Production of 20-HETE and its role in autoregulation of cerebral blood flow.

In the brain, pressure-induced myogenic constriction of cerebral arteriolar muscle contributes to autoregulation of cerebral blood flow (CBF). This study examined the role of 20-HETE in autoregulation of CBF in anesthetized rats. The expression of P-450 4A protein and mRNA was localized in isolated cerebral arteriolar muscle of rat by immunocytochemistry and in situ hybridization. The results of reverse transcriptase-polymerase chain reaction studies revealed that rat cerebral microvessels express cytochrome P-450 4A1, 4A2, 4A3, and 4A8 isoforms, some of which catalyze the formation of 20-HETE from arachidonic acid. Cerebral arterial microsomes incubated with [(14)C]arachidonic acid produced 20-HETE. An elevation in transmural pressure from 20 to 140 mm Hg increased 20-HETE concentration by 6-fold in cerebral arteries as measured by gas chromatography/mass spectrometry. In vivo, inhibition of vascular 20-HETE formation with N-methylsulfonyl-12, 12-dibromododec-11-enamide (DDMS), or its vasoconstrictor actions using 15-HETE or 20-hydroxyeicosa-6(Z),15(Z)-dienoic acid (20-HEDE), attenuated autoregulation of CBF to elevations of arterial pressure. In vitro application of DDMS, 15-HETE, or 20-HEDE eliminated pressure-induced constriction of rat middle cerebral arteries, and 20-HEDE and 15-HETE blocked the vasoconstriction action of 20-HETE. Taken together, these data suggest an important role for 20-HETE in the autoregulation of CBF.

Important role of carotid chemoreceptor afferents in control of breathing of adult and neonatal mammals.

This review provides a summary and prospective on the importance of carotid/peripheral chemoreceptors to the control of breathing during physiologic conditions. For several days after carotid body denervation (CBD), adult mammals hypoventilate (+10 mmHg increase in Pa(CO(2))) at rest and during exercise and CO(2) sensitivity is attenuated by about 60%. In addition, if the rostral ventrolateral medulla is cooled during NREM sleep after CBD, a sustained apnea is observed. Eventually, days or weeks after CBD, a peripheral ventilatory chemoreflex redevelops and there is a normalization of breathing (rest and exercise) and CO(2) sensitivity. The site (s) of the regained chemosensitivity has not been established. This plasticity/redundancy after CBD appears greater in neonates than in adult mammals. These data suggest the carotid and other peripheral chemoreceptors provide an important excitatory input to medullary respiratory neurons that is essential for breathing when wakeful stimuli and central chemoreceptors are absent.

Effects on breathing of carotid body denervation in neonatal piglets.

The purpose of these studies was to test the hypothesis that carotid chemoreceptor activity is necessary for postnatal maturation of the ventilatory control system. By using a lateral surgical access, 17 piglets were carotid body denervated (CBD) and 14 were sham denervated at 3-25 days of age. After surgery, there was no irregular breathing in any group. There was no significant hypoventilation when CBD was performed at less than 5 days of age (n = 5) and only a mild (arterial PCO(2) 5 Torr; P < 0.05) to moderate, transient (arterial PCO(2) 8 Torr; P < 0.5) hypoventilation in piglets denervated at 10-15 (n = 6) and 20-25 (n = 6) days of age, respectively. Three weeks after surgery, both breathing of a hypoxic gas mixture and jugular venous NaCN injections elicited a hyperpnea in the CBD piglets that was attenuated compared with that in sham CBD piglets. In the CBD piglets, there was no response to injections of NaCN in the carotid arteries, but there was a response to NaCN injected into the proximal descending aorta, suggesting the residual peripheral chemosensitivity was of aortic origin. Carotid chemoreceptor-intact piglets had carotid and aortic NaCN chemosensitivity by 2 days of age. The carotid response persisted for the 40 days of the study, but the aortic reflex persisted only until approximately 8 days of age. We conclude that 1) the major effect of CBD per se in neonatal piglets is age-dependent hypoventilation and 2) there is a high degree of plasticity in peripheral chemosensitivity in neonates that may contribute to minimizing the changes in breathing after CBD.

Effect of carotid body denervation on breathing in neonatal goats.

The objective of the present study was to determine in goats whether carotid body denervation (CBD) at 1-3 days of age causes permanent changes in breathing greater than those that occur after CBD in adult goats. Goats underwent CBD (n = 6) or sham CBD (n = 3) surgery at 1-3 days of age. In addition, one unoperated control animal was studied. Bolus intravenous injections of NaCN 2 days postsurgery verified successful CBD surgery. However, at 3, 11, and 18 mo of age, the CBD goats had regained a NaCN response that did not differ (P > 0.10) from that of intact goats. Intracarotid NaCN injections elicited a hyperpnea in the sham CBD but not the CBD goats. Only one animal exhibited highly irregular breathing [characterized by prolonged (>9-s) apneas] after CBD, and the irregularity disappeared by 3 mo of age. One CBD goat died at 35 days of age, and autopsy revealed that death was associated with pneumonia. After 3 mo of age, there were no statistically significant differences (P > 0.10) between sham and CBD goats in eupneic breathing, hypoxia and CO(2) sensitivity, and the exercise hyperpnea. It is, therefore, concluded that CBD at 1-3 days of age in goats does not appear to affect selected aspects of respiratory control after 3 mo of age, conceivably because of the emergence of other functional chemoreceptors that compensate for the loss of the carotid chemoreceptor.

Important role of carotid afferents in control of breathing.

The purpose of the present study was to determine the effect on breathing in the awake state of carotid body denervation (CBD) over 1-2 wk after denervation. Studies were completed on adult goats repeatedly before and 1) for 15 days after bilateral CBD (n = 8), 2) for 7 days after unilateral CBD (n = 5), and 3) for 15 days after sham CBD (n = 3). Absence of ventilatory stimulation when NaCN was injected directly into a common carotid artery confirmed CBD. There was a significant (P < 0.01) hypoventilation during the breathing of room air after unilateral and bilateral CBD. The maximum PaCO2 increase (8 Torr for unilateral and 11 Torr for bilateral) occurred approximately 4 days after CBD. This maximum was transient because by 7 (unilateral) to 15 (bilateral) days after CBD, PaCO2 was only 3-4 Torr above control. CO2 sensitivity was attenuated from control by 60% on day 4 after bilateral CBD and by 35% on day 4 after unilateral CBD. This attenuation was transient, because CO2 sensitivity returned to control temporally similar to the return of PaCO2 during the breathing of room air. During mild and moderate treadmill exercise 1-8 days after bilateral CBD, PaCO2 was unchanged from its elevated level at rest, but, 10-15 days after CBD, PaCO2 decreased slightly from rest during exercise. These data indicate that 1) carotid afferents are an important determinant of rest and exercise breathing and ventilatory CO2 sensitivity, and 2) apparent plasticity within the ventilatory control system eventually provides compensation for chronic loss of these afferents.

Breathing of awake goats during prolonged dysfunction of caudal M ventrolateral medullary neurons.

Cooling the caudal M ventrolateral medullary (VLM) surface for 30 s results in a sustained apnea in anesthetized goats but only a 30% decrease in breathing in awake goats. The purpose of the present study was to determine, in the awake state, the effect of prolonged (minutes, hours) caudal M neuronal dysfunction on eupneic breathing and CO2 sensitivity. Dysfunction was created by ejecting excitatory amino acid receptor antagonists or a neurotoxin on the VLM surface through guide tubes chronically implanted bilaterally on a 10- to 12-mm2 portion of the caudal M VLM surface of 12 goats. Unilateral and bilateral ejections (1 microliter) of selective antagonists for N-methyl-D-aspartic acid or non-N-methyl-D-aspartic acid receptors had no significant effect on eupneic breathing or CO2 sensitivity. Unilateral ejection of a nonselective excitatory amino acid receptor antagonist generally had no effect on eupneic breathing or CO2 sensitivity. However, bilateral ejection of this antagonist resulted in a significant 2-Torr hypoventilation during eupnea and a significant reduction in CO2 sensitivity to 60 +/- 9% of control. Unilateral ejection of the neurotoxin kainic acid initially stimulated breathing; however, breathing then returned to near control with no incidence of apnea. After the kainic acid ejection, CO2 sensitivity was reduced significantly to 60 +/- 7% of control. We conclude that in the awake state a prolonged dysfunction of caudal M VLM neurons results in compensation by other mechanisms (e.g., carotid chemoreceptors, wakefulness) to maintain near-normal eupneic breathing, but compensation is more limited for maintaining CO2 sensitivity.

Effect on breathing of surface ventrolateral medullary cooling in awake, anesthetized and asleep goats.

In adult and neonatal goats, we chronically implanted thermodes on the ventrolateral (VLM) medullary surface to create reversible neuronal dysfunction and thereby gain insight into the role of superficial VLM neurons in control of breathing in anesthetized, awake and asleep states. Consistent with data of others, cooling caudal area M and rostral area S caused sustained apnea under anesthesia. However, in the awake and NREM sleep states, cooling at this site caused only a modest reduction in breathing, indicating that neurons at this site are not critical for respiratory rhythm in these states. Moreover, data in the awake state over multiple conditions suggest neurons at this site do not integrate all intracranial and carotid chemoreception. The data suggest though that neurons at this site have a facilitatory-like effect on breathing both unrelated and related to intracranial chemoreception. We believe that this facilitation serves a function similar to the facilitation provided by the carotid chemoreceptors and by sources associated with wakefulness. Accordingly, elimination/attenuation of any one of these three influences (caudal M rostral S VLM, wakefulness, carotid chemoreception) results in a slight decrease in breathing, removal of two of the three results in a greater decrease in breathing, and removal of all three results in sustained apnea.

Effects of cooling the ventrolateral medulla on diaphragm activity during NREM sleep.

Dysfunction through cooling of neurons near the ventrolateral medullary (VLM) surface results in apnea in the anesthetized state, whereas similar neuronal dysfunction in the awake state only modestly decreases breathing. The purpose of this study was to investigate effects on breathing, as measured by diaphragm electromyogram (EMGdi), of VLM neuronal dysfunction during NREM sleep, a naturally occurring change in state. In six goats, thermodes for cooling were chronically implanted between the first hypoglossal rootlet and the pontomedullary junction (area M and area S). During wakefulness and NREM sleep, bilateral VLM cooling (thermode temp = 20 degrees C) for 30 sec decreased EMGdi mean activity and minute EMGdi (p < 0.05) and lengthened the time between diaphragm contractions. During NREM sleep, reductions in mean and minute EMGdi during cooling tended to be greater than during waking, but not significantly. However, following carotid body denervation. VLM cooling caused prolonged apnea during NREM sleep but only a brief apnea in the awake state. The data suggest that either intact VLM neuronal mechanisms or intact carotid afferents are necessary for sustained EMGdi activity during NREM sleep.

Effect on breathing of ventral medullary surface cooling in neonatal goats.

The present study was designed to determine whether neurons near the ventral medullary surface (VMS) that are important to control of breathing in adult mammals are also important to control of breathing in neonates. In 7-day-old goats (n = 22), the VMS was surgically exposed under halothane anesthesia. Stainless steel thermodes (2 x 2 mm) were used to cool (20 degrees C) and thereby create neuronal dysfunction of discrete VMS sites. Bilateral cooling under anesthesia 0-2 or 2-4 mm lateral to the midline between the exit of cranial nerves VI and XII resulted in a reduction (P < 0.05) of breathing and most often in apnea. Cooling caudal or rostral to this area did not have a consistent effect on breathing. In 7-day-old goats (n = 8), 3 x 3-mm thermodes were chronically implanted bilaterally on the VMS surface between the exit of cranial nerves VI and XII. The goats recovered and were studied over several days thereafter. VMS cooling while the goats were awake caused breathing to decrease (P < 0.05), but apnea was never observed. The decrease was less (P < 0.05) than while the goats were anesthetized. After 10 s of cooling, the hypopnea while the goats were awake was uniform during eupnea, hypercapnia, hyperoxia, and hypoxia, but after 10 s of cooling, the decrease was relatively greater (P < 0.05) during hyperoxia and hypercapnia. These effects of VMS cooling are qualitatively the same as in adult goats; thus the data are consistent with mature VMS contribution to the control of breathing in neonatal goats.

Effect of theophylline on ventilatory roll-off during hypoxia in goats.

The increase in pulmonary ventilation (VE) during the first minutes of hypoxia is not sustained as after several minutes VE decreases or "rolls-off" toward control levels. We hypothesized that intravenous infusion of theophylline, by blocking the central inhibitory effects on breathing of adenosine, would attenuate the hypoxic VE roll-off. Twelve unanesthetized adult goats were exposed for 20 min to a 12% O2-88% N2 gas mixture. In some studies,theophylline was infused intravenously (IV) for 20 min before and during the hypoxia. The highest infusion rate of 6.0--8.0 mg/min was sufficient to totally prevent the arterial hypertension and bradycardia that occurred with IV infusion of 4 mg min (-1) of adenosine. Nine of the 12 goats demonstrated VE roll-off without the theophylline infusion. In goats that demonstrated VE roll-off without theophylline, a significant (P < 0.05) VE roll-off was observed even at the highest theophylline infusion rate. We therefore conclude that the VE roll-off during hypoxia is not primarily or critically mediated by adenosine in awake, adult goats.

Ventral medullary surface activity during hypoxia in awake and anesthetized goats.

The rostral ventrolateral medullary surface (VMS) plays a major state-dependent role in the control of breathing; its role during hypoxia remains speculative. We therefore assessed activity within the rostral VMS by measuring reflectance of scattered light in 5 goats during normoxia, hypoxia, and hyperoxia in awake and halothane anesthetic states. Within the first minute of hypoxia, light reflectance began to decrease in the awake state; reflectance reached a stable nadir within 30 min about 10 and 17% below control values (P < 0.01), at 12 and 10% inspired O2, respectively. In the anesthetized state, reflectance decreased (P < 0.01) by 6% at 10% inspired O2. After 30 min in the awake state, reflectance returned (P < 0.01) toward control values, reaching a stable level at 7 and 11% below control at 12 and 10% inspired O2, respectively (P < 0.05). Hyperoxia resulted in a 1% increase (P < 0.05) in reflectance. Changes in reflectance during hypoxia did not consistently parallel changes in breathing, heart rate, or arterial blood pressure. We conclude that, a) decreased reflectance during hypoxia results, in part, from increased neural activity, and b) state exerts a substantial effect on the response of VMS areas to hypoxia.

Effect of dichloroacetate on PaCO2 responses to hypoxia in awake goats.

To gain insight into the role of cerebral lactic acidosis in the hypoxic ventilatory response, we administered dichloroacetate (DCA) intravenously to inhibit lactic acid production in 7 awake goats (40-70 kg) during 0.5 h of normoxia (inspired O2 fraction = 0.209) and 5 h of poikilocapnic hypoxia (inspired O2 fraction = 0.125). On separate days, these goats were also studied with a continuous saline infusion (18 ml/h iv) during 5 h of normoxia and hypoxia. Arterial PCO2 (PaCO2) did not change during the 5-h normoxic period. During hypoxia, arterial PO2 fell significantly (P < 0.05) with both saline (from 111.3 to 39.0 Torr) and DCA (from 111.8 to 42.0 Torr) infusions. PaCO2 decreased (P < 0.05) during the first 0.5 h of both the saline and DCA hypoxia protocols. The decrease was greater (P < 0.05) during DCA (from 36.5 to 33.5 Torr) than during saline infusion (from 37.7 to 36.3 Torr). With saline infusion, PaCO2 decreased (P < 0.05) by 4.9 Torr between 0.5 and 5.0 h of hypoxia. However, over this period of DCA hypoxia, PaCO2 did not significantly decrease (P > 0.05). We conclude that the enhanced hyperventilation with DCA during acute hypoxia is consistent with brain lactic acidosis depressing breathing. Absence of additional significant hyperventilation after 0.5 h of DCA hypoxia suggests that a time-dependent alleviation of brain lactic acidosis might normally contribute to ventilatory acclimatization to hypoxia.

Imaging of VMS activity during blood pressure challenges in awake and anesthetized goats.

We examined scattered-light changes in a rostral ventral medullary surface (VMS) area from five goats after blood pressure challenges during waking and halothane anesthesia. Reflected 660-nm images were digitized at 1/s after baseline; intravenous saline; 5, 10, or 15 micrograms/kg phenylephrine administration; or sodium nitroprusside infusion sufficient to lower blood pressure by 50%. Phenylephrine elicited a dose-dependent, blood pressure elevation during both states and a substantial transient reflectance increase (interpreted as activity decline) during anesthesia, but only a minimal, long-latency, slow-reflectance decrease activity increase) during waking. Sodium nitroprusside elicited lowering of blood pressure and decreased reflectance in the rostral site during anesthesia. The magnitude of the reflectance change to depressor challenge increased 30%, and the onset latency shortened during waking. Isolated regions of enhanced reflectance changes appeared during both challenges. Activity in this rostral VMS site differentially responds to blood pressure elevation or lowering, and state markedly alters the responses. We speculate that VMS responses to depressor challenge represent reflex activation of respiratory regions.

Effect on breathing of neuronal dysfunction in the caudal ventral medulla of goats.

It has been reported that the caudal ventrolateral medulla (VLM) is important in central chemoreception and the control of breathing. The objective of this study was to determine in adult goats the effects on breathing of neuronal dysfunction of this caudal VLM region (area L; caudal to rostral hypoglossal nerve rootlet). Thermodes were chronically implanted on the VLM to cool neurons and thereby cause neuronal dysfunction. During awake and (halothane) anesthetized states, cooling the caudal VLM for 20 s to 20 degrees C did not alter breathing (P > 0.10). However, between 20 and 30 s of cooling and during recovery from cooling 0-4 mm caudal to the rostral hypoglossal rootlet, there was a 12 (awake) to 25% (anesthetized) increase (P < 0.05) in breathing. This tachypneic hyperpnea was uniform over conditions of eucapnia, hypercapnia, and hypoxia and resulted from reduced inspiratory time that increased frequency. We conclude that in goats inhibitory neurons are located in area L and the lateral caudal ventral medulla.

Effect of carotid chemoreceptor denervation on breathing during ventrolateral medullary cooling in goats.

It has been postulated that the so-called area S of the ventrolateral medulla (VLM) integrates peripheral chemoreceptor activity; thus cooling-induced dysfunction of neurons in this VLM area should functionally eliminate carotid chemoreceptor stimulation of breathing. Accordingly, carotid chemoreceptor denervation (CBD) should not alter the breathing effects of VLM neuronal dysfunction. To test this hypothesis in awake goats, chronically implanted thermodes were used to cool the VLM and thereby cause reversible neuronal dysfunction in all or portions of VLM areas M and S. Within 5 s after initiation of cooling approximately 60-100% of areas M and S in (P < 0.05) uniformly over conditions of eupnea, hypercapnia, and hypoxia. Between 10 and 20 s of cooling, the reduction in VI was approximately 10% greater (P < 0.05) during hypercapnia than during eupnea and hypoxia. For the remaining 10 s of cooling and for approximately 1 min after cooling, VI increased to and above control for all conditions. For all conditions, CBD accentuated the depression of VI during cooling, causing VI to decrease (P < 0.05) 10-40% more than before CBD. After CBD, the greatest effect on VI of cooling was again during hypercapnia. Thus the carotid bodies in intact goats appear to sense blood gas errors caused during VLM cooling to minimize the decreases in VI. We conclude that the data from this study do not support the concept that the VLM integrates carotid chemoreceptor activity.

Effect of multiple denervations on the exercise hyperpnea in awake ponies.

In three previously reported studies, we had documented that the normal exercise hyperventilation in ponies is accentuated by carotid body denervation (CBD), not affected by hilar nerve pulmonary vagal denervation (HND), and mildly attenuated by spinal cord ablation of the dorsal lateral columns at L2 (SA). In the present study, we hypothesized that if redundancy of control existed in exercising ponies, then multiple denervations of theoretically important pathways in the same animal might attenuate the ventilatory response to exercise in a way not predictable by the individual lesion experiments alone. There were three major findings in the various combinations of CBD, HND, and SA in ponies during treadmill exercise. First, the combination of CBD with HND or SA resulted generally in an accentuation of the hypocapnia during exercise that was predictable on the basis of CBD alone. However, in one pony that showed a hypercapnic exercise response after SA alone, CBD subsequently caused a greater exercise hypercapnia. Second, HND in a CBD or SA pony did not affect the exercise arterial PCO2 response, which is consistent with previous data showing the lack of an HND effect in otherwise intact ponies. Third, in ponies with all three denervations together, the predominant response was an increase, not a decrease, in the exercise hyperventilation; this increase was greater than that predicted from the individual lesions. We conclude that these data do not provide evidence of redundancy in mechanism for the exercise hyperpnea other than instances of carotid chemoreceptor error sensing when hypercapnia occurs during exercise.

Rostral ventral medullary surface activity during hypercapnic challenges in awake and anesthetized goats.

Regions within the rostral ventral medullary surface (RVMS) play an important role in cardiorespiratory responses to CO2 during anesthesia. Activity within a RVMS area, in which local cooling elicited marked ventilatory and blood pressure reductions, was measured as 660 nm scattered light changes in 5 goats following 5% CO2 challenges during waking and anesthetic states. During wakefulness, hypercapnia elicited a substantial, short latency transient (1-1.5 min) activity increase, followed by a sustained decrease. Stimulus cessation elicited a large and rapid off-transient activity increase which persisted for approximately 20 min. In contrast, during halothane anesthesia, the initial activation was absent, and the later activity decline and off-response were much reduced. We conclude that biphasic RVMS activity responses emerge to CO2 stimulation, and are state-dependent.

Differential effect of ventrolateral medullary cooling on respiratory muscles of goats.

The objective was to determine whether there is an inhomogeneous response of respiratory muscles during cooling-induced ventrolateral medullary (VLM) neuronal dysfunction in anesthetized and awake goats. Thermodes for cooling were chronically implanted on all or portions of rostral, intermediate, and caudal areas of the VLM of 16 adult goats. Electromyograms (EMGs) were obtained from chronically implanted wires in the diaphragm (di), transversus abdominis (TA), and triangularis sterni (TS) muscles. During some periods of cooling in 9 of 16 anesthetized airway-intubated goats, complete cessation of EMGdi coincided with a reduced yet sustained inspiratory flow. In six awake tracheotomized goats, VLM cooling decreased (P < 0.05) EMGdi duration and minute activity more than inspiratory duration and minute ventilation. Cooling thus decreased activation of the diaphragm more than activation of other respiratory muscles. On the other hand, during VLM cooling in 3 of 10 airway-intact awake goats, cessation of inspiratory flow coincided with sustained EMGdi, suggesting that cooling decreased stimulation of the upper airway muscles more than stimulation of the diaphragm. Finally, VLM cooling in a majority of goats decreased EMGTA and EMGTS more than EMGdi. We conclude that VLM neuronal dysfunction has a differential effect on respiratory muscles of adult anesthetized and awake goats.

Ventilatory responses to cooling the ventrolateral medullary surface of awake and anesthetized goats.

The ventrolateral medulla (VLM) has been reported to be important as a source of tonic facilitation of dorsal respiratory neurons and as a site critical for respiratory rhythmogenesis. We investigated these theories in awake and anesthetized goats (n = 13) by using chronically implanted thermodes to create reversible neuronal dysfunction at superficial VLM sites between the first hypoglossal rootlet and the pontomedullary junction (area M (rostral) and area S). During halothane anesthesia (arterial PCO2 = 57.4 +/- 4.5 Torr), bilateral cooling (thermode temperature = 20 degrees C) of 60-100% of areas M and S for 30 s produced a sustained apnea (46 +/- 4 s) that lasted beyond the period of cooling. While the animals were awake (arterial PCO2 = 36.0 +/- 1.9 Torr), cooling the identical region in the same goats resulted in a decrease (approximately 50%) in pulmonary ventilation, with a brief apnea seen only in one goat. Reductions in both tidal volume and frequency were observed. Qualitatively similar responses were obtained when cooling caudal area M-rostral area S and rostral area M, but the responses were less pronounced. Minimal effects were seen in response to cooling caudal area S. During anesthesia, breathing is critically dependent on superficial VLM neurons, whereas in the awake state these neurons are not essential for the maintenance of respiratory rhythm. Our data are consistent with these superficial VLM neuronal regions providing tonic facilitation to more dorsal respiratory neurons in both the anesthetized and awake states.

Effects on breathing of ventrolateral medullary cooling in awake goats.

Our objective was to investigate the role of the ventrolateral medulla (VLM) in the control of breathing during the awake state. In 17 awake adult goats, chronically implanted thermodes were used to cool the VLM and thereby cause reversible neuronal dysfunction in all or portions of the area between the first hypoglossal rootlet and the ponto-medullary junction (so-called area M (rostral) and area S). Within 5 s after the initiation of cooling, 60-100% of areas M and S, pulmonary ventilation (VE) decreased uniformly over conditions of eucapnia, hypercapnia, hypoxia, and exercise (P < 0.05). Between 10 and 20 s of cooling, the reduction in VE was approximately 10% greater during eucapnia and hypercapnia than during hypoxia and exercise (P < 0.05). For the remaining 10 s of cooling and for about 1 min after cooling, VE increased to and above control level. Cooling only rostral area M or only caudal area M-rostral area S affected breathing qualitatively in the same manner as when 60-100% of areas M and S were cooled. However, cooling caudal area S had effects that differed significantly (P < 0.05) from more rostral cooling in that the initial decrease in VE was attenuated and the subsequent increase was accentuated. The initial uniform decreased VE during cooling suggests that superficial VLM nonchemoreceptor neurons facilitate breathing. The subsequent relatively greater effect of cooling during eucapnia and hypercapnia probably reflects dysfunction of chemoreceptor-related neurons that normally stimulate breathing. The stimulation of breathing during the later stages and after cooling may suggest that some VLM neurons inhibit breathing.