Aminophylline increases ventilation and diaphragm contractility in awake canines.

Traditional theophylline bronchodilators are still used clinically, especially in COPD. However, the effect of theophyllines on ventilation and respiratory muscles remains uncertain and these effects have not been measured directly in any awake, intact mammal. We hypothesized that aminophylline in the usual therapeutic dosage range, would elicit in the awake mammal, a significant increase in ventilation, and a significant increase in costal diaphragm shortening and contractility as recorded directly from the muscle. Therefore, we studied 13 awake canines, which had been chronically implanted with fine-wire EMG electrodes and sonomicrometer crystals in the costal segment of the diaphragm. Ventilatory parameters, moving average muscle EMG activity and muscle length and shortening, were measured at baseline and with aminophylline, during resting and hypercapnic stimulated breathing. Experiments were carried out prior to administration of aminophylline (baseline), and 1.5h after loading and ongoing infusion with aminophylline. Minute ventilation, tidal volume and respiratory frequency all increased significantly with aminophylline, both during resting breathing and at equivalent levels of hypercapnic stimulated breathing. Costal diaphragm baseline muscle length was entirely unchanged with aminophylline. Costal diaphragm shortening increased significantly with aminophylline while corresponding costal diaphragm EMG activity remained constant, consistent with increased diaphragm contractility. Thus, in awake, intact mammals, aminophylline in usual therapeutic dosage elicits increased ventilation and increased contractility of respiratory muscles.

Effect of salbutamol on respiratory muscle function and ventilation in awake canines.

The effect of the beta-agonist bronchodilator salbutamol on respiratory muscles and ventilation is uncertain. The presence of beta2 receptors on skeletal muscles and increased diaphragm contractility in vitro with salbutamol predict a significant effect that has not been confirmed, in vivo in non-fatigued diaphragm or in clinical studies using standard bronchodilator dosages. Therefore, we infused salbutamol at a higher dosage (23.3 microg/min) used clinically for treatment of respiratory emergencies, while measuring directly the length, shortening and EMG activation of costal and crural diaphragm, parasternal intercostal and transversus abdominis muscles, in 10 awake canines. At this salbutamol dosage, ventilation and tidal volume increased significantly during both resting and CO2-stimulated breathing. Salbutamol elicited significant increases in respiratory muscle shortening with much smaller increases in EMG activity, so the proportionally greater muscle shortening per unit EMG showed increased muscle contractility. The effects of salbutamol were not extinguished by inspiratory flow resistance or fluid challenge but were reversed specifically by the beta-blocker, propranolol. This study demonstrates that, in sufficient intravenous dosage, the beta-agonist salbutamol elicits increased ventilation and a beta2 receptor-mediated increase in contractility of respiratory muscles.

Aminophylline increases parasternal intercostal muscle activity during hypoxia in humans.

To clarify the mechanism of action of aminophylline on the hypoxic ventilatory response in humans, we analyzed the effects of aminophylline on respiratory neural output. To evaluate the respiratory neural output, we analyzed the electromyogram (EMG) of the parasternal intercostal muscle, one of the major inspiratory muscles, in eight healthy subjects. Both before and during aminophylline administration, measurements of ventilatory parameters with EMG recordings were conducted in room air, mild hypoxia (F(I)(o)(2) 0.15), and severe hypoxia (F(I)(o)(2) 0.11). Before administering aminophylline, hypoxic stimulation elicited ventilatory augmentation in a hypoxia-intensity dependent manner. Administration of aminophylline caused significant increases in ventilation (V (I)), tidal volume (V(T)), respiratory frequency (f(R)), and the respiration-related phasic moving averaged EMG amplitude (tidal EMG), at corresponding levels of hypoxia compared to before aminophylline. Augmentation patterns of hypoxia-induced increases in V(T) and tidal EMG showed close similarity. These results indicate that augmentation of hypoxic ventilatory response by aminophylline is mainly mediated by an increase in the respiratory neural drive in healthy humans.

Neck and abdominal muscle activity during a sniff.

Measurement of sniff nasal inspiratory pressure (SNIP) is now used widely as a simple, non-invasive assessment of global respiratory muscle strength, even though the technique evolved originally from measurements of trans-diaphragmatic pressure (Pdi) that reflect the status of the diaphragm. The relative participation of major respiratory muscles, apart from the diaphragm, in the generation of SNIP is not known. Therefore, we examined the activity during a sniff of both neck and abdominal "accessory" muscles. In seven young adults we implanted fine wire EMG electrodes under direct vision with high-resolution ultrasound into scalene, sternocleidomastoid, trapezius, and transversus abdominis. SNIP was measured during sniffs that were short and sharp, from low to maximal intensity, in both standing and supine postures. Mean maximum SNIP was -105.6cmH2O (SD 32.9) in supine and -94.5cmH2O (26.6) in the standing posture, (difference NS). In every subject, scalene activity appeared even at the lowest SNIP, and increased linearly with increasing SNIP. Sternomastoid activity appeared at higher SNIP levels in three of seven subjects. By contrast, trapezius activity was never present at low SNIP, and appeared in only 2 subjects at maximum SNIP. Sniff abdominal expiratory activity was inconsistent with no activity of transversus in four of seven subjects even at greatest SNIP. Thus, we observed differential activation among these non-diaphragm respiratory muscles during SNIP; while some accessory muscles were very active, others were unlikely to contribute to generation of SNIP. Clinically, this indicates SNIP will be impacted unequally by loss of function of specific respiratory muscles.

Geniohyoid muscle function in awake canines.

The geniohyoid (Genio) upper airway muscle shows phasic, inspiratory electrical activity in awake humans but no activity and lengthening in anesthetized cats. There is no information about the mechanical action of the Genio, including length and shortening, in any awake, nonanesthetized mammal during respiration (or swallowing). Therefore, we studied four canines, mean weight 28.8 kg, 1.5 days after Genio implantation with sonomicrometry transducers and bipolar electromyogram (EMG) electrodes. Awake recordings of breathing pattern, muscle length and shortening, and EMG activity were made with the animal in the right lateral decubitus position during quiet resting, CO2-stimulated breathing, inspiratory-resisted breathing (80 cmH2O. l-1. s), and airway occlusion. Genio length and activity were also measured during swallowing, when it shortened, showing a 9.31% change from resting length, and its EMG activity increased 6.44 V. During resting breathing, there was no phasic Genio EMG activity at all, and Genio showed virtually no movement during inspiration. During CO2-stimulated breathing, Genio showed minimal lengthening of only 0.07% change from resting length, whereas phasic EMG activity was still absent. During inspiratory-resisted breathing and airway occlusion, Genio showed phasic EMG activity but still lengthened. We conclude that the Genio in awake, nonanesthetized canines shows active contraction and EMG activity only during swallowing. During quiet or stimulated breathing, Genio is electrically inactive with passive lengthening. Even against resistance, Genio is electrically active but still lengthens during inspiration.

Simulation of acute spinal cord injury: effects on respiration.

The respiratory effects of acute spinal injury and paralysis are difficult to study. Urgent medical needs of human spinal cord injury victims usually preclude study, while induction of spinal cord lesions in awake animals is not feasible ethically. We utilized controlled, segmental infusion of epidural anesthetic in awake, highly trained, implanted canines to reversibly simulate the effects of thoracic and cervical (paraplegic and quadriplegic) spinal cord injury. We studied six animals, an average of 29 days after implantation with electromyogram and sonomicrometry transducers in transversus abdominis, external intercostal, parasternal intercostal and costal diaphragm muscles. Anesthetic was infused through an epidural catheter inserted percutaneously, under fluoroscopic guidance. Asymmetrical motor blockade was prevented using repositioning during epidural infusions. By sequential infusion we were able to induce three distinct, functional levels of spinal paralysis showing cumulative paralysis of abdominal, external intercostal, and parasternal intercostal muscles. Paralysis of the abdomen and chest wall, sparing only the diaphragm, showed unexpected bradypnea and failure to maintain minute ventilation.

Postinspiratory activity of the parasternal and external intercostal muscles in awake canines.

Previous studies have shown in awake dogs that activity in the crural diaphragm, but not in the costal diaphragm, usually persists after the end of inspiratory airflow. It has been suggested that this difference in postinspiratory activity results from greater muscle spindle content in the crural diaphragm. To evaluate the relationship between muscle spindles and postinspiratory activity, we have studied the pattern of activation of the parasternal and external intercostal muscles in the second to fourth interspaces in eight chronically implanted animals. Recordings were made on 2 or 3 successive days with the animals breathing quietly in the lateral decubitus position. The two muscles discharged in phase with inspiration, but parasternal intercostal activity usually terminated with the cessation of inspiratory flow, whereas external intercostal activity persisted for 24.7 +/- 12.3% of inspiratory time (P < 0.05). Forelimb elevation in six animals did not affect postinspiratory activity in the parasternal but prolonged postinspiratory activity in the external intercostal to 45.4 +/- 16.3% of inspiratory time (P < 0.05); in two animals, activity was still present at the onset of the next inspiratory burst. These observations support the concept that muscle spindles are an important determinant of postinspiratory activity. The absence of such activity in the parasternal intercostals and costal diaphragm also suggests that the mechanical impact of postinspiratory activity on the respiratory system is smaller than conventionally thought.

Postinspiratory activity of costal and crural diaphragm.

Because the first stage of expiration or "postinspiration" is an active neurorespiratory event, we expect some persistence of diaphragm electromyogram (EMG) after the cessation of inspiratory airflow, as postinspiratory inspiratory activity (PIIA). The costal and crural segments of the mammalian diaphragm have different mechanical and proprioceptive characteristics, so postinspiratory activity of these two portions may be different. In six canines, we implanted chronically EMG electrodes and sonomicrometer transducers and then sampled EMG activity and length of costal and crural diaphragm segments at 4 kHz, 10.2 days after implantation during wakeful, resting breathing. Costal and crural EMG were reviewed on-screen, and duration of PIIA was calculated for each breath. Crural PIIA was present in nearly every breath, with mean duration 16% of expiratory time, compared with costal PIIA with duration -2. 6% of expiratory time (P < 0.002). A linear regression model of crural centroid frequency vs. length, which was computed during the active shortening of inspiration, did not accurately predict crural EMG centroid frequency values at equivalent length during the controlled relaxation of postinspiration. This difference in activation of crural diaphragm in inspiration and postinspiration is consistent with a different pattern of motor unit recruitment during PIIA.

Screening for hypothyroidism in sleep apnea.

Primary sleep apnea-hypopnea syndrome (obstructive sleep apnea [OSA]) and hypothyroidism have many signs and symptoms in common. The overlap in clinical presentation, and the sleep-disordered breathing that can accompany hypothyroidism, create a significant risk of misdiagnosis of sleep apnea among patients referred to sleep clinic who have undiagnosed hypothyroidism. We determined the point prevalence of hypothyroidism in our sleep clinic patients with newly diagnosed sleep-disordered breathing. Of 290 sequential patients referred to sleep clinic, 200 (69%) patients judged at high risk for OSA underwent polysomnography (PSG) and biochemical screening for hypothyroidism. Of these, 124 (62%) were judged to have sleep apnea. This included three patients (1.5% of patients undergoing PSG or 2. 4% of those judged to have OSA) who were also discovered to have previously undiagnosed hypothyroidism. These three patients with "secondary" sleep apnea were treated with thyroxine therapy alone, and followed with sequential sleep studies and serum thyroid hormone assays; symptoms, sleep-disordered breathing, nocturnal hypoxia, and thyroid deficiency resolved simultaneously. We conclude that biochemical screening for hypothyroidism is required to prevent inadvertent misdiagnosis of hypothyroid sleep-disordered breathing as primary sleep apnea, and that it is a cost-effective component of the investigation of sleep apnea.

Posture effects on timing of abdominal muscle activity during stimulated ventilation.

In humans during stimulated ventilation, substantial abdominal muscle activity extends into the following inspiration as postexpiratory expiratory activity (PEEA) and commences again during late inspiration as preexpiratory expiratory activity (PREA). We hypothesized that the timing of PEEA and PREA would be changed systematically by posture. Fine-wire electrodes were inserted into the rectus abdominis, external oblique, internal oblique, and transversus abdominis in nine awake subjects. Airflow, end-tidal CO2, and moving average electromyogram (EMG) signals were recorded during resting and CO2-stimulated ventilation in both supine and standing postures. Phasic expiratory EMG activity (tidal EMG) of the four abdominal muscles at any level of CO2 stimulation was greater while standing. Abdominal muscle activities during inspiration, PEEA, and PREA, were observed with CO2 stimulation, both supine and standing. Change in posture had a significant effect on intrabreath timing of expiratory muscle activation at any level of CO2 stimulation. The transversus abdominis showed a significant increase in PEEA and a significant decrease in PREA while subjects were standing; similar changes were seen in the internal oblique. We conclude that changes in posture are associated with significant changes in phasic expiratory activity of the four abdominal muscles, with systematic changes in the timing of abdominal muscle activity during early and late inspiration.

Eliminating ultrasonic interference from respiratory muscle EMG.

Fine wire recordings of the respiratory muscle electromyogram are often employed to represent muscle activity, and recently ultrasound-sonomicrometry has become a common method of measuring length of respiratory muscles in both acute and chronic preparations. Although recording both EMG and sonomicrometry simultaneously has become standard practice, there has not been any consideration of the potential confounding influence of ultrasound noise upon the recorded EMG spectrum. Activation of the sonomicrometry-ultrasound tranducer introduces a high frequency, high amplitude voltage pulse plus harmonics, which can contaminate the EMG spectrum directly, as well as through aliasing when EMG is sampled directly digitally. We describe the use of a new, combined, wing stabilized sonomicrometry- and EMG measurement transducer to characterize exactly the influence of ultrasound upon the crural diaphragm EMG spectrum, and the development of digital filtering techniques which effectively eliminate the ultrasound interference. Two alternative methods of avoiding ultrasound-EMG interference are also considered. The isolation and elimination of ultrasound-sonomicrometry signal interference may be important in studies where EMG and length are measured together.

Diaphragm function during sighs in awake dogs after laparotomy.

Pulmonary complications after upper abdominal surgery are usually ascribed to temporary postoperative impairment of diaphragm function, which may not originate from intrinsic, structural injury but from reflex inhibition of diaphragm contractility. Spontaneous breathing is interrupted periodically by sighs, even after upper abdominal surgery. If postoperative dysfunction of the diaphragm arises from a reflexic inhibition, then the sigh should temporarily override the inhibition and restore normal diaphragm function. We implanted sonomicrometer and electromyogram transducers chronically in six dogs by laparotomy, then directly measured length, shortening, and electromyogram activity of costal and crural diaphragm segments, parasternal intercostal, and transversus abdominis muscles an average of 8.7 (range, 1-16) d later during resting tidal breathing and sighs. In each animal we analyzed a sequence of breaths, including a sigh, when costal or crural diaphragm contractility was abnormal. With each sigh, the shape and amplitude of costal and crural diaphragm segmental shortening improved abruptly, from 0.9 and 1.4% of baseline length (% LBL) during resting breathing to 12.1 and 11.1% LBL, respectively, during sighs. The sighs were compared to CO2-stimulated breaths of equivalent tidal volume, which did not show either pattern or amplitude of shortening equivalent to sighs. We conclude that diaphragm dysfunction after laparotomy arises from a reflex inhibition, which is overridden abruptly to return diaphragm function briefly to normal during each spontaneous sigh.

A modified Bessel filter for amplitude demodulation of respiratory electromyograms.

We studied a device that is commonly used for amplitude demodulation of respiratory muscle electromyograms (EMG). This device contains a rectifier and a low-pass filter called a modified third-order Paynter filter. We characterized this filter and found that it has good transient characteristics that suit its task as an EMG demodulator, but it has poor high-frequency attenuation that passes interfering, higher frequency components to the output waveform. Therefore, we designed and constructed a new filter with transient characteristics that are comparable to those of the modified Paynter filter but with superior high-frequency attenuation. This new filter is a modified seventh-order Bessel filter. We also identified a simple technique to convert an existing modified Paynter filter back to an original Paynter filter. The original Paynter filter has a wider pass band than the modified Paynter filter but superior stop-band attenuation.

Differential respiratory activity of four abdominal muscles in humans.

Together the abdominal muscles contribute significantly to ventilation under some conditions, but there is little information regarding individual recruitment and timing of activation of the four abdominal muscles in humans. Fine-wire electrodes were inserted under direct vision guided by high-resolution ultra-sound into the rectus abdominis (Rectus), external oblique (Extern), internal oblique (Intern), and transversus abdominis (Transv) in nine awake healthy subjects. Airflow, end-tidal CO2, and moving-average EMG signals were recorded during 1) supine resting and CO2-stimulated ventilation and 2) resting ventilation in the standing position. During resting supine breathing, Transv showed significant phasic EMG activity during expiration. As posture changed from supine to standing, phasic activity during resting ventilation was greatest in Transv, with lesser activity in Intern and Extern, while Rectus remained inactive. As CO2 began to increase, Transv was activated first, followed by Intern, the Extern, and finally Rectus. With moderate CO2 stimulation, Transv and Intern were more active than was Extern and Rectus remained least active. EMG activities in the expiratory muscles after cessation of expiratory flow (postexpiratory expiratory activity) and in expiratory muscle activity preceding expiratory flow were observed consistently during supine stimulated ventilation and standing resting ventilation. These activities before and after expiratory airflow were prominent with stimulated ventilation during a substantial portion of inspiration, suggesting dual control of inspiratory pump action by both inspiratory and expiratory muscles, which provide acceleration and braking actions, respectively. These results suggest that in awake humans 1) during resting ventilation, expiration is an active process; 2) abdominal muscles are activated differentially; 3) Transv is the most active, Intern and Extern are intermediate, and Rectus is the least active expiratory muscle; and 4) during stimulated ventilation, inspiratory and expiratory muscles contribute dually to inspiratory pump action.

Activity of costal and crural diaphragm during progressive hypoxia or hypercapnia.

Because costal and crural diaphragm segments have different functional characteristics, ventilatory stimulation with hypoxia or hypercapnia may elicit differential segmental function. We report measurements of diaphragm segmental length, shortening, and electromyogram (EMG) activity from 11 canines that were chronically implanted with sonomicrometry transducers and EMG electrodes and then studied a mean of 18 days postimplantation while awake and breathing spontaneously during CO2 rebreathing and progressive isocapnic hypoxia. Ventilatory responses to hypercapnia and progressive hypoxia were moderate at 1.13 +/- 0.31 (SD) 1. min-1. mm-1 arterial Pco2 and -0.98 +/- 0.51 l. min-1.%arterial O2 saturation-1. When tidal values for breathing pattern and segmental function were compared at matching tidal volumes that correspond to mean CO2 of 49.4 arterial Pco2 and 77% arterial O2 saturation, there was no significant difference in resting length, tidal shortening, or tidal EMG of costal or crural segments. Intrabreath profiles of flow, shortening, and EMG activity at matched tidal volumes showed that 1) inspiratory flow during hypoxia was significantly greater during early inspiration, 2) crural EMG activity preceded costal EMG activity in early inspiration during both hypercapnia and hypoxia, 3) both segments showed increased postinspiratory inspiratory activity with stimulated ventilation, and 4) postinspiratory shortening and EMG were greatest for the crural segment during hypoxia. These results suggest that costal and crural diaphragm segments exhibit differential function during chemical stimulation, especially during postinspiration.

Respiratory muscle compensation for unilateral or bilateral hemidiaphragm paralysis in awake canines.

In humans and some animals, the surviving respiratory muscles are able to compensate fully for unilateral, and partially for bilateral, hemidiaphragm paralysis. To examine differential activity of individual respiratory muscles after unilateral or bilateral diaphragm paralysis, length and electromyogram (EMG) of left costal and crural diaphragm segments, parasternal intercostal, and transversus abdominis were measured directly in five awake canines after implantation with sonomicrometry transducers and bipolar EMG electrodes under three conditions: during normal breathing (NOFRZ), after infusion of local anesthetic (bupivacaine) through a cervical phrenic nerve cuff to induce reversible contralateral hemidiaphragm (CNFRZ), and after bilateral diaphragm (BIFRZ) paralysis. From NOFRZ to CNFRZ, costal, crural, parasternal, and transversus abdominis increased shortening and EMG activity to compensate for contralateral diaphragm paralysis, but the increase in activity was not equivalent for each muscle. With BIFRZ, parasternal and transversus abdominis showed further increases in activity, coordinated between both inspiration and expiration. Normalized intrabreath profiles revealed dynamic differences in development of muscle activity within each breath as paralysis worsened. Review of simultaneous muscle activities showed coordinated interactions among the compensating muscles: passive shortening of transversus, and lengthening of costal and crural, coincided with increased active inspiratory shortening of parasternal. We conclude that an integrated strategy of respiratory muscle compensation for unilateral or bilateral diaphragm paralysis occurs among chest wall, abdominal, and diaphragm segmental muscles, with relative contributions of individual muscles adjusted according to the degree of diaphragm dysfunction.

Costal and crural diaphragm function during panting in awake canines.

During natural panting for thermal regulation, the pattern of activation of the major respiratory muscles, including costal and crural diaphragm segments, is not known. We measured diaphragm segmental length, shortening, and electromyographic (EMG) activity in five chronically implanted canines awake and breathing spontaneously at rest and during a mild dry heat stress. During panting, minute ventilation increased fourfold from 5.07 l/min and respiratory rate increased from 16.9 to 192.8 breaths/min or 3.2 Hz. During panting, end-expiratory length of both costal and crural segments decreased, concurrent with significant increases in end-expiratory EMG. With the onset of panting, tidal costal shortening decreased significantly from 6.29% of end-expiratory length to 3.54%, whereas crural shortening decreased from 6.04 to 2.46%. Meanwhile, segmental EMG tended to increase during panting. During panting, intrabreath costal and crural segmental function revealed differential activation; the costal segment shortened in concert with inspiratory flow, whereas peak crural shortening occurred in expiration, almost 180 degrees out of phase with costal. The divergence in segmental shortening during panting was accompanied by a lesser shift in timing of segmental EMG. In the awake spontaneously panting canine, asynchronous costal and crural shortening may enhance gas mixing in a manner analogous to high-frequency ventilation.

Respiratory muscle function during emesis in awake canines.

Emesis requires a coordinated differential recruitment of gastrointestinal smooth muscle, upper airway muscles, and several muscles involved in respiration. In seven awake intact canines we measured the electrical activity (electromyogram) and shortening of costal and crural diaphragm segments, parasternal intercostal, and transversus abdominis during emesis that was induced by instillation of apomorphine into the lower conjunctival fornix. The process of emesis was tightly coordinated with ventilation and showed four respiratory phases: baseline ventilation (Base), initial preemetic hyperventilation (Hyperv), prodromal ventilation associated with salivation and probable nausea (Prodrome), and finally retching and expulsion (Expel) of gastric contents. Ventilation was suppressed during expulsive events, but a small inspiratory airflow was interjected between expulsions. Resting electromyogram of all four muscles increased during the process of emesis, with costal and crural segments showing a marked decrease in resting length through Prodrome and Expel. To produce an expulsive maneuver, both inspiratory and expiratory muscles were activated synchronously, unlike their usual sequential activation during ventilation, with costal and crural segments and transversus abdominis showing the most shortening. The crural segment showed a biphasic length change with initial shortening and then lengthening to assist esophageal sphincter function during Expel. These results indicate a strong coordinated interaction between brain stem centers responsible for control of respiration and of emesis.

Differential function of the costal and crural diaphragm during emesis in canines.

In six mongrel dogs under thiopental anesthesia, piezoelectric transducers and bipolar electromyographic (EMG) wires were installed onto left costal, medial crural and lateral crural segments of the diaphragm. During CO2 rebreathing, shortening and EMG activity increased significantly in all three regions of the diaphragm compared to resting breathing. During emesis, (1) both shortening and EMG activity significantly increased compared to resting in costal segment; however, (2) lateral crural shortening was not increased in spite of significant increase in EMG activity; furthermore, (3) the medial crural segment lengthened without any increased EMG activity. These results demonstrate a differential recruitment of costal and crural diaphragm segments, and an additional differential activity within the crural segment between medial and lateral crural regions, during emesis. This activity of the canine diaphragm is consistent with a central influence of emesis upon individual regions of the diaphragm.

Costal and crural diaphragm function during CO2 rebreathing in awake dogs.

If costal and crural diaphragm segments can perform as separate muscles, then CO2-stimulated ventilation may elicit differential segmental function. We studied diaphragm segmental length, shortening, and electromyogram (EMG) activity in 10 awake dogs chronically implanted with sonomicrometer transducers and EMG electrodes. During CO2 rebreathing, segmental shortening and EMG activity per whole tidal breath progressively increased, but segmental responses could not be differentiated at any level of CO2. With increasing CO2, resting end-expiratory length of both diaphragm segments increased. During the complete intrabreath inspiratory-expiratory cycle, costal and crural diaphragm revealed distinctive segmental function. At rest, crural shortening exceeded costal shortening in earliest inspiration, costal and especially crural shortening persisted into early expiration, and EMG activity of the crural segment was greater than that of the costal segment in earliest inspiration and showed more end-inspiratory/early expiratory [post-inspiratory inspiratory activity (PIIA)] activity. During CO2-stimulated breathing, neither segment shortened during the inspiratory flow of earliest inspiration. During CO2 rebreathing, shortening of the crural segment exceeded that of the costal segment during early inspiration and outlasted costal shortening during expiration; for both segments, shortening persisted after termination of inspiratory airflow. With increased CO2, EMG activity of the crural segment preceded that of the costal segment in earliest inspiration and was dominant into expiration, whereas costal EMG activity terminated abruptly with inspiratory flow. Thus, costal EMG PIIA was not evident during hypercapnia, whereas crural EMG PIIA was significant.(ABSTRACT TRUNCATED AT 250 WORDS)