Characterization of respiratory-modulated activities of hypoglossal motoneurons

In decerebrate, vagotomized, paralyzed, and ventilated cats, activities of the phrenic nerve and single hypoglossal nerve fibers were monitored. The great majority of hypoglossal neuronal activities were inspiratory (I), discharging during a period approximating that of phrenic. Many were not active at normocapnia but were recruited in hypercapnia or hypoxia. Once recruited, discharge frequencies, which rose quickly to near maximal levels in early to midinspiration, significantly increased with further augmentations of drive. Also, the onset of activities became progressively earlier, compared with phrenic discharge, in hypercapnia or hypoxia. Smaller numbers of hypoglossal fiber activities, having inspiratory-expiratory (I-E), expiratory (E), expiratory-inspiratory (E-I), or tonic discharge patterns, were also recorded. Activities of E, I-E, and those I fibers that became I-E in high drive may underlie the early burst of expiratory activity of the hypoglossal nerve. It is concluded that the firing and recruitment patterns of hypoglossal neurons differ from those of phrenic motoneurons. However, responses to chemoreceptor stimuli are similar among the two neuronal groups.     

Comparison of phrenic motoneuron activity during eupnea and gasping

Single-fiber phrenic nerve action potentials were recorded together with activity of contralateral whole phrenic nerve rootlets during eupnea and gasping in decerebrate, cerebellectomized, vagotomized, paralyzed, and ventilated cats. Gasping was reversibly produced by cooling a fork thermode positioned through the pontomedullary junction. In eupnea, phrenic motoneurons were distributed into "early" and "late" populations relative to their onset of activity during inspiration. During gasping, however, both fiber types typically commenced activity at the beginning of the phrenic nerve burst. Moreover, late fibers, but not early units, exhibited an augmentation of discharge frequency with the onset of gasping. The concentration of activity of all phrenic motoneurons at the beginning of inspiration and the increase in late-unit discharge frequency account for the faster rise of the gasp as compared with the eupneic breath. It is concluded that the pattern of phrenic nerve activation during gasping differs fundamentally from that during eupnea. These results support the concept that mechanisms underlying the neurogenesis of gasping and eupnea may not be identical.     

Influence of pulmonary inflations on discharge patterns of phrenic motoneurons

The purpose of this study was to assess the influence of pulmonary inflations on activities of single phrenic motoneurons. Studies were performed in decerebrate and paralyzed cats; activities of phrenic nerve and single phrenic motoneurons were recorded. Animals were ventilated with a servo-respirator which produced alterations in tracheal pressure in parallel with changes in integrated activity of the phrenic nerve. At end-tidal fractional concentrations of CO2 of 0.05, phrenic motoneurons were distributed into "early" and "late" populations, depending on time of onset of activity. During the late stages of neural inspiration, differences in levels of integrated activity of the phrenic nerve became evident between cycles with and without lung inflations. At a time approximating 90% of the inspiratory duration during inflations, integrated phrenic activity was higher for cycles with inflation. Concomitantly, with lung inflations, the discharge frequencies of early phrenic motoneurons were lower, and late motoneurons began to discharge sooner than when inflations were withheld. Similar results were obtained in hypercapnia. We conclude that reflexes activated by pulmonary inflations may produce augmentation, as well as inhibition of phrenic motoneuronal activities. Factors responsible for eliciting these reflex augmentations and inhibitions are discussed.     

Synchronization of respiratory frequency by somatic afferent stimulation.

In cats anesthetized with chloralose-urethan, vagotomized, paralyzed, and artifically ventilated, superficial radial (cutaneous) and hamstring (muscle) nerve afferents were stimulated while phrenic nerve electrical activity was recorded. The results obtained with both types of nerves were similar. Stimulation in mid and late expiration advanced the onset of the next inspiration, shortening its duration. Stimulation in early inspiration advanced, while that in late inspiration delayed, the onset of the next expiration. These effects were often accompanied by changes in phrenic motoneuron firing patterns (earlier recruitment, increased discharge frequency, increased slope of integrated phrenic neurogram). Repetitive somatic afferent stimulation produced sustained increases in respiratory frequency in all cats and in half of them entrainment of respiratory frequency to the frequency of stimulation occurred at ratios such as 4:3, 4:5, 1:2, 1:3, 1:4, and 1:7. The lowest stimulus intensity required for evoking these phase shifts was between 5 and 10T (threshold of most excitable fibers) for muscle afferents and between 1 and 2T for cutaneous afferents. These results demonstrate the existence of a reflex mechanism capable of locking respiratory frequency to that of a periodic somatic afferent input. They also provide an experimental basis for the hypothesis that reflexes are resposible for the observed locking between step or pedal frequency and respiratory rate during exercise in man.     

Inspiratory rhythm in airway smooth muscle tone

In anesthetized paralyzed open-chested cats ventilated with low tidal volumes at high frequency, we recorded phrenic nerve activity, transpulmonary pressure (TPP), and either the tension in an upper tracheal segment or the impulse activity in a pulmonary branch of the vagus nerve. The TPP and upper tracheal segment tension fluctuated with respiration, with peak pressure and tension paralleling phrenic nerve activity. Increased end-tidal CO2 or stimulation of the carotid chemoreceptors with sodium cyanide increased both TPP and tracheal segment tension during the increased activity of the phrenic nerve. Lowering end-tidal CO2 or hyperinflating the lungs to achieve neural apnea (lack of phrenic activity) caused a decrease in TPP and tracheal segment tension and abolished the inspiratory fluctuations. During neural apnea produced by lowering end-tidal CO2, lung inflation caused no further decrease in tracheal segment tension and TPP. Likewise, stimulation of the cervical sympathetics, which caused a reduction in TPP and tracheal segment tension during normal breathing, caused no further reduction in these parameters when the stimulation occurred during neural apnea. During neural apnea the tracheal segment tension and TPP were the same as those following the transection of the vagi or the administration of atropine (0.5 mg/kg). Numerous fibers in the pulmonary branch of the vagus nerve fired in synchrony with the phrenic nerve. Only these fibers had activity which paralleled changes in TPP and tracheal tension. We propose that the major excitatory input to airway smooth muscle arises from cholinergic nerves that fire during inspiration, which have preganglionic cell bodies in the ventral respiratory group in the region of the nucleus ambiguus and are driven by the same pattern generators that drive the phrenic and inspiratory intercostal motoneurons.     

Analysis of postinspiratory activity of phrenic motoneurons with chemical and vagal reflexes

We examined the effects of chemical and reflex drives on the postinspiratory inspiratory activity (PIIA) of phrenic motoneurons using a single-fiber technique. Action potentials from "single" fibers were recorded from the C5 phrenic root together with contralateral mass phrenic activity (also from C5) in anesthetized, paralyzed, and artificially ventilated cats with intact vagus and carotid sinus nerves. Nerve fibers were classified as "early" or "late" based on their onset of discharge in relation to mass phrenic activity during hyperoxic ventilation. Only the early fibers displayed PIIA but not the late fibers, even when their activity began earlier in inspiration with increased chemical drives. Isocapnic hypoxia increased, whereas hyperoxic hypercapnia shortened the duration of PIIA. Pulmonary stretch and "irritant" receptors inhibited PIIA. Hypercapnia and stimulation of peripheral chemoreceptors by lobeline excited both early and late units to the same extent, but hypoxic ventilation had a less marked excitatory effect on late fiber activity. Irritant receptor activation increased the activity of early more than late fibers. Hyperoxic hyperventilation eliminated late phrenic fiber activity, whereas early fibers became tonically active. Bilateral vagotomy abolished this sustained discharge in eight of nine early units, suggesting the importance of vagal afferents in producing tonic firing during hyperventilation. These results suggest that early and late phrenic fibers have different responses to chemical stimuli and to vagally mediated reflexes; late units do not discharge in postinspiratory period, whereas early fibers do; the PIIA is not affected in the same way by various chemical and vagal inputs; and early units that exhibit PIIA display tonic activity with hyperoxic hypocapnia.     

Differing activities of medullary respiratory neurons in eupnea and gasping

Our purpose was to compare further eupneic ventilatory activity with that of gasping. Decerebrate, paralyzed, and ventilated cats were used; the vagi were sectioned within the thorax caudal to the laryngeal branches. Activities of the phrenic nerve and medullary respiratory neurons were recorded. Antidromic invasion was used to define bulbospinal, laryngeal, or not antidromically activated units. The ventilatory pattern was reversibly altered to gasping by exposure to 1% carbon monoxide in air. In eupnea, activities of inspiratory neurons commenced at various times during inspiration, and for most the discharge frequency gradually increased. In gasping, the peak discharge frequency of inspiratory neurons was unaltered. However, all commenced activities at the start of the phrenic burst and reached peak discharge almost immediately. The discharge frequencies of all groups of expiratory neurons fell in gasping, with many neurons ceasing activity entirely. These data are consistent with the hypothesis that brain stem mechanisms controlling eupnea and gasping differ fundamentally.     

Some effects of superior laryngeal nerve stimulation on sympathetic preganglionic neuron firing

Repetitive electrical stimulation of afferent fibers in the superior laryngeal nerve (SLN) evoked depressant or excitatory effects on sympathetic preganglionic neurons of the cervical trunk in Nembutal-anesthetized, paralyzed, artifically ventilated cats. The depressant effect, which consisted of suppression of the inspiration-synchronous discharge of units with such firing pattern, was obtained at low strength and frequency of stimulation (e.g. 600 mV, 30 Hz) and was absent at end-tidal CO2 values below threshold for phrenic nerve activity. The excitatory effect required higher intensity and frequency of stimulation and was CO2 independent. The depressant effect on sympathetic preganglionic neurons with inspiratory firing pattern seemed a replica of the inspiration-inhibitory effect observed on phrenic motoneurons. Hence, it could be attributed to the known inhibition by the SLN of central inspiratory activity, if it is assumed that this is a common driver for phrenic motoneurons and some sympathetic preganglionic neurons. The excitatory effect, on the other hand, appears to be due to connections of SLN afferents with sympathetic preganglionic neurons, independent of the respiratory center.     

Discharge patterns of phrenic motoneurons during fictive coughing and vomiting in decerebrate cats.

In decerebrate, paralyzed, and ventilated cats, we recorded the activity of 100 spontaneously active phrenic motor axons during the increased phrenic discharges characteristic of fictive vomiting (FV) and coughing (FC). During control respiratory cycles, approximately one-half the neurons were recruited in the first decile of inspiration; recruitment continued throughout inspiration. During FV, the duration of phrenic discharge was halved; 20 of 26 motoneurons studied were recruited in the first decile of the burst. During FC, recruitment times did not change compared with control, although the duration of the phrenic burst doubled. Discharge frequencies increased and recruitment order of phrenic motoneurons was virtually unaffected during FC and FV. Limited recruitment of previously inactive neurons in the filaments from which we recorded was found during FV and FC. During FV, 1 previously inactive motoneuron was recruited in 16 filaments containing 25 spontaneously active motor axons. During FC, 3 new motoneurons were recruited in addition to the 64 already active in 35 filaments. Recruitment during FV and FC was absent even when recording from filaments known, on the basis of antidromic activation, to contain inactive motor axons. During FV, 10 of 26 motoneurons began their discharges with doublets (interspike interval < 10 ms); doublets occurred in only 4 of 67 motoneurons during FC. Already active phrenic motoneurons contributed to the intense phrenic activity associated with both respiratory (coughing) and nonrespiratory (vomiting) behavior by increases in discharge frequency, earlier recruitment, and doublets; the contribution of previously quiescent motoneurons remains uncertain.     

家兔脊髓蛛网膜下腔注射乙酰胆碱对膈神经放电的影响

实验在50例麻醉、麻痹和切断双侧迷走神经的家兔上进行。用双极银丝电极引导膈神经放电,经放大、幅度积分而后记录在 X-Y 记录仪上。脊髓蛛网膜下腔注射乙酰胆碱(ACh)50μg 后,膈神经吸气放电幅度增加,平均增加30.6±11.6%.ACh 的这一效应可被阿托品阻断,但不能被六甲双铵、酚妥拉明和心得安阻断。但单独注射上述任何一种受体阻断剂均不能改变膈神经的放电活动。上述结果提示,脊髓蛛网膜下腔注射 ACh 可兴奋膈运动神经元,而且 ACh 的这一兴奋效应是由 Μ 受体中介的,但在正常生理情况下,胆碱能递质系统对膈神经元似无紧张性兴奋作用。因此,ACh 递质可能在脊髓水平对高级中枢下传的呼吸驱动信息的整合中起兴奋性调制作用。     

Firing properties and hypercapnic responses of single phrenic motor axons in the rat

Inspiratory phase activity was recorded from 33 phrenic motoneuron (PM) axonal fibers in anesthetized, vagotomized, artificially ventilated adult rats. During control conditions (no inspired CO2 added), the population of PM fibers could be separated into early and late onset types based on the time of firing onset relative to the onset of whole phrenic nerve activity. Mean discharge frequencies of both types were not significantly different. Compared with late PM's, early PM's had more spikes per inspiration, fired for a longer period, and the last spike occurred later and during the postinspiratory period. Further, the mean minimal interspike interval was shorter for early PM's than for late PM's. Increasing inspired CO2 to 0.03 and 0.05 resulted in earlier firing onsets and a greater number of spikes per inspiration, particularly for late PM's. Increases in mean firing frequency occurred for both PM types. Mean minimal interspike intervals for both types of PM's showed progressive reductions as CO2 rose. For almost all of the firing properties examined in this study, responses of rat PM axons were similar to those previously reported for the cat.     

Recurrent laryngeal nerve activity exhibits a 5-HT-mediated long-term facilitation and enhanced response to hypoxia following acute intermittent hypoxia in rat

A progressive and sustained increase in inspiratory-related motor output ("long-term facilitation") and an augmented ventilatory response to hypoxia occur following acute intermittent hypoxia (AIH). To date, acute plasticity in respiratory motor outputs active in the postinspiratory and expiratory phases has not been studied. The recurrent laryngeal nerve (RLN) innervates laryngeal abductor muscles that widen the glottic aperture during inspiration. Other efferent fibers in the RLN innervate adductor muscles that partially narrow the glottic aperture during postinspiration. The aim of this study was to investigate whether or not AIH elicits a serotonin-mediated long-term facilitation of laryngeal abductor muscles, and if recruitment of adductor muscle activity occurs following AIH. Urethane anesthetized, paralyzed, unilaterally vagotomized, and artificially ventilated adult male Sprague-Dawley rats were subjected to 10 exposures of hypoxia (10% O(2) in N(2), 45 s, separated by 5 min, n = 7). At 60 min post-AIH, phrenic nerve activity and inspiratory RLN activity were elevated (39 ± 11 and 23 ± 6% above baseline, respectively). These responses were abolished by pretreatment with the serotonin-receptor antagonist, methysergide (n = 4). No increase occurred in time control animals (n = 7). Animals that did not exhibit postinspiratory RLN activity at baseline did not show recruitment of this activity post-AIH (n = 6). A repeat hypoxia 60 min after AIH produced a significantly greater peak response in both phrenic and RLN activity, accompanied by a prolonged recovery time that was also prevented by pretreatment with methysergide. We conclude that AIH induces neural plasticity in laryngeal motoneurons, via serotonin-mediated mechanisms similar to that observed in phrenic motoneurons: the so-called "Q-pathway". We also provide evidence that the augmented responsiveness to repeat hypoxia following AIH also involves a serotonergic mechanism.     

Respiratory-related hypoglossal nerve activity: influence of anesthetics

In decerebrate, vagotomized, paralyzed, and ventilated cats, phrenic and respiratory-related hypoglossal discharges were evident at normocapnic normoxia or hyperoxia. Both increased progressively in hypercapnia or hypoxia. With increasing drive, onset of inspiratory hypoglossal activity began earlier relative to phrenic onset; an early expiratory hypoglossal burst was also observed. Following subanesthetic doses of chloralose, halothane, ketamine, or pentobarbital, hypoglossal activity was depressed much more than phrenic discharge. In moderate hypercapnia or hypoxia, phrenic activity increased more than hypoglossal, whereas, at high drive, the latter rose more sharply in some cats. Electromyograms of the diaphragm and genioglossus were recorded in intact awake cats to determine if their responses and those of decerebrates are comparable. Respiratory-related genioglossal discharge was evident in normocapnia. We conclude that anesthesia suppresses hypoglossal motor activities much more than those of the bulbospinal-phrenic system. Data for decerebrate cats and unanesthetized cats or humans provide no evidence of a differential distribution of chemoreceptor afferents on hypoglossal and bulbospinal-phrenic neurons, as suggested by results in anesthetized animals.     

Neural drive to tongue protrudor and retractor muscles following pulmonary C-fiber activation.

Hypoglossal (XII) nerve recordings indicate that pulmonary C-fiber (PCF) receptor activation reduces inspiratory bursting and triggers tonic discharge. We tested three hypotheses related to this observation: 1) PCF receptor activation inhibits inspiratory activity in XII branches innervating both tongue protrudor muscles (medial branch; XIImed) and retractor muscles (lateral branch; XIIlat); 2) reduced XII neurogram amplitude reflects decreased XII motoneuron discharge rate; and 3) tonic XII activity reflects recruitment of previously silent motoneurons. Phrenic, XIImed, and XIIlat neurograms were recorded in anesthetized, paralyzed, and ventilated rats. Capsaicin delivered to the jugular vein reduced phrenic bursting at doses of 0.625 and 1.25 mug/kg but augmented bursting at 5 mug/kg. All doses reduced inspiratory amplitude in XIImed and XIIlat (P < 0.05), and these effects were eliminated following bilateral vagotomy. Single-fiber recordings indicated that capsaicin causes individual XII motoneurons to either decrease discharge rate (n = 101/153) or become silent (n = 39/153). Capsaicin also altered temporal characteristics such that both XIImed and XIIlat inspiratory burst onset occurred after the phrenic burst (P < 0.05). Increases in tonic discharge after capsaicin were greater in XIImed vs. XIIlat (P < 0.05); single-fiber recordings indicated that tonic discharge reflected recruitment of previously silent motoneurons. We conclude that PCF receptor activation reduces inspiratory XII motoneuron discharge and transiently attenuates neural drive to both tongue protrudor and retractor muscles. However, tonic discharge appears to be selectively enhanced in tongue protrudor muscles. Accordingly, reductions in upper airway stiffness associated with reduced XII burst amplitude may be offset by enhanced tonic activity in tongue protrudor muscles.     

正常猫膈神经单纤维电活动特征

在麻醉猫和麻痹的切断迷走神经的清醒猫,观察了膈神经单纤维电活动特征。1.电活动类型:按膈神经单纤维放电与其总干放电的相位关系分为三种类型。(1)完全同步型,即单纤维放电与总干放电同时开始并同时停止,占76.9％。(2)部分同步型占15.4％,其中早期同步,即单纤维放电与总干放电同时开始,但提前终止,占1.9％,中期同步,即单纤维放电较总干放电开始晚,又提前终止,占5.8％,晚期同步,即单纤维放电较总干放电开始晚,但两者同时终止,占7.7％。(8)非同步型,即吸气相和呼气相都有放电,但呼气相时冲动频率较低,占7.7％。前两型为单纯的吸气性放电,共占92.3％。2.单纤维放电平均参数值:麻醉猫每次吸气发放11个冲动,其频率为21次/秒,清醒猫每次吸气发放18个冲动,其频率为34次/秒。结果表明:猫膈神经单纤维放电类型和文献上报导的直接记录膈神经运动神经元放电一致,即以单纯的吸气性放电为最多。     

Attenuation of phrenic motor discharge by phrenic nerve afferents

Short latency phrenic motor responses to phrenic nerve stimulation were studied in anesthetized, paralyzed cats. Electrical stimulation (0.2 ms, 0.01-10 mA, 2 Hz) of the right C5 phrenic rootlet during inspiration consistently elicited a transient reduction in the phrenic motor discharge. This attenuation occurred bilaterally with an onset latency of 8-12 ms and a duration of 8-30 ms. Section of the ipsilateral C4-C6 dorsal roots abolished the response to stimulation, thereby confirming the involvement of phrenic nerve afferent activity. Stimulation of the left C5 phrenic rootlet or the right thoracic phrenic nerve usually elicited similar inhibitory responses. The difference in onset latency of responses to cervical vs. thoracic phrenic nerve stimulation indicates activation of group III afferents with a peripheral conduction velocity of approximately 10 m/s. A much shorter latency response (5 ms) was evoked ipsilaterally by thoracic phrenic nerve stimulation. Section of either the C5 or C6 dorsal root altered the ipsilateral response so that it resembled the longer latency contralateral response. The low-stimulus threshold and short latency for the ipsilateral response to thoracic phrenic nerve stimulation suggest that it involves larger diameter fibers. Decerebration, decerebellation, and transection of the dorsal columns at C2 do not abolish the inhibitory phrenic-to-phrenic reflex.     

The crossed phrenic phenomenon: a model for plasticity in the respiratory pathways following spinal cord injury.

Hemisection of the cervical spinal cord rostral to the level of the phrenic nucleus interrupts descending bulbospinal respiratory pathways, which results in a paralysis of the ipsilateral hemidiaphragm. In several mammalian species, functional recovery of the paretic hemidiaphragm can be achieved by transecting the contralateral phrenic nerve. The recovery of the paralyzed hemidiaphragm has been termed the "crossed phrenic phenomenon." The physiological basis for the crossed phrenic phenomenon is as follows: asphyxia induced by spinal hemisection and contralateral phrenicotomy increases central respiratory drive, which activates a latent crossed respiratory pathway. The uninjured, initially latent pathway mediates the hemidiaphragm recovery by descending into the spinal cord contralateral to the hemisection and then crossing the midline of the spinal cord before terminating on phrenic motoneurons ipsilateral and caudal to the hemisection. The purpose of this study is to review work conducted on the crossed phrenic phenomenon and to review closely related studies focusing particularly on the plasticity associated with the response. Because the review deals with recovery of respiratory muscles paralyzed by spinal cord injury, the clinical relevance of the reviewed studies is highlighted.     

Influences from laryngeal afferents on expiratory bulbospinal neurons and motoneurons

The purpose of this study is to analyze the reflex effects of laryngeal afferent activation on respiratory patterns in anesthetized, vagotomized, paralyzed, ventilated cats. We recorded simultaneously from the phrenic nerve, T10 internal intercostal nerve, and single bulbospinal expiratory neurons of the caudal ventral respiratory group (VRG). Laryngeal afferents were activated by electrical stimulation of the superior laryngeal nerve (SLN) or by cold-water infusion into the larynx. Both types of stimuli caused inhibition of phrenic activity and facilitation of internal intercostal nerve activity, indicating expiratory effort. The activity of 46 bulbospinal expiratory cells was depressed during SLN electrical stimulation, and 13 of them were completely inhibited. In 44 of 56 neurons tested, mean firing frequency (FFmean) was decreased in response to cold-water infusion and 8 others responded with increased FFmean; in the remaining 4 neurons, FFmean was unchanged. Possible reasons for different neuronal responses to SLN electrical stimulation and water infusion are discussed. We conclude that bulbospinal expiratory neurons of VRG were not the source of the reflex motoneuronal expiratory-like activity produced by SLN stimulation. Other, not yet identified inputs to spinal expiratory motoneurons are activated during this experimental condition.     

Inhibition of inspiratory upper airway motoneuron activity by phasic volume feedback

The effects of phasic volume feedback on efferent hypoglossal, recurrent laryngeal and phrenic nerve activity were studied in decerebrate, paralyzed intubated cats ventilated with a phrenic-driven servo-respirator. The gain of the respirator was altered for single inspirations, and the resulting changes in neural activities were quantified by comparison with respective neural activities without phasic volume feedback. The volume thresholds for suppression of hypoglossal and recurrent laryngeal activities were time independent. Above these two thresholds and extending over a substantial range, volume feedback caused graded inhibition of upper airway motoneuron outputs. At any particular time during inspiration the relationships between hypoglossal or recurrent laryngeal inhibition and volume were concave to the volume axis. Rate of airflow appeared to exert an effect on upper airway motoneuron activity independent of volume. These results indicate that for hypoglossal and recurrent laryngeal efferent activity 1) volume feedback can cause a sustained graded inhibition throughout inspiration; 2) the volume thresholds are time independent; and 3) partial inhibition decreases susceptibility to additional inhibition. These actions of volume feedback on upper airway motoneuron output differ from those on phrenic efferent discharge and show that phasic vagal volume feedback has a marked and differential effect on upper airway motoneuron activity. The vagus, in this preparation, appears to play a critical role in the regulation of upper airway motoneuron activity and therefore maintenance of upper airway patency.     

Single motor unit activity in the diaphragm of cat during pressure breathing

In this study we analyzed the breath-by-breath activity of single motor units in the diaphragm slip of allobarbital-anesthetized cats during quiet breathing and during continuous positive- and negative-pressure breathing. Our objective was to determine whether single motor units, on the basis of their activities, can be separated into discrete subpopulations or whether they fall on a continuum analogous to that of motor units of hindlimb muscles. The firing profiles of each unit were characterized for each pressure level by the onset and peak firing frequencies, onset latency, duration of firing, number of impulses per breath, and minimal frequency, when appropriate. Units with shorter onset latencies had higher peak frequencies, longer firing durations, and increased firing frequencies than did units with longer onset latencies. These comparative relationships persisted even though the activity of every motor unit was altered during pressure breathing. During positive-pressure breathing onset latencies were lengthened, and durations of firing were shortened with little change in onset or peak frequencies. Late units might be silenced. During negative-pressure breathing onset latencies were shortened, and durations of firing were lengthened, sufficiently in some cases to fill the expiratory pause. In addition, previously inactive units were recruited late in inspiration for short, relatively high frequency bursts during inspiration. The results support the concept that the phrenic motoneuron pool is comprised of three discrete subpopulations.