Journal of the Autonomic Nervous System, 41 (1992) 129-140
© 1992 Elsevier Science Publishers B.V. All rights reserved 0165-1838/92/$05.00 JANS 01356
Role of the rostral ventrolateral medulla in the generation of synchronized sympathetic rhythmicities in the rat Andrzej Trzebski and Stanislaw Baradziej Department of Physiology, Medical Academy, Warsaw, Poland
Key words: Rostral ventrolateral medulla; Sympathetic oscillator; Power density spectral analysis; Synaptic block; Calcium antagonists Abstract In artificially ventilated, paralyzed rats anesthetized with Nembutal or urethane, power density spectral analysis (PDS), using direct FFT algorithm, was used to quantify rhythmicities in the sympathetic cervical and lumbar nerves after bilateral microinjections into rostral ventrolateral medulla (RVLM) of CoCI 2 and MgCI2--unspecific synaptic transmission blockers. Later overall sympathetic activity, phrenic nerve discharge, heart rate and arterial blood pressure were recorded. Block of synaptic transmission in RVLM was tested by elimination of sympathoinhibitory or sympathoexcitatory reflex responses to aortic nerve and vagal afferents stimulation respectively. In animals vagotomized with bilateral section of carotid sinus nerve the power in all frequency bands was very significantly reduced to a level not different from that which remained after spinal cord transsection. If carotid baroreceptors were intact, a small peak corresponding to cardiac frequency band persisted. Overall, non-synchronized sympathetic activity and arterial blood pressure increased. All effects were transient, lasted up to 15 min, and could be reproduced several times in one experiment. Respiratory rhythmic activity was unchanged yet respiratory-sympathetic synchronization was abolished. It is concluded that RVLM reticulospinal sympathoexcitatory neurons are responsible for non-synchronized tonic sympathetic activity but are not able to generate synchronized sympathetic rhythms. Synaptic input, presumably inhibitory, either from local neuronal circuits within ventral medulla a n d / o r from other brain stem neuronal circuities is needed to shape out the flexible pattern of sympathetic oscillations.
A simplified view on the classic paradigm of homeostasis, a cornerstone of the past and present physiology and biology laid by Walter B. Cannon , has been challenged recently by the theories of chaotic dynamics [20,36], synergetics  and concept of dissipative structures in irre-
Correspondence to: A. Trzebski, Department of Physiology, Medical Academy, Krakowskie Przedmiescie 26/28, 00-325 Warsaw, Poland.
versible thermodynamic processes being applied to biology . Instability and fluctuations are no more disregarded as a stochastic noise, uncompensated error or imperfection of closed loop cybernetical models of regulation  but emerge as fundamental characteristics and preconditions of living systems. Chandler McC. Brooks, a pupil of Cannon, was aware that realities of the totality of interacting functional systems of the body are beyond the reach of classical tools of system analysis. In 1984 he asked: "The rhythms that a computer could identify must be legion--the question is and will be: what rhythms are func-
tionally significant?" . This is the question which physiology of the autonomic nervous system has been confronted with before and still is. Cardiac and respiratory-related oscillations were discovered just as the first direct recording of the sympathetic nerve activity by Adrian, Bronk and Phillips was made, sixty years ago . As periodicities of sympathetic discharge  are not abolished by elimination of rhythmical peripheral inputs from pulmonary, baro- or chemoreceptors, the hypothesis that a central oscillator accounts for sympathetic rhythm was proposed [5,17]. More recently, a model of a central probabilistic multisynaptic oscillator entrained to rhythmical peripheral inputs was presented . Such a central oscillator would reflect a flexible superimposition of the rhythms arising from multiple circuits of brain stem neurons of changing anatomical configuration and characteristics related to the functional state of the body [19,27]. A similar, yet more general, idea of a common reticular network comprising multiple coupled oscillators responsible for generation and sliding coordination of rhythmicities in the autonomic nervous system and in respiratory and motor outputs has been proposed before by Koepchen [28,29] and his group . In the search for specific brain areas critical for generation and synchronization of sympathetic rhythmicities, the rostral ventrolateral medulla (RVLM) deserves particular attention. RVLM is not merely the main supraspinal population of reticulospinal excitatory neurons, a final common pathway for most sympathetic reflexes and central inputs  but represents an ensemble of synaptically organized subsets of neurons. Some of them project to functionally different spinal preganglionic neurons . Some RVLM neurons are synaptically interconnected with dorsal ventrolateral , rostral ventromedial medulla  and with other neuronal circuits involved in sympathetic control [23,27]. Our previous studies  demonstrated that a restricted neuronal pool in RVLM (paraambigual area) represents an interphase between respiratory oscillator and reticulospinal sympathoexcitatory neurons critical for generation of respiratory-related sympathetic oscillatory activity and for the timing of sympathetic
rhythms within each phase of individual respiratory cycle. This conclusion was confirmed by more refined pharmacological techniques . Some RVLM neurons studied both in vivo  and in vitro  exhibit the properties of pacemaker cells generating intrinsic rhythmicity in the absence of synaptic input. These findings opened a new avenue of speculations on the sympathetic central oscillator, very similar to the problem which respiratory rhythmogenesis is confronted with at present : what is the role of single cell intrinsic oscillators versus more complex neuronal synaptic circuitry in the generation of sympathetic rhythms and patterns? (for discussion, see ). The present experiments were conducted to study the effect of blocking synaptic transmission in RVLM by calcium antagonists, divalent Mg 2+ and Co 2+ cations , on rhythmicities exhibited in sympathetic discharge. The purpose was to elucidate if pacemaker intrinsic activity within RVLM reticulospinal cells [39,40] could account for rhythmical synchronization of activity in sympathetic nerves by itself. Some preliminary results were presented at international symposia [3,4,42].
Materials and Methods General Twenty male Wistar (280-360 g) rats were anaesthetized intravenously (i.v.) with Nembutal or Sagatal. In another group of experiments, five male Wistar rats (280-360 g) were anaesthetized with an intraperitoneal (i.p.) injection of urethane Fuka AG 1.2 g / k g in order to compare the effects of different anesthesia on the results. As no significant difference between the two groups was found, pooled data were used for statistical analysis. Details of methods for maintaining body temperature at 37°C, artificial pump ventilation with oxygen-enriched room air, administering muscle relaxant (Pavulon Organon Hesse, i.v.) and recording arterial blood pressure have been previously described . Heart rate was monitored by R waves of ECG. Monitoring and maintaining blood gases were performed by AVL 995Hb Automatic Blood Gas System in 0.1 ml blood sample taken from the femoral artery, and re-
placed by the same amount of donor blood or plasma. In addition, the end tidal CO 2 concentration was monitored by Beckman LB-2 Gas Analyser and maintained at the level of 4% by appropriate change of ventilatory pump frequency. The rat was placed in a stereotaxic frame in a supine position. In all experiments animals were vagotomized, and carotid sinus nerves cut bilaterally. In five animals the effects were observed' before and after carotid sinus denervation. The central ends of aortic nerve and vagus nerve were dissected and placed on bipolar platinum electrodes in a paraffin pool for stimulation. In some experiments the spinal cord was exposed from C 1 to C a prior to experiment. Spinal transsection was performed very slowly to minimize damage to the spinal cord.
Nerve recordings and stimulation Surgical procedure and details of sympathetic nerve dissection are published elsewhere . Efferent activity was recorded in the cervical nerve and in the lumbar sympathetic trunk or its branches. Phrenic nerve activity was recorded in parallel with sympathetic nerve activity. Nerve activity was amplified, integrated (time constant 0.2 s) fed into a window discriminator which passed discharge above electrical noise levels. The number of spikes was counted every 10 s for 16 periods and computed as a mean in 1 s according to M R A T E program (Cambridge Electronic Design Ltd.). Nerve activity and blood pressure were recorded on the Racall 7DS tape recorder and simultaneously displayed on paper by Mingograph 7 ink injector recorder (Siemens) and on 5103N Tektronix oscilloscope. The central end of the separated aortic nerve and the vagal trunk was stimulated by Digitimer 4030 programmed stimulator with the 3-10 V pulses of 0.5-1.0 ms duration and 5-30 Hz frequency. Lack of sympathoinhibitory and depressor responses to baroreceptor aortic nerve stimulation and lack of sympathoexcitatory and pressor responses to high frequency vagal afferent stimulation were accepted as criterion of synaptic transmission block in RVLM neurons as both sympathoinhibitory and sympathoexcitatory reflexes are synaptically mediated by R V L M neurons .
Microinjections into RVLM The ventral surface of the basooccipital skull was exposed by removing muscles and soft tissues. A hole 6 mm wide and 8 mm long was drilled in the bone to create a window large enough to expose the ventral medulla. Under binoculars, the dura and arachnoid was opened and caudal range of trapezoid bodies, basiliar artery and its branches, pyramids and nerve rootlets visualized. The window cavity was filled with cerebrospinal fluid. In most experiments bilateral, symmetrical microinjections were performed. Small bore pipettes of 4 0 - 1 0 0 / z m internal tip diameter were applied. PV 830 Pneumatic PicoPump WPI microinjector connected to a 100 atmospheric pressure reservoir was used. The volume programmed by microinjector was, in addition, measured directly by micrometer observation of the changing meniscus level within the pipette of known internal diameter. For functional mapping of RVLM, 10-20 nl of 0.5 M glutamate sodium was microinjected according to the procedure described in detail previously . Pressor and sympathoexcitatory response were used as criteria for the proper location of the micropipette tip. The same sites were subsequently used for microinjections of other agents. For histological verification, the brain was slowly perfused through the left ventricle with 10% formalin solution under 20 mmHg pressure after the end of the experiment. The brains were then removed and medulla-sectioned on Kryostat Microtom into 60 Izm transverse sections. Pontamine sky blue dye added to microinjected solutions marked injection sites located by using the atlas of Paxinos and Watson . The following solutions in saline were used: 4 mM and 12 mM MgC12 or CoC12 respectively, and 0.01 and 0.5 M sodium kainate. All solutions were buffered in the range 7.2-7.5 pH at 37°C. The volumes varied between 150-300 nl and were injected very slowly up to 180 s depending on the volume. Saline solution of the same volume and speed of injection was used for control. Data collection and analysis Recorded nerve activity was rectified with time constant 0-0.2 s in 1 s intervals. Integrated sig-
nals were digitized and averaged with sampling frequency 1000-2000 Hz by Cambridge Electronic Design Ltd. 1401 real-time computer triggered by the constant level of increasing integrated phrenic nerve burst. Electrical noise level was determined by recording from the nerves after the animal had been killed with an overdose of Nembutal. In frequency domain for power density spectral analysis (PDS) a direct Fast Fourier transform (FFT) algorithm was applied,
and Waterfall program 1989 (Cambridge Electronic Design Ltd.) and IBM A T computer for data processing were used. Briefly, nerve signals were picked up off-line from a Racall magnetic tape recorder at a frequency range u p to 3000 Hz. Integrated activity was filtered and sampled at 100 Hz. A slow constant component was eliminated by long-pass filtering. The signal was analyzed in the filter passband up to 100 Hz. The spectra obtained for the different data sets were
I ~..--~ 4-, ~ I
~='-~ g=_'7-~---=--'-.=. . .
W=~r-e ......e ~
.. . . . . . .
" r ~ > - - - - = .. . . . . . . . .
--"- - -
12. Ha Fig. 1. Power spectra of the cervical symapthetic nerve activity (upper records) and lumbar nerve activity/trunk (below) b~fore and aRer unilateral microinjection of 150 nl 12 mM CoCl a solution m a r ~ d by solid dots. $ ~ c t r a were con'q~l~d during lff-s periods of recording. Solid points (upper record) mark 10-s intervals between sequential spectra. In the bottom figure, lO-s intervals are marked on the vertical axis. 8. B
after ascertaining equality of variance between the groups. The value of probability used to determine significance in statistical analysis was 0.05. Variability was expressed as means + S.E.
averaged so as to attenuate variable noise contribution and then computed in sequential windows of 10 s each. It was assumed that the limited duration of 10 s enabled us to compute nonstationary processes (transients) in the same way as the power spectra of stationary processes in frequency domain . Spectral values were exhibited as relative amplitudes, the value in each frequency bin representing the fraction of the total energy that was present in the band around that frequency. The bands around individual frequencies appeared to be broad. Such a reduced strength of rhythms characterizes mass activity in which superimposition of activities in individual fibers occurs, and the resulting aggregate rhythm is less regular. In order to minimize this effect, for quantification the areas within selected frequency bands instead of peak amplitude were measured. Respective values for each nerve spectra were compared by a Student's paired t-test
PDS in vagotomized and chemoreceptor denervated animals exhibited power in the broad frequency range between 0.4-7.5 Hz. Three bands were distinguished and computed: 0.7-2 Hz which corresponded to respiratory frequency and its harmonics, 5.5-7.5 Hz in the range of cardiac frequency, and a broad band between 2.5-5 Hz with irregular peaks not related to either respiratory or cardiac frequency. In this band maximal fraction of power was exhibited. In the lower frequency range of this intermediate band occasional peaks, corresponding to harmonics of res-
o.o ,3 / f
8. 8 1 0
,~ V I
Fig. 2. Power spectra of activity in lumbar sympathetic trunk in the rat with carotid baroreceptors intact. Solid circles indicate respiratory frequency, open cirles mark cardiac frequency. Abscissa: frequency in Hz; ordinate: fraction of power. (A) control; (B) after ipsilateral microinjection of 150 nl 4 mM MgCI 2 solution into RVLM; (C) the same microinjection into contralateral RVLM; (D) spectrum computed 20 min later.
134 piratory rhythm, appeared. Relative power related to cardiac rhythm was small compared to power represented by other frequencies. In order to check the relation between respiratory-related/cardiac-related bands in PDS averaged power spectra of the frequency of integrated phrenic activity and frequency of R-R intervals of ECG signal were correlated with respective frequency bands in the PDS of the sympathetic nerve. The general pattern of PDS in cervical and lumbar sympathetic nerves was in agreement with our earlier findings . Multiple peaks in the respiratory rhythm-related band are presumably accounted for by distinct phases of sympathetic activity which occur rhythmically in each individual respiratory cycle in the rat [11,12]. In PDS they are represented by harmonics of respiratory rhythm.
Effects of MgCl 2 and CoCl 2 applied into RVLM on the PDS of sympathetic rhythmicities MgCl 2 and CoCl 2 produced a similar effect in both sympathetic nerves. Unilateral microinjection was followed by significant reduction of relative power in all frequency bands (Fig. 1). Bilateral microinjection reduced the total and frequency-related power in PDS (Fig. 2) to a level not much different from that which remained after spinal cord section (Fig. 3). Effects appeared within 5 - 1 0 s after the onset of microinjections, developed and lasted for 160 s up to 3 min depending on the volume and concentration of MgCI 2 and CoCl 2 solution. All effects were reversible. R e a p p e a r a n c e of rhythmicities was manifested in the spectrum occasionally by more pronounced power and additional speaks (Figs. 1, 2). Reproducible transient effects could be repeated after several microinjections into the same site in 15-20 min intervals. In animals in which the carotid sinus nerve remained intact, a small peak in PDS in the cardiac frequency band remained, even after bilateral microinjection of bivalent calcium ion antagonists (Fig. 2). Figures 4 and 5 present summarized data. Blood pressure and overall sympathetic response A slight yet jsignificant increase in the blood pressure followed CoCI 2 and MgCI 2 application
A |.ta 0.1 Lm IL tO
L Im LIO O, m l
C L IM
I. I m
. . . .
Fig. 3. On the left and right parallel power spectra of activity recorded in two single branches dissected from the lumbar trunk. (A) control; (B) 1 rain after bilateral microinjectionsof 300 nl 4 mM MgCIz into RVLM; (C) after spinal cord transsection. Rat vagotomizedwith both carotid sinus nerves cut. Same representations as in Fig. 2.
into RVLM (Fig. 6). The rise in pressure reached maximal value about 5 rain after microinjeetion. Sympathetic activity expressed as number of spikes counted over 160 s in 10 s periods and averaged per 1 s augmented significantly in the cervical and the lumbar trunks or in its dissected filaments (Fig. 7). No significant difference between sympathoexcitatory response in the cervical and in lumbar nerves was observed. Sympathoexcitatory effect lasted for about 15 rain parallel to the rise in pressure and outliving it slightly. SympathoimtE~fitory or sympathoexeitatoty effect of aortic nerve or vagus trunk stimulation respectively was abolished in the period of reduced power in PDS and ausmented syml~thetie activity which followed bilateral mierCnnjeefions of CoCI 2 or M ~ 2 into R V L M (Fig. 8).
T 150nlCo2"/ 150 nl salln
IS0 nl Mg=" 150 I11Co =. 150 nl mllne
2.5 - 5.0
5.5 - 7.5
Fig. 4. Summarized m e a n data of the relative power in three frequencies bands: 0.4-2.0 Hz, 2.5-5.0 Hz and 5.5-7.5 Hz. Triangles indicate m e a n power of spectra computed after bilateral saline microinjections into R V L M as a control; crosses mark the power computed after bilateral CoCI 2 microiniections into RVLM. All differences were highly significant. Maximal power was concentrated in the intermediate band of frequencies (2.5-5.0 Hz), therefore the effect of CoC12 was the most conspicious in this frequency band. There was no significant difference of r e m n a n t power between the three frequency bands remaining after CoCI 2 microinjections into RVLM.
Phrenic nerve response and respiratory-sympathetic synchronization
No significant effect of MgCI 2 and CoG12 microinjections upon phrenic nerve activity was observed. Some tendency to reduced peak amplitude of integrated phrenic discharge was insignificant. Synchronization of integrated sympathetic nerve discharge with the respiratory cycle disappeared totally. Consequently, a species- and
Fig. 6. Summarized m e a n blood pressure M A P response after bilateral microinjections of MgCI 2 solution (solid circles), CoCI 2 solution (crosses) and saline (triangles) into RVLM. S.E. marked by vertical bars. Blood pressure rise was significant after MgCI 2 and CoCI 2 microinjections into RVLM. No significant difference was found between pressor effects produced by MgC12 and CoCI 2 microinjections into RVLM. Ordinate: time in min. Abscissa: m m H g .
strain-dependent timing of sympathetic activity within the individual respiratory cycle characteristic to rats [11,12] was absent. A total dissociation of the phrenic nerve rhythmical respiratory activity from the augmented sympathetic nerve activity lasted as long as the depression of power in PDS spectra. Respiratory rhythm/sympathetic activity synchronization recovered after about 20 min following MgC12 or CoCI 2 microinjections into RVLM along with the return of PDS to control pattern. 30
• x •
150 nl Mg 2. 150 nl Co 2° 150 nl sollne
150 *'11M g 2. 150 nl sallne 2O
13 0.4 - 2.0
2.5 - 5.0
5.5 - 7.5
Fig. 5. Summarized m e a n values of the relative power in frequency bands 0.4-2.0 Hz, 2.5-5.0 Hz and 5.5-7.5 Hz in the control and after bilateral microinjections of MgCl 2 into RVLM. Same representation as in Fig. 4.
Fig. 7. Summarized m e a n response of the activity in the cervical and lumbar sympathetic nerves after CoCl 2 and MgCl 2 bilateral microinjections into RVLM. Ordinate: time in min. Abscissa: n u m b e r of spikes per s. Highly significant sympathoexcitatory response. No significant difference was found between the effects of MgC12 and CoC12 microinjections.
136 30 25 20 I£ .-~
lO I e 150 N M92÷
1S0 nl ¢ o ~
• 1S0 nl sc~lne Control
Aortic nerve stim.
Vagal afferen slim.
Fig. 8. Mean values of effects produced by baroreceptor aortic nerve stimulation on the activities in the cervical and in lumbar sympathetic nerves expressed as mean number of spikes per s. Control mean value of sympathetic activity was calculated after bilateral saline mieroinjections into RVLM (triangle) and is presented in the left section of the figure. This value is a reference point to the middle and right section of the figure. Control mean sympathetic activities recorded after CoCI 2 or MgCI 2 bilateral microinjections into RVLM are presented in the same left section. In the middle section the cross and solid circle mark sympathetic activity during aortic nerve stimulation on the background of effects produced by CoC12 and MgCI 2 microinjections into RVLM. In the right section mean sympathetic activity after CoCI z and MgCI 2 microinjection was so close that respective values overlapped in the figure and are marked as a single solid square. Very significant sympathoinhibitory effect of aortic nerve stimulation after saline mieroiniection into RVLM is evident in the middle section; The same stimulation on the background of effects produced by CoC1z or MgCI 2 microinjections into RVLM was entirely ineffective as the mean sympathetic activities during aortic nerve stimulation do not differ significantly from the control values prior to stimulation presented in the left section. The right section demonstrates a less pronounced yet significant sympathoexcitatory effect of vagal afferents stimulation after control saline microinjection in comparison with the control value marked by the triangle in the left section. The response was absent on the background of effects produced by CoCI z and MgCI 2 bilateral microinjections into RVLM.
Effect of kainic acid microinjected into RVLM on sympathetic rhythmicities and phrenic nerve activity 0.01 or 0.5 M of sodium kainate volumes (the same as used for MgC12 or CoCl 2) produced a powerful, transient sympathoexcitatory and pressure response after microinjection into RVLM. Respiratory rhythm became irregular, and after a n initial increase phrenic nerve discharge gradually diminished and ceased, along with a fall in
blood pressure and significant reduction of sympathetic activity. The effect was irreversible, and apnea lasted for hours. At the initial excitatory phase of kainic microinjection, PDS exhibited augmented power in all bands with conspicuous peaks within the respiratory frequency band and in the intermediate, non-respiratory and non-cardiac frequency range. After this transient excitatory phase, the total power in all frequency bands declined along with respiratory depression, blood pressure fall and significant reduction of non-synchronized overall sympathetic activity.
The major finding of this study is a dissociation between the tonic and the oscillatory activities in the sympathetic nerves by local block of synaptic transmission within the rostral ventrotateral medulla effectuated by calcium antagonists. Mg 2+ and Co 2+. As the total, yet non-synchronized, sympathetic nerve activity and arterial blood pressure increased under these conditions our results support the view of Guyenet [23,24] on intrinsic pacemaker activity of RVLM reticulospinal sympathoexcitatory neurons as the origin of the sympathetic vasomotor tone. Divalent Mg 2÷ and Co/+ cations are unspecific blockers of different kinds of calcium channels, including possibly postsynaptic ones which might be involved in the generation of pacemaker potential. As regenerative calcium spikes do not seem to occur in RVLM pacemaker cells  such a hypothesis appears unlikely. Besides, such an effect of calcium antagonists could hardly account for augmented sympathetic discharge, a response incompatible with the assumption of any functional impairment of RVLM reticulospinal sympathoexcitatory neurons. In contrast, a neurotoxic glutamate analogue, kainic acid, applied locally into RVLM produced a known depressor and sympatho-inhibitory effect. Kynurenate, a NMDA glutaminergic receptor antagonist, abolished sympathetic reflexes transmitted synaptically to RVLM slightly augmenting sympathetic activity , a finding similar to our
results obtained by the use of Mg 2+ and C o 2+ a s blockers of synaptic transmission. Although they are very unspecific blockers, they appear useful for the problem studied as the effects are transient and reversible, easily reproduced in the same experiments and limited to a restricted area by local microinjection. Our results are consistent with Guyenet's concept, as to his conclusion that nonsynaptic intrinsic pacemaker reticulospinal cells activity is mainly responsible for tonic sympathetic discharge. However, this concept and respective supporting evidences do not provide information crucial for the major subject of this study. Intrinsic pacemaker activity of sympathoexcitatory reticulospinal neurons is not by itself a likely candidate as a source of synchronized oscillatory pattern of sympathetic discharge. It would be difficult to imagine how single cell pacemaker activity of high frequency with a main discharge of 22 spikes/s  would be synchronized and integrated into slow frequency sympathetic efferent discharge unless some mechanisms of coordination of single cell rhythms were involved. Some increase in tonic, yet non-synchronized sympathetic activity after blocking local synaptic connections in RVLM suggests that an inhibitory synaptic component is necessary to generate the pattern of sympathetic oscillations. In general, our results support hypotheses [18,19,28-30] that neuronal circuitries are the source of centrally generated sympathetic rhythmicity. However, the organization and anatomical location of such oscillatory networks cannot be deduced from our findings. RVLM neurons may be a target of multiple interacting synaptic inputs, transmitting a flexible rhythmical pattern generated within neuronal networks located outside RVLM in various parts of the brain stem reticular system, including the dorsal segmental area [18,19] and other subsets of reticular network oscillators [29,30]. Another alternative hypothesis to be considered is that a basic pattern of synchronized sympathetic activity is generated in the ventral medulla by local synaptic interconnections between different functional neuronal subsets. These putative local circuitries could be activated, set going a n d / o r modulated from other brain regions involved in
various functions. Local inhibitory multisynaptic circuits between caudal and rostral medullary regions were suggested and reported including baroreceptor inhibitory input [13,22] which was used as a criterion of synaptic block in the present experiments. In order to determine which of the proposed hypotheses is correct, many studies over the years to come are required. An interesting observation was the presence of a cardiac-frequency related band in PDS despite bilateral blocking of synaptic transmission in RVLM in those animals in which carotid baroreceptors were intact (Fig. 2). This periodicity may reflect a direct rhythmical inhibition of sympathetic preganglionic neurons on spinal level by baroreceptor reflex input as reported in our earlier study  (see also [13,32]). A small fraction of power exhibited in PDS remained after bilateral blocking of synaptic transmission in RVLM and was not different from that which persisted after spinal transsection (Fig. 3). The pattern of this remnant PDS was similar to that in the, spinal animal: the power was concentrated in the intermediate band of 2-5 Hz, i.e. between the respiratory-like and cardiac-like frequency range. We suggest that remnant periodicities were generated by sympathetic oscillators on the spinal cord level . Bilateral block of synaptic transmission within RVLM appears to remove all periodicities generated on the supraspinal level. A surprising feature of the response to microinjection of divalent calcium antagonists into RVLM was an absence of any significant change in respiratory phrenic nerve discharge and frequency. In the rat, respiratory-related neurons are located mainly in the ventral and not in the dorsal medulla [14,37]. The volumes microinjected into RVLM were large and sites of penetration overlapped with the ventral respiratory group. As similar volumes of kainic acid resulted in apnea it seems unlikely to assume that a block of synaptic transmission did not affect the ventral respiratory group. Further research is required to elucidate this intriguing observation. Whatever the reason may be, divalent Ca 2+ antagonists proved to be a useful tool for separation of respiratory oscillator from the reticulospinal sympa-
thoexcitatory neurons. This demonstrates that sympathoexcitatory vasomotor tone generating cells are active, although synaptically separated from the respiratory oscillator, and corroborates our previous finding of the difference between sympathoexcitatory tone generation activity and respiratory oscillator in their respective threshold sensitivities to central CO 2 action . In conclusion, the present results support the view that the rostral ventrolateral medulla plays a crucial role as a neuronal pool where the patterns of centrally generated sympathetic periodicities are shaped out at the background of tonic nonsynchronized activity of the reticulospinal sympathoexcitatory neurons. An inhibitory mechanism seems to contribute to the rhythmical pattern generation within RVLM. Furthermore, respiratory and sympathetic central oscillators appear to be represented by neuronal subsets which can be uncoupled by the block of synaptic transmission within RVLM. We wish to emphasize that synaptic complexity of multiple neuronal circuits within the brain stem which control differently segmental frequencies at spinal level  presents a challenge for future long-term goal studies. Only functional tracing of synaptically organized neuronal networks may bring us closer to Dr. Brooks' question: what rhythms are functionally most significant? This fundamental question still awaits an answer.
Acknowledgement We thank Dr. Jolanta Sieminska for technical assistance and Mrs. Zofia Szymborska-Zydek for her work and generous help in preparing this manuscript. This work was supported by trd3N Grant 4 1728 9101p/03.
References 1 Adrian, E.D., Bronk, D.W. and Phillips, G., Discharges in mammalian sympathetic nerves, J. Physiol. (Lond.), 74 (1932) 115-133. 2 Baradziej, S. and Trzebski, A., Specific areas of the ventral
medulla controlling sympathetic and respiratory activities and their functional synchronization in the rat, Prog. Brain Res., 81 (1989) 193-204. Baradziej, S. and Trzebski, A., The role of the rostral ventrolateral medulla in the synchronization of respiratory and sympathetic functions. In G. Kunos and J. Ciriello (Eds.), Central Neural Mechanisms in Cardiovascular Regulation, Birkhauser, Boston, Basel, Berlin, 1991, pp. 69-82. Baradziej, S. and Trzebski, A., Neuronal population controlling the pattern of respiratory-sympathetic synchronization and sympathetic rhythmieities: parambigual area of the rostral ventrolateral medulla. In M. Yoshikawa, M. Uono, H. Tanabe and S. Ishikawa (Eds.), New Trends in Autonomic Nervous System Research, Excerpta Medica International Congress Series 951, Amsterdam, London, New York, Tokyo, 1991, pp. 316-319. Barman, S.M. and Gebber, G.L., Sympathetic nerve rhythm of brain stem origin, Am. J. Physiol., 239 (1980) R42-R47. Bendat, J.S. and Piersch, A.G., Random Data Analysis and Measurement Procedures, Wiley and Sons, New York, 1986, Ch. 12. Brooks, C. McC., Prospectives: relative to studies of blood pressure and its fluctuations. In K. Miyakawa, H.P. Koepchen, and C. Polosa (Eds.), Mechanisms of Blood Pressure Waves, Japan Sci. Soc. Press, Springer-Verlag, Tokyo/Berlin, Heidelberg, New York, 1984, pp. 351-355. Cannon, W.B., Organization for physiological homeostasis, Physiol. Rev., 9 (1929) 399-431. Ciriello, J., Caverson, M.M. and Polosa, C., Function of the ventrolateral medulla in the control of the circulation, Brain Res. Rev., 11 (1986) 359-391. Cohen, M.J. and Gootman, P.M., Periodicities in efferent discharge of splanchnic nerve of the cat, Am. J.' Physiol., 218 (1970) 1092-1101. Czyzyk, M.F., Fedorko, L, and Trzebski, A., Pattern of the respiratory modulation of the sympathetic activity is species dependent: synchronization of the sympathetic outflow over the respiratory cycle in the rat. In J. Cirielto, F.R. Calaresu, L.P. Renaud and C. Polosa (Eds~), Organization of the Autonomic Nervous System, Central and Peripheral Mechanism, Neurology and Neurobiology, Vol. 31, Alan R. Lis, New York, 1987, pp. 143-152. Czyzyk-Krzeska, M. and Trzebski, A., Respiratory-related discharge pattern of sympathetic nerve activity in the spontaneously hypertensive rat, J. Physiol. (Lond.), 426 (1990) 355-368. Dembowsky, K., Czachurski, J. and Seller, H., Some properties of the sympathoinhibition from the caudal ventrolateral medulla oblongata in the cat, Prog. Brain Res., 8t (1989) 143-158. Ezure, K., Manabe, M and Yamada, H., Distribution of medullary respiratory neurons in the rat, Brain Res., 455 (1988) 262-270. Fedorko, L., Lioy, F. and Trzebski, A., Synaptic inhibition of preganglionic neurons induced by the stimulation of the aortic nerve in the cat, J. Physiol. (Londl), 360 (I985) 45P.
139 16 Feldman, J.R., Smith, J.C., Ellenberger, H.H., Connely, C.A., Liu, G., Greer, J.J., Lindsay, A.D. and Otto, M.R., Neurogenesis of respiratory rhythm and pattern: emerging concepts, Am. J. Physiol., 259 (1990) 879-886. 17 Gebber, G.L., Central oscillators respo/lsible for sympathetic nerve discharge, Am. J. Physiol., 239 (1980) H143H155. 18 Gebber, G.L. and Barman, S.M., A physiologically-based model of the brain stern generator of sympathetic nerve discharge, Prog. Brain Res., 81 (1989) 131-139. 19 Gebber, G.L., Central determinants of sympathetic nerve discharge. In A.D. Loewy and K.M. Spyer (Eds.), Central Regulation of Autonomic Functions, Oxford Univ. Press, 1990, pp. 126-144. 20 Gleick, J., Chaos Making a New Science, Penguin, New York, 1987. 21 Gootman, P.M. and Cohen, M.J., Sympathetic rhythms in spinal cats, J. Auton. Nerv. Syst., 3 (1981) 379-387. 22 Granata, A.R., Numao, Y., Kumada, M. and Reis, D.J., A1 noradrenergic neurons tonically inhibit sympathoexcitatory neurons of C1 area in rat brainstem, Brain Res., 377 (1986) 127-146. 23 Guyenet, P.G., Role of the ventral medulla oblongata in blood pressure regulation. In A.D. Loewy and K.M. Spyer (Eds.), Central Regulation of Autonomic Functions, Oxford Univ. Press, 1990, pp. 145-167. 24 Guyenet, P.G., Haselton, J.R. and Sun, M.K., Sympathoexcitatory neurons of the rostroventral medulla and the origin of the sympathetic vasomotor tone, Progr. Brain Res., 81 (1989) 105-116. 25 Guyenet, P.G., Darnal, R.A. and Riley, T.A., Rostral ventro-lateral medulla and sympathorespiratory integration in rats, Am. J. Physiol., 259 (1990) R1063-R1074. 26 Haken, H., Advanced Synergetics, Springer, Berlin, Heidelberg, New York, 1983. 27 Kenney, M.J., Barman, S.M., Kocsis, B. and Gebber, G.L., Flexibility of the map of brainstem neurons with sympathetic nerve-related activity. In G. Kunos, and J. Ciriello (Eds.), Central Neural Mechanisms in Cardiovascular Regulation, Birkhauser, Boston, Basel, Berlin, 1991, pp. 37-54. 28 Koepchen, H.P., Respiratory and cardiovascular "centres": functional entitity or separate structures? In M.E. Schlafke, H.P. Koepchen, and W.R. See (Eds.), Central Neurone Environment and the Control Systems of Breathing and Circulation, Springer-Verlag Berlin, Heidelberg, New York, 1983, pp. 221-235. 29 Koepchen, H.P., Abel, H.H. and Klussendort, D., Brain stem generation of specific and non-specific rhythms. In J. Ciriello, F.R. Calaresu, L.P. Renaud and C. Polosa (Eds.), Organization of the Autonomic Nervous System. Central and Peripheral Mechanisms, Neurology and Neurobiology, Vol. 31, Alan Liss, New York 31, 1987, pp. 179-188. 30 Langhorst, P., Schultz G. and Lambertz, M., Oscillating neuronal network of the "common brain stem system". In
35 36 37
K. Miyakawa, H.P. Koepchen and C. Polosa (Eds.), Mechanisms of Blood Pressure Waves, Japan Sci. Soc. Press Springer-Verlag, Tokyo, Berlin, Heidelberg, New York, 1984, pp. 257-275. McAllen, R.M. and Dampney, R.A.L., The selecting of descending vasomotor control by subretrofacial neurons, Prog. Brain Res., 81 (1989) 233-242. McLachlan, E.M. and Hirst, G.D.S., Some properties of the preganglionic neurons in upper thoracic spinal cord of the cat, J. Neurophysiol., 43 (1980) 1251-1265. Nicolis, G. and Prigogine, J., Self-Organization in Nonequillibrium Systems, New York, 1977. Paton, J.F., Rogers, W.T. and Schwaber, J.S., Tonically rhythmic neurons within a cardiorespiratory region of the nucleus tractus solitarii of the cat, J. Neurophysiol., 66 (1991) 824-838. Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Coordinates, 2nd Edn., Academic Press, Sydney, 1986. Schuster, H.G., Deterministic Chaos, VCH, Weinheim, 1988. Smith, J.C., Greer, J.J., Liu, G. and Feldman, J.L., Neural mechanisms generating respiratory pattern in mammalian brain stem-spinal cord in vitro. Spatiotemporal patterns of motor and medullary neuron activity, J. Neurophysiol., 64 (1990) 1149-1169. Sun, M.K. and Guyenet, P.G., Arterial baroreceptor and vagal inputs to sympathoexcitatory neurons in rat medulla, Am. J. Physiol., 252 (1987) R69%-R709. Sun, M.K., Hackett, J.T. and Guyenet, P.G., Sympathoexcitatory neurons of rostral ventrolateral medulla exhibit pacemaker properties in presence of a glutamate-receptor antagonist, Brain Res., 438 (1988) 23-40. Sun, M.K., Young, B.S., Hackett, J.T. and Guyenet, P.G., Reticulospinal pacemaker neurons of the rat rostral ventrolateral medulla with putative sympathoexcitatory function: an intracellular study in vitro, Brain Res., 442 (1988) 229-239. Trzebski, A. and Kubin, L., Is the central inspiratory activity responsible for pCO2-dependent drive of the sympathetic discharge?, J. Auton. Nerv. Syst., 3 (1981) 401420. Trzebski, A. and Baradziej, S., Role of the rostroventral medulla in the generation of rhythmicities of the sympathetic activity. In H. Haken and H.P. Koepchen (Eds.), Rhythms in Physiological Systems, Springer-Verlag, 1991, pp. 61-68. Varner, K.J., Vasquez, E.C., Lewis, S.J., Machado, B.H., Grosskreutz, C.L., Simon, J.S. and Brody, M.J., Regulation of autonomic cardiovascular function by the rostral ventromedial medulla. In G. Kunos and J. Ciriello (Eds.), Central Neural Mechanisms in Cardiovascular Regulation, Birkhauser, Boston, Berlin, 1991, pp. 29-36. Wiener, N., Cybernetics or Control and Communication in the Animal and in the Machine, John Wiley, Paris, Cambridge, New York, 1948.