Medullary respiratory neurons in the guinea pig: localization and firing patterns

Brain Research, 591 (1992) 79-87 © 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.P~}

79

BRES 18085

Medullary respiratory neurons in the guinea pig: localization and firing patterns G e o r g e B. R i c h e r s o n a,b a n d P e t e r A. G e t t i n g

a

a Department of Physiology and Biophysics, Unirersity of Iowa, College of Medicine, Iowa City, 1,4 52242 (USA) and h Department of Neurology, Yale Unicersity School of Medicine and VAMC, New Hacoi, CT 06510 (USA) (Accepted 14 April 1992)

Key words: Inspiration; Expiration; Electrophysiology; Nucleus tractus solitarius; Nucleus ambiguus; Dorsal respiratory group; Ventral respiratory group

The location and firing patterns of medullary respiratory neurons have been described in a small number of species. The cat has been the most widely studied species, but some potentially important differences have recently been noted in others. A more complete survey of species is required to determine the significance of these differences. We describe the location and firing patterns of respiratory neurons in the medulla of anesthetized, paralyzed and mechanically ventilated adult guinea pigs. Extracellular single-unit recordings were made from the medulla, their phase relationship with phrenic nerve activity used to define them as respiratory and their location marked with fast green. Respiratory units were concentrated ventrolateral to the nuck,us tractus solitarius (NTS) and within and surrounding the nucleus ambiguus (NA), corresponding to the dorsal respiratory group (DRG) and ventral respiratory group (VRG) of the cat, respectively. Most DRG respiratory units were inspiratory, while the VR(3 contained equal numbers of inspiratory and expiratory units. The DRG and VRG both contained early, late and constant-frequency inspiratory and expiratory units. In general, these findings are similar to those in other mammalian species examined, consistent with these basic aspects of the respiratory network being highly conserved.

INTRODUCTION The brainstem respiratory centers are believed to be similar in their general organization across most mammalian species; however, some important species differences have recently been noted. For example, the cat has a prominent DRG which has been studied extensively t4. In the rat, Saether et al. 2t' found a large number of respiratory neurons in a region ventrolateral to the NTS, corresponding to the DRG of cats; however, several other groups have subsequently found only a small number of DRG respiratory neurons in the adult and neonatal rat ~t'1~'27. The number of other species whose medullary respiratory neurons have been studied is relatively small. If other species are found to lack a DRG, this would suggest that the basic cellular mechanisms of respiration are not conserved across mammalian species. The location and firing patterns of respiratory neurons in the guinea pig have not previously been studied

in vivo, although the guinea pig has proven useful for in vitro preparations for electrophysiology, including the perfused guinea pig preparation 2~ and the in vitro brain slice x''~'l°'~7.The activity of respiratory neurons in vivo has been best studied in the cat, but this species cannot be routinely used for in vitro preparations. We have examined the location and firing patterns of bralnstem respiratory neurons of the guinea pig in vivo to allow comparison with other mammalian species, as well as to provide a foundation for use of in vitro preparations to study the cellular properties of respiratory neurons. M A T E R I A L S AND M E T H O D S Surgical technique Adult guinea pigs (300-1000 g) were anesthetized with 0.2-1.0% penthrane or 1-2% halothane, in 100% oxygen or 50% oxygen/50% nitrous oxide. Animals were paralyzed (gailamine) and mechanically ventilated. Rectal temperature was maintained greater than 36°C by a heating pad. Blood pressure was monitored with a cannula in the carotid artery. Maintenance fluids were provided using lactated

Correspondence: G. Richerson, Department of Neurology, VAMC, Building 34, Room 123, 950 Campbell Ave, West Haven, CT 06516, USA. Fax: (1) (203) 785-5694.

80 Ringer infused through a venous catheter. Animals were placed in a stereotaxic frame and the dorsal surface of the medulla was exposed. Tracheal pressure was monitored with a pressure transducer. The ventilator could be operated in three modes: (I)electronically triggered by the phrenic nerve recording, thus inflating the lungs only during the period of phrenic nerve activity (cycle-triggered ventilation): (2) triggered by a free running oscillator with inspiratory and expiratory durations which were independently controlled (autonomous ventilation); and (3) manually turned off to withhold lung inflation every other breath during cycle-triggered ventilation. All three ventilator modes were used for each recording from medullary respiratory neurons. In three animals, arterial blood was withdrawn from the carotid cannula and PCO 2 was measured with a blood gas analyzer to be 39.8 + 1.67 mmHg (mean + S.D.) at a rate and depth of mechanical ventilation that was typical of the parameters used in other experiments. These parameters were then used in all further experiments to maintain eucapnia.

Data recording Respiratory motor output was recorded from the phrenic nerve with a bipolar cuff electrode. Phrenic nerve activity was recorded on an FM tape recorder and processed with a phrenic nerve integrator (time constant = 100 ms). Single unit activity was recorded from medullary neurons using glass microelectrodes filled with either 150 mM NaCI, or 4 M NaCI (im)'cdance--1-8 MO at I kHz). The extracellular recording was amplified, filtered and recorded on an FM tape recorder. Extracelh:ldr recordings were judged to be action

A

Data analysis The onset of inspiration and expiration was detected by an electronic circuit which measured the rate of change of the integrated phrenic neurogram (IPN). Detection of the change in respiratory phase by the respiratory controller sometimes lagged the true onset of inspiration and expiration by a small amount, never exceeding 100 ms. The lag can be seen as a shift to the right of the bar representing inspiration. This lag did not alter the results of the analysis, because visual inspection of the original recordings was always used to verify the classification of firing patterns. Cycle triggered histogramss (CTHs; 10 ms interval) were used to analyze firing patterns. Neurons were classified u~ing objective criteria for classification based on previously d=scribed patterns of activity from the cat ~4.

B

.z

k

potentials from a single cell (single-unit recording) if there was a homogeneous population of action potential waveforms (entire waveform visualized on an oscilloscope on-line) and the minimum interspike interval was greater than 4 ms. Microelectrodes were inserted into the medulla using a stereotaxic micromanipulator. The location of respiratory neurons was initially determined using stereotaxic coordinates: rostrocaudally relative to the obex (Fig. 1; defined as the junction of the teniae rentrictdi quarti21); laterally relative to the midline; and vertically relative to the dorsal surface of the medulla at the point of entry of the microelectrode. In many cases, the location of the tip of the microelectrode was verified histologically after recording (see below).

"""""""

"'.':""B

2,0 - 2.5 mm

C

D 4th

(

vt NA

1 mm

•

w

,J

0.25-

1.25 mm

Fig. I. Location of respiratory units in the DRG and VRG. Dorsal view of the medulla (Part A) and composite drawings at three levels of the medulla (Parts B-D) of all unit recordings histologically localized from different animals. Coordinates are rostral to the obex. Inspiratory units (circles) are drawn on the right side of the sections and expiratory units (squares) are drawn on the left. Open symbols, VRG units; filled symbols, DRG units (both defined by stereotaxic coordinates). AP, area postrema: cc, central canal; cp, inferior cerebellar peduncle; Cu, cuneate nucleus; Gr, gracile nucleus; IO, inferior olive; mlf, medial longitudinal fasciculus; NA, nucleus arnbiguus; Pyr, pyramids; TS, tractus solitarius; V, spinal trigeminal nucleus; Vlll, vestibular nucleus; X, dorsal motor nucleus of the vagus; XII, hypoglossal nucleus; XII n., fibers of the hypoglossal nerve; 4th, 4th ventricle.

81 Respiratory units were classified into groups based on the time of peak firing (TPF) and the percentage of the inspiratory phase that the cell fired action potentials at a frequency greater than 50% of the peak firing frequency (duration of firing; DF), determined-from the CTHs. Inspiratory units were thus grouped into constant Ire. quency, early and late inspiratory units. Constant frequency inspira. tory units were classified as those units with DF greater than 70% of Ti, regardless of TPF. Early inspiratory units were classified as those units with TPF less than 50% of Ti and DF less than 70% of Ti. Late

inspiratory units were classified as those units with TPF greater than 50% of Ti and DF less than 70% of Ti. The value of 70% for DF was determined empirically after examining plots of DF vs TPF to give the clearest segregation into distinct subgroups. The same classification was used for expiratory units, with Tc substituted for Ti.

Histology Microelectrodes were filled with 150 mM NaCI, 90% saturated with fast green FCF. After recording, negative current was passed

Unit Recording

A

13 1oo IPN

o

!

-4oo 0.5 II

-20O

0

200

40O

6OO

I

a

Tlme (ms)

C

D

i /lm.

0 •2OO

0

•

200 III

0

.

400

-8OO

-8OO

-4OO

-200

Time (ml)

L~00

Time (ms)

E

F

!°

i

I0

0 -2O0O

0

I

o -1000

0

1000

.looo

o

lOOO

200o

3ooo

m i n i

Tlmo (ms) Tlme (ms) Fig. 2. Subtypes of inspiratory units in the DRG. A: coastant frequency inspiratory unit. Top trace, microelectrode recording. Bottom trace, integrated phrenic neurogram (IPN). Dotted lines mark the onset of inspiration and expiration. Scale, 0.5 s. B: constant frequency inspiratory unit. CTH derived from 52 respiratory cycles of the unit in Part A. The beginning of inspiration was used as the trigger event for averaging. The bar in all CTHs represents the inspiratory phase detected by the respirator controller and the time scale is relative to the trigger event. Time of peak firing frequency (TPF)= 80%. Duration of firing greater than 50% of maximum (DF) = 100%. C: early inspiratory unit. CTH derived from 108 respiratory cycles. TPF = 10%; DF = 20%. D: one type of late inspiratory unit. CTH derived from 60 respiratory cycles, The beginning of expiration was used as the trigger event for averaging. TPF = 95%; DF = 10%. E: second type of late inspiratory unit, with an augmenting pattern. CTH derived ~/rom 32 respiratory cycles. TPF = 90%: DF = 60%. F: phase-spanning expiratory-inspiratory unit. CTH derived from 12 respiratory cycles.

82 for 10 minutes (10 p.A). Animals were perfused transpericardially with 10% phosphate buffered formalin. The brainstem was removed and placed in formalinfor at least 12 h. Transverse sections(40 izm) were cut using a freezing microtome (AO Scientific Instruments, Buffalo, NY) and stained with neutral red and metanil yellow. Injection sites were usually50 lzm in diameter. Composite maps of respiratory units within the DRG and VRG were constructed from the injection sites and from neurons recorded ~v.;thin0.7 mm of injection sites using stereotaxic coordinates. Transverse sections were first normalized to a standard size, then the distance and direction from each unit to the center of either the tractus solitarilts or NA was measured. Each unit was then plotted on the composite maps (inspiratory on the right side, expiratory on the left side), the same distance and direction from the center of the tractus solitarius or NA. RESULTS

TABLE !

Number of respirator)' units from the dorsal and central respiratory groups All inspiratory and expiratory units within the stereotactic coordinates used to define the VRG and DRG were included. Pump units and pump inhibited units were not included.

R / C Let'el

(from obex) DRG

Spontaneous respiratory motor output was recorded from the phronic nerve. The inspiratory burst showed a sudden onset, slow augmentation and rapid termination, Post-inspiratory activity 25 was observed in a few experiments. The respiratory rate had a mean of 28 breaths/min, with a mean duration of inspiration (T i) of 0.5 s and a mean duration of expiration (T¢) of 2.2 s. The least squares fit for these data followed the relationship Ti = 0.08. Tc + 0.33 s (n = 16 animals). The effect of lung stretch receptor feedback on T~ and T~ was studied by withholding lung inflation during cycle-triggered ventilation. The average T~ for each individual animal was then calculated for the cycles with lung inflation and without lung inflation and T¢ was calculated from the expiratory phase following the corresponding Tj, The averages from each animal were then used to calculate the average across animals. Withholding lung inflation resulted in an average increase in TI ~f 294 ms (59% of the mean) and To increased by 457 ms (18% of the mean; n - 9 animals). Distribution o f respiratory units

Most respiratory neurons were located within and adjacent to the NA and NTS (Fig. 1) corresponding to the VRG and DRG of the cat t4. Attempts to record extracellular activity from respiratory neurons were restricted to a region from 0.5 mm caudal to 3,0 mm rostrai to the obex, from 0,0-2.5 m1:1 lateral to the midline and to 4,0 mm below the dorsal surface of the medulla. This region included the NTS and the NA and extends from the caudal 1/3 of the medulla to the pontomedullary junction. DRG respiratory units were found in a relatively compact region ventral to the tractus solitarius, Pump units were not histologically localized, but they tended to lie immediately dorsal to respiratory units of the DRG, VRG respiratory units were located within or adjacent to the NA,

Expiratory n (%)

Total n

0.0-0.7 mm 0.7-1.3 mm 1.3-2.5 mm

14 (87) 32 (78) 20 (87)

2 (13) 9 (22) 3 (13)

16 41 23

Total VRG

66 (82)

14 (18)

80

mm 1.4-2.5 mm Total

16 (59) I I (46)

11 (41) 13 (54)

27 24

27 (53)

24 (47)

51

0.0- !.4

Respiratory motor output

lnspiratoty n (%)

Dorsal respiratory group

Within the DRG (0.0-2.5 mm rostral to the obex, 0.7-1.5 mm lateral to the midline and 0.6-2 mm from the dorsal surface of the medulla), 66 out of 80 (82%) respiratory neurons were inspiratory (Table I). Inspiratory units in the DRG were classified based on their temporal pattern of firing. For example, some units fired throughout the inspiratory phase with a constant frequency (Fig. 2A, B), while other units had their peak firing frequency during late inspiration or early inspiration (Fig. 2C). Late inspiratory units could be further grouped into those which fired only during late inspiration and those which fired throughout inspiration, but with a ramp firing frequency (augmenting unit) (Fig. 2D,E). The onset of firing of some constant-frequency units (e.g. Fig. 2A) preceded the onset of inspiration by 100-200 ms. When the onset of activity was this close to the onset of inspiration, these units were still considered inspiratory units rather than phase-spanning units. Three phase-spanning expiratory-inspiratory units ~7 were recorded in the DRG of the guinea pig (Fig. 2F). This class of unit was silent for approximately the first third of expiration, The firing rate slowly increased during the last two-thirds of expiration. During inspiration the firing rate rapidly increased, reaching a maximum during the last half of inspiration. Expiratory units in the DRG with early, late and constant frequency expiratory patterns were also found (Fig. 3). Using plots of DF vs TPF, the classification of firing patterns resulted in segregation of units with an early inspiratory pattern, but there was a continuum of cell types between constant-frequency inspiratory units and late inspiratory units. Expiratory units of the D R G fell more clearly into three distinct groups. In the DRG, respiratory units that were recorded long enough

83 to be classified i n t o subgroups based on CTHs included 29 inspiratory units (14 constant frequency, six early and nine late); seven expiratory units (two constant frequency, three early and two late) and three phase-spanning expiratory-inspiratory units.

Unit Recording

Effect of lung inflation on dorsal respiratory group units

II,!ll

The effect of lung inflation on neuronal firing pattern was studied by withholding lung inflation during cycle-triggered inflations. Lung inflation during inspiration resulted in two different responses in D R G inspiratory neurons. One type of unit showed a decrease in DF and TPF when the lungs were inflated, without a change in the rate of rise of activity, analogous to the response of the whole phrenic nerve recording with lung inflation (Hering-Breuer inflation reflex). This response has been described by Cohen 5 as an inflation (0) unit. Seven out of 11 units tested were inflation (0) units and included early, late and constant-frequency inspiratory patterns. The second type of response to lung inflation during inspiration was an increase in the rate of rise of firing frequency and an increase in the maximum firing fre-

40

,nn

¢/)

-1000

o

-500

5o0

Tlmo (ms)

B

r~

n a

c,) 0 .lOOO

o

lOOO Time (ms)

° 40 ca. (O

I;t:

fl

II ,h ll),ill151J

i i

,1: li,

Tracheal Pressure

10 mm Hg I

Fig. 4. Pump-inhibitedunit. Animalwas ventilated usingautonomous ventilation. Bars are drawn above the extracellular recording during the central inspiratory phase and below the extracellular recording during lung inflation. Scale, 2 s.

quency. Neurons with this type of response are called inflation ( + ) units 5. Four of 11 units in the DRG responded to lung inflation in this manner. All inflation ( + ) units were constant-frequency inspiratory units. Units were also found in the DRG that fired during lung inflation independent of the phrenic nerve discharge. Units with this firing pattern are called pump units "~and are not considered a type of inspiratory unit. Within the DRG, 16 units with this firing pattern were found. Units with an unusual pattern, which we have called pump.inhibited units 22, were also recorded from the DRG. Pump-inhibited units fired with an expiratory pattern during cycle-triggered inflation; however, during autonomous ventilation pump.inhibited units were completely silent during the inspiratory phase of phrenic nerve activity and also during the time of lung inflation (Fig. 4). Five units of this type were found in the DRG.

Ventral respiratory group

0

-500

0

5O0 Time (ms)

Fig. 3. Subtypes of expiratory units in the DRG. A: constant frequency expiratory unit. CTH derived ttom 28 respiratory cycles. TPF = 30%; DF -- 100%. B: early expiratory unit. CTH derived from 41 respiratory cycles. TPF = 10%; D F - 10%. C: late expiratory augmenting unit. CTH derived from 94 respiratory cycles. TPF = 85%; DF = 30%.

Within the VRG (1.0-2.5 mm lateral to the midline and 2.0-4.0 mm from the dorsal surface), recordings were only attempted from 0.0-2.5 mm rostral to the obex, corresponding approximately to the intermediate VRG of cats. Table I lists the numbers of inspiratory and expiratory units found within these coordinates in 19 animals. Overall, the intermediate VRG contained approximately equal numbers of inspiratory and expiratory units. There was a small difference between the

84

B

A

¢T

i

~' .==

w

so

o.

m

0 -500

0

500

10C10 Time (ms)

2s

D

C A

,o

40

g e~ L/]

20

0 _=

-4OO

-200

0

200

,,, 400

g

20

~

0

n

-500

-.lo~

6OO

0 Time (ms)

Time (ms)

Fig. 5. Subtypes of inspiratory units in the VRG. A: constant frequency inspiratory unit. Bars are drawn above the IPN during the first two inspiratory phases. Scale, 2 s, B: CTH derived from 34 respiratory cycles of the unit in Part A. TPF = 80%; DF = 90%. C: early inspiratory unit, CTH derived from 67 respiratory cycles, TPF = 20%; DF = 40%, D: late inspiratory unit. CTH derived from 26 respiratory cycles. TPF = 75%; DF = 33%.

A

B

I

J .1000

o

0 .1000

looo

0

m

1000

I

Time (ms)

C

Time (me)

D 15

I '= so

.o ,~0

o

~oo

III

.1000

o

1~00

2OOO

m~=~ll~B Timo (ms)

Time

(ms)

Fig. 6, Subtypes of expiratory units in the VRG, A: constant frequency expiratory unit, CTH derived from 14 respiratory cycles. TPF--95%; DF = 100%. B: decrementing early expiratory unit, CTH derived from 29 respiratory cycles. TPF = 5%, DF = 12%. C: post-inspiratory unit. CTH derived from 30 respiratory cycles. TPF = 5%; DF = 20%. D: late expiratory unit. CTH derived from 37 respiratory cycles. TPF = 80%; DF = 50%,

85 number of inspiratory and expiratory units at different rostro-caudal levels, but the number of recordings were not sufficient to clearly determine if there was a rostrocaudal segregation of inspiratory and expiratory units. In general, the same types of inspiratory and expiratory units were found in the VRG as in the DRG, including early, late and constant frequency inspiratory units (Fig. 5) and early, late and constant frequency expiratory units (Fig. 6). One type of early expiratory unit in the VRG, called a post-inspiratory unit 24, fired a short burst of action potentials at very high frequency at the beginning of expiration and then the firing frequency decreased to near zero within 400 ms after the onset of expiration (Fig. 6C). O~ly one unit with a post-inspiratory firing pattern was found and was located in the rostral portion of the regior, examined (2.0 mm rostral to the obex). Another type of early expiratory unit in the VRG, called a decrementing early expiratory unit (Fig. 6B), reached a maximum firing frequency at the beginning of expiration, but continued to fire action potentials throughout expiration. Despite the continued activity throughout expiration, the maximum firing frequency was much less than that for the post-inspiratory unit. Classification of VRG firing patterns resulted in a clear grouping of units with early inspiratory and early expiratory patterns. The other firing patterns were not as clearly grouped. In particular, the late expiratory units had only a modest degree of augmentation, thus resembling the constant-frequency expiratory units. In the VRG, respiratory units that could be classified based on CTHs included 14 inspiratory units (three constant frequency, seven early and four late) and 13 expiratory units (seven constant frequency, three early decrementing, one post-inspiratory and two late). DISCUSSION The goal of this study was to characterize the phrenic nerve output, localize the regions with high concentration of respiratory neurons and define the patterns of respiratory unit firing in the medulla of the guinea pig. A limited number of studies of medullary respiratory neurons have been done on mammalian species other than the cat. Defining the similarities and differences between species may help to understand the function and importance of different components of the respiratory system. These results are also important as a foundation for in vitro studies of the guinea pig respiratory network. Phrenic nerve output

Phrenic nerve activity of the guinea pig had the same stereotypical waveform and temporal pattern

characteristic of other species TM. In some cases, there was post-inspiratory activity present, as previously described in the cat 25 and rat 27. In the cat there is a larger slope of the Ti vs. Te relationship 4, although differences in this relationship have been noted previously within the same species and probably depend on the experimental conditions (e.g., anesthetic level3). Pulmonary stretch receptor feedback caused a large decrease in Ti, with a less consistent and proportionately smaller effect on Te. On the average, the effect of pulmonary stretch receptor feedback was to shift the Ti vs Te curve towards lower values of Ti, consistent with the Hering-Breuer reflex ~4. Presence o f a dorsal and ventral respiratory group

In the guinea pig, we found the majority of respiratory units concentrated in two regions. The DRG was located immediately ventral to the tractus solitarius, which is slightly different than the, cat ~4 and rat 26 where the DRG is ventrolateral to the tractus solitarius. The portion of the VRG examined (corresponding to the intermediate region of the VRG ~a) was located within and surrounding the NA. There was also a band of respiratory units in the reticular formation between the DRG and VRG. The location of medullary respiratory neurons and their firing patterns have been extensively studied in the cat TM. Limited studies have also been done in the rabbit 12, rat 11,16,26,27, perinatal sheep I, piglet 19, opossum t3, dog and monkey 2. The DRG and VRG of the cat are both well developed and are thought to play a role in development or modulation of normal motor output. Respiratory neurons have also been found in large numbers in the region of the NA and NTS of the rat 26 and rabbit ~2. Smaller studies found respiratory neurons in the same regions of the opossum I"~ and sheep I. Recently, several investigators have reported a relative paucity or frank absence of respiratory neurons in the DRG of the rat t1'27. This and evidence from lesion studies ~6 has led to questions regarding the role of the DRG and the validity of the cat as a model for mammalian respiration. The guinea pig was found to have a very prominent DRG; however, we found two sources of selection bias when attempting to localize the DRG. First, respiratory units in the VRG were considerably easier to isolate as single units than in the DRG. Frequently, as an electrode passed through the region suspected to be the DRG (based on coordinates), a large amount of background inspiratory activity was seen that was not pump related, without single units being recorded. The ability to record single-unit activity was highly dependent on the configuration of the electrode. This prob-

86

lem was not encountered while recording from the VRG. One explanation for this difficulty could be that the neurons of the guinea pig DRG are smaller (range of diameters 10-40 pm (ref. 10) than those of the VRG (range of diameters = 10-82 pm (ref. 17)), making them more difficult to record from. A second reason that respiratory units were more difficult to locate in the DRG was that the DRG is a more compact cell group than the VRG in the guinea pig (Fig. 1). Respiratory units were not recorded from the DRG if the electrode tract was displaced by only 200 ~m from the center of the DRG. Similar factors may be responsible for the disparate results obtained in previous attempts to localize a DRG in the rat. The presence of a DRG in the guinea pig, rabbit and cat (and possibly the opossum, sheep and rat) suggests that the DRG is a general feature of the mammalian respiratory network.

Firing patterns in the dorsal respiratory group lnspiratory, expiratory and pump units were found in the DRG of the guinea pig. Inspiratory and expiratory units each fell into several different categories which closely parallel those seen in the cat j4, rabbit 12 and rat 2~. Constant frequency inspiratory and late inspiratory units were both common in the guinea pig and appeared to blend into each other based on the classification scheme used here. Early inspiratory units were very common, making up 19% of the inspiratory units in the DRG, as opposed to the cat where they are uncommon ¢'. Expiratory units made up 18% of the non.pump respiratory units in the DRG, which is higher than has been reported in the cat (4-5%), but similar to that reported in the dog (20%) s and in the rat (13%) ~. Phase-spanning expiratory-inspiratory units were seen in the guinea pig DRG, which have been reported previously in the rat VRG 27. Based on this and other evidence, respiration has been separated into three phases: inspiration; post-inspiration or stage I expira. tion; and stage !1 expiration associated with active contraction of some expiratory muscles 2s. There was previous evidence for three phases of respiration in the rat, cat and piglet 27. The presence in the guinea pig of phase-spanning expiratory-inspiratory neurons, post-inspiratory neurons and post-inspiratory activity in the IPN add an additional species to this list and supports the hypothesis that the presence of three phases of respiration may be a general property of the mammalian respiratory network. Four of 11 neurons tested in the DRG of the guinea pig were inflation ( + ). This is similar to the cat, where 40% of the inspiratory units of the cat DRG are

inflation (+)5. All the units examined in the VRG of the guinea pig responded to lung inflation with an inflation (0) response. In the cat, most inspiratory units in the VRG are inflation (0), although some are inflation (.)5. A cell type was found in the guinea pig DRG which has not been reported in other species, called a pumpinhibited unit. During cycle-triggered ventilations, pump-inhibited units were indistinguishable from constant-fl'equency expiratory units. During autonomous ventilation, pump-inhibited units stopped firing completely during the expiratory phase whenever the lungs were inflated. Expiratory neurons showing a mild degree of inhibition by lung inflation during the expiratory phase have previously been reported in the caudal ~ and rostra115 nucleus retroambigualis. During hypocapnic apnea, the tonic activity of some expiratory neurons from the Botzinger complex are also inhibited by lung inflation 2°. However, pump inhibition of DRG neurons has not previously been reported and the complete inhibition of firing seen here has not previously been reported from any eucapnic medullary expiratory neurons. Pump units are thought to be second-order sensory neurons which receive afferent input from pulmonary stretch receptors ~4 and may be involved in termination of inspiration during the Hering-Breuer inflation reflex. The pump-inhibited unit may receive inhibitory input from pump units and the decreased activity of these neurons during expiration may play a role in delaying the ons~'*, of inspiration during the Hering-Breuer expiratory prolongation reflex.

Firing patterns in the ventral respiratory group In the guinea pig, inspiratory and expiratory units were found in approximately equal numbers in the region of the VRG examined. Respiratory units in the VRG had early, late and constant-frequency inspiratory firing patterns and decrementing early, late and constant frequency expiratory patterns. Each of these types of firing patterns have been reported in the VRG of the cat 14 and rat z6.zT. One unit with a post-inspiratory firing pattern was found in the VRG. Neurons with this type of firing pattern have been described in the cat 2"~,rat 27 and opossum Is. CONCLUSIONS We found that the major features of the guinea pig respiratory network are similar to the cat and other species, with some minor features of the guinea pig which appear to be unique. These results suggest that the general organization and the cellular mechanisms

87 of respiratory neurons are highly conserved across mammalian species. Acknowledgements. This work was supported by NIH Grants NS15350 (P.A.G.), GM07337 & MHlSI?2 (GBR) and constituted a portion of a Ph.D. thesis (G.B.R.). G.B.R. is a Research Associate in the VAMC Career Development Program. We wish to thank Stephen Johnson, Gabriel Haddad and Jeffery Kocsis for critical review of this manuscript.

REFERENCES 1 Bystrzycka, E., Nail, B.S. and Purves, M.J., Central and peripheral neural respiratory activity in the mature sheep foetus and newborn lamb, Respir. PhysioL, 25 (1975) 199-215. 2 Carregal, E.J.A., Williams, B. and Birzis, L., Respiratory centers in the dog ard squirrel monkey: a comparative study, Respir. Physiol., 3 (1967) 333-348. 3 Clark, F.J. and Von Euler, C., On the regulation of depth and rate of breathing, J. Physiol., 222 (1972) 267-295. 4 Cohen, M.I., Phrenic and recurrent laryngeal discharge patterns and the Hering-Breuer reflex, Am. J. Physiol., 228 (1975) 14891496. 5 Cohen, M.I., Neurogenesis of respiratory rhythm in the mammal, Physiol. Rev., 59 (1979) 1105-1173. 6 Cohen, M.I. and Feldman, J.L., Discharge properties of dorsal medullary inspiratory neurons: relation to pulmonary afferent and phrenic efferent discharge, J. Neurophysiol., 51 (1984) 753776. 7 Cohen, M.i., Feldman, J.L. and Sommer, D., Caudal medullary expiratory neurone and internal intercostal nerve discharges in the cat: effects of lung inflation, J. Physiol., 368 (1985) 147-178. 8 Dekin, M.S., Rieherson, G.B. and Getting, P.A., Thyrotropin-releasing hormone induces rhythmic bursting in neurons of the m~cleus tractus solitarius, Science, 229 (1985) 67-69. 9 Dekin, M.S, and Getting, P.A,, in vitro characterization of neurons in the ventral part of the nucleus tractus solitarius, ll. Ionic basis for repetitive firing patterns, J. NeurophysioL, 58 (1987) 215-229. I(I Dekin, M,S,, Getting, P,A. and Johnson, S.M., in vitro characteriz~ltion of neurons in the ventral part of the m¢ch,us tractt¢.~' solitariu,v, !, identification of neutomd types and repetitive firing properties, J. Netm#d~viol., 58 (1987) 195-214.

i I Ezure, K., Manabe, M. and Yamada, H., Distribution of medullary respiratory neurons in the rat, Bra#a Res., 455 (1988) 262-270. 12 Fallert, M. and Wassermeyer, B., Spatial distribution of various types of bulbar respiratory neurons in the rabbit, Exp. Brain Res., 30 (1977) 339-351. 13 Farber, J.P., Medullary inspiratory activity during opossum development, Am. J. Physiol., 254 (1988) R578-R584. 14 Feldman, J.L., Neurophysiology of breathing in mammals. In: F.E. Bloom (Ed.), Handbook Of Physiology: Section: The nerrous system IV," Volume: Intrinsic regulatory systems of the brain, APS 1986, pp. 463-524. 15 Feldman, J.L. and Cohen, M.I., Relation between expiratory duration and rostral medullary expiratory neuronal discharge, Brain Res., 141 (1978) 172-178. 16 Hilaire, G., Monteau, R., Gauthier, P., Rega, P. and Morin, D., Functional significance of the dorsal respiratory group in adult and newborn rats: in vivo and in vitro studies, Neuroscience Lett., 111 (1990) 133-138. 17 Johnson, S.M. and Getting, P.A., Electrophysiological properties of neurons within the nucleus ambiguus of adult guinea pigs, J. Neurophysiol., 66 ( 1991) 744- 761. 18 Kalia, M., Neurohistochemical methods in tracing central respiratory mechanisms, Fed. Proc., 40 (1981) 2365-2371. 19 Lawson, E.E., Richter, D.W. and Bischoff, A., lntracellular recordings of respiratory neurons in the lateral medulla of piglets, J. Appl. Physiol., 66 (1989) 983-988. 20 Pantaleo, T. and Corda, M., Expiration-related neurons in the region of the retrofacial nucleus: vagal and laryngeal inhibitory influences, Brain Res., 359 (1985) 343-346. 21 Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Comdinares, 2nd edn., Academic Press, New York, 1986. 22 Richerson, G.B. and Getting, P.A., Characteristics of respiratory neurons in the guinea pig, NeuroscL Abstr., 9 (1983) 1163. 23 Richerson, G.B. and Getting, P.A., Preservation of integrative function in a perfused guinea pig brain, Brain Res., 517 (1990) 7-18. 24 Richter, D.W. and Ballantyne, D., Central rhythm generation by i,,spiratory neurones in the anaesthetized cat, Neurosci. Lett., 7 (1981) S117. 25 Richter, D.W., Generation and maintenance of the respiratory rhythm, J. Exp. BioL, 100 (1982)93-107. 26 Saether, K., Hilaire, G. and Monteau, R., Dorsal and ventral respiratory groups of neurons in the medulla of the rat, Brain R~:~., 41.o (1987) 87-96. 27 Schwarzacher, S.W.,Wilhelm, Z., Anders, K. and Richter, D.W., The medullary respiratory network in the rat, J. PhysioL, 43:~ (1991) 631-644.