Brain stem neural mechanisms for vocalization in decerebrate cats

Neuroanatomical frameworks for volitional control of breathing and orofacial behaviors

Breathing is the only vital function that can be volitionally controlled. However, a detailed understanding how volitional (cortical) motor commands can transform vital breathing activity into adaptive breathing patterns that accommodate orofacial behaviors such as swallowing, vocalization or sniffing remains to be developed. Recent neuroanatomical tract tracing studies have identified patterns and origins of descending forebrain projections that target brain nuclei involved in laryngeal adductor function which is critically involved in orofacial behavior. These nuclei include the midbrain periaqueductal gray and nuclei of the respiratory rhythm and pattern generating network in the brainstem, specifically including the pontine Kölliker-Fuse nucleus and the pre-Bötzinger complex in the medulla oblongata. This review discusses the functional implications of the forebrain-brainstem anatomical connectivity that could underlie the volitional control and coordination of orofacial behaviors with breathing.

Forebrain projection neurons target functionally diverse respiratory control areas in the midbrain, pons, and medulla oblongata

Eupnea is generated by neural circuits located in the ponto-medullary brainstem, but can be modulated by higher brain inputs which contribute to volitional control of breathing and the expression of orofacial behaviors, such as vocalization, sniffing, coughing, and swallowing. Surprisingly, the anatomical organization of descending inputs that connect the forebrain with the brainstem respiratory network remains poorly defined. We hypothesized that descending forebrain projections target multiple distributed respiratory control nuclei across the neuroaxis. To test our hypothesis, we made discrete unilateral microinjections of the retrograde tracer cholera toxin subunit B in the midbrain periaqueductal gray (PAG), the pontine Kölliker-Fuse nucleus (KFn), the medullary Bötzinger complex (BötC), pre-BötC, or caudal midline raphé nuclei. We quantified the regional distribution of retrogradely labeled neurons in the forebrain 12-14 days postinjection. Overall, our data reveal that descending inputs from cortical areas predominantly target the PAG and KFn. Differential forebrain regions innervating the PAG (prefrontal, cingulate cortices, and lateral septum) and KFn (rhinal, piriform, and somatosensory cortices) imply that volitional motor commands for vocalization are specifically relayed via the PAG, while the KFn may receive commands to coordinate breathing with other orofacial behaviors (e.g., sniffing, swallowing). Additionally, we observed that the limbic or autonomic (interoceptive) systems are connected to broadly distributed downstream bulbar respiratory networks. Collectively, these data provide a neural substrate to explain how volitional, state-dependent, and emotional modulation of breathing is regulated by the forebrain.

Activity of respiratory neurons in the rostral medulla during vocalization, swallowing, and coughing in guinea pigs

To examine the relationship between the neuronal networks underlying respiration and non-respiratory behaviors such as vocalization and airway defensive reflexes, we compared the activity of respiratory neurons in the ventrolateral medulla during breathing with that during non-respiratory behaviors including vocalization, swallowing, and coughing in guinea pigs. During fictive vocalization the activity of augmenting expiratory neurons ceased, whereas the other types of expiratory neurons did not show a consistent tendency of increasing or decreasing activity. All inspiratory neurons discharged in synchrony with the phrenic nerve activity. Most of the phase-spanning neurons were activated throughout the vocal phase. During fictive swallowing, many expiratory and inspiratory neurons were silent, whereas many phase-spanning neurons were activated. During fictive coughing, many expiratory neurons were activated during the expiratory phase of coughing. Most inspiratory neurons discharged in parallel with the phrenic nerve activity during coughing. Many phase-spanning neurons were activated during the expiratory phase of coughing. These findings indicate that the medullary respiratory neurons help shape respiratory muscle nerve activity not only during breathing but also during these non-respiratory behaviors, and thus suggest that at least some of the respiratory neurons are shared among the neuronal circuits underlying the generation of breathing and non-respiratory behaviors.

Coordination of breathing with nonrespiratory activities

Many articles in this section of Comprehensive Physiology are concerned with the development and function of a central pattern generator (CPG) for the control of breathing in vertebrate animals. The action of the respiratory CPG is extensively modified by cortical and other descending influences as well as by feedback from peripheral sensory systems. The central nervous system also incorporates other CPGs, which orchestrate a wide variety of discrete and repetitive, voluntary and involuntary movements. The coordination of breathing with these other activities requires interaction and coordination between the respiratory CPG and those governing the nonrespiratory activities. Most of these interactions are complex and poorly understood. They seem to involve both conventional synaptic crosstalk between groups of neurons and fluid identity of neurons as belonging to one CPG or another: neurons that normally participate in breathing may be temporarily borrowed or hijacked by a competing or interrupting activity. This review explores the control of breathing as it is influenced by many activities that are generally considered to be nonrespiratory. The mechanistic detail varies greatly among topics, reflecting the wide variety of pertinent experiments.

Avian nucleus retroambigualis: cell types and projections to other respiratory-vocal nuclei in the brain of the zebra finch (Taeniopygia guttata)

In songbirds song production requires the intricate coordination of vocal and respiratory muscles under the executive influence of the telencephalon, as for speech in humans. In songbirds the site of this coordination is suspected to be the nucleus retroambigualis (RAm), because it contains premotor neurons projecting upon both vocal motoneurons and spinal motoneurons innervating expiratory muscles, and because it receives descending inputs from the telencephalic vocal control nucleus robustus archopallialis (RA). Here we used tract-tracing techniques to provide a more comprehensive account of the projections of RAm and to identify the different populations of RAm neurons. We found that RAm comprises diverse projection neuron types, including: 1) bulbospinal neurons that project, primarily contralaterally, upon expiratory motoneurons; 2) a separate group of neurons that project, primarily ipsilaterally, upon vocal motoneurons in the tracheosyringeal part of the hypoglossal nucleus (XIIts); 3) neurons that project throughout the ipsilateral and contralateral RAm; 4) another group that sends reciprocal, ascending projections to all the brainstem sources of afferents to RAm, namely, nucleus parambigualis, the ventrolateral nucleus of the rostral medulla, nucleus infra-olivarus superior, ventrolateral parabrachial nucleus, and dorsomedial nucleus of the intercollicular complex; and 5) a group of relatively large neurons that project their axons into the vagus nerve. Three morphological classes of RAm cells were identified by intracellular labeling, the dendritic arbors of which were confined to RAm, as defined by the terminal field of RA axons. Together the ascending and descending projections of RAm confirm its pivotal role in the mediation of respiratory-vocal control.

Neuronal activity in the medulla oblongata during vocalization. A single-unit recording study in the squirrel monkey

In six squirrel monkeys (Saimiri sciureus), the medulla oblongata was explored with microelectrodes, looking for vocalization-correlated activity. The vocalizations were elicited by microinjections of glutamate agonists into the periaqueductal grey of the midbrain. Vocalization-related cells were found in greater numbers in the nucl. ambiguus (Ab) and retroambiguus (RAb), in the parvocellular, magnocellular and central reticular formation as well as in the solitary tract nucleus and spinal trigeminal nucleus. Small numbers were also found in the vestibular complex, cuneate nuclei, inferior olive and lateral reticular nucleus. A differentiation of the neuronal responses into 12 reaction types reveals that the frequency of each reaction type varies from brain structure to brain structure, thus allowing a specification of the different vocalization-related areas. According to this specification, it is proposed that initiation of vocalization takes place via the parvocellular reticular formation; vocal pattern control is mainly brought about by the parvocellular reticular formation, Ab, solitary tract nucleus and spinal trigeminal nucleus; expiratory control and respiratory-laryngeal coordination is carried out by the RAb, Ab and central nucleus of the reticular formation; vocalization-specific postural adjustments are carried out via the vestibular and cuneate nuclei.

Synchronized fast rhythms in inspiratory and expiratory nerve discharges during fictive vocalization

In precollicular decerebrate and paralyzed cats, respiratory nerve activities were recorded during fictive vocalization (FV), which consisted of a distinctive pattern of 1) decreased inspiratory (I) and expiratory (E) phase durations, 2) marked increase of phrenic activity and moderate changes of recurrent laryngeal (RL) and superior laryngeal (SL) I activities, and 3) massive recruitment of laryngeal and abdominal (ABD; lumbar) E activities. FV was produced by electrical stimulation (100 Hz) in the midbrain periaqueductal gray (PAG) or its putative descending pathways in the ventrolateral pons (VLP). Spectral and correlation analyses revealed three types of effect on fast rhythms during FV. 1) I activities: the coherent high-frequency oscillations in I (I-HFO, 60-90 Hz) present in phrenic and RL discharges during the control state did not change qualitatively, but there was an increase of power and a moderate increase (4-10 Hz) of frequency. Sometimes a distinct relatively weak stimulus-locked rhythm appeared. 2) RL and SL activities during E: in recruited discharges, a prominent intrinsic rhythm (coherent E-HFOs at 50-70 Hz) appeared; sometimes a distinct relatively strong stimulus-locked rhythm appeared. 3) ABD activities during E: this recruited activity had no intrinsic rhythm but had an evoked oscillation locked to the stimulus frequency. Thus FV is characterized by 1) appearance of prominent coherent intrinsic rhythms in RL and SL E discharges, which presumably arise as a result of excitation and increased interactions in laryngeal networks; 2) modification of intrinsic rhythmic interactions in inspiratory networks; and 3) evoked rhythms in augmenting-E neuron networks without occurrence of intrinsic rhythms.