardial function (Ben-Tal et al., 2012). Further studies are thus necessary to more thoroughly elucidate functional utility and mechanisms underlying the genesis and functional utility of autonomorespiratory coupling (Molkov et al., 2017).

M.G.Z.G.: conception and design, acquisition of data, analysis and interpretation of data, drafting the article, and revising critically for intellectual content; approval of the final version of the manuscript.

Compliance with ethical standards.

We appreciate the helpful comments from anonymous reviewers for improving the quality of this review.

We have no conflicts of interest to disclose.