A study of renal-efferent neurones and their neural connexions within cat renal ganglia using intracellular electrodes

Differential Distribution of Renal Nerves in the Sympathetic Ganglia of the Rat

The renal nerve plexus comprises efferent and afferent fibers. It controls urine production and bodily fluid homeostasis. Efferent fibers to the kidney include sympathetic nerve fibers from their main ganglia, the prevertebral suprarenal ganglia (SrG), and the paravertebral sympathetic chain ganglia (ChG). In the present study, we examined topological innervation from these ganglia to the renal parenchymal segments of the left kidney of the rat. Fluoro-Gold was injected into the rostral or caudal poles of the left kidney. Approximately 50% of the cells in the SrG of rats injected in the rostral pole were labeled, while 60% of the cells in the ChG T13 of rats injected in the caudal pole were labeled. In addition, we performed dual-probe retrograde tracing of the nerves using two kinds of fluorescent-conjugated cholera toxins (f-CTbs) injected into the rostral and caudal poles of the left kidney. The cells labeled with each f-CTb were distributed differently in the left SrG and the lower ChGs; no dual-labeled cells were found in these ganglia. Anterograde tracing with pCAGGS-tdTomato vector transfected into the left SrG showed that tdTomato-labeled nerve varicosities extended to the cortical arterioles and urinary tubules. Immunohistochemistry revealed that they were positive to tyrosine hydroxylase and synaptophysin, suggesting that they possessed sympathetic nerve endings. Our results show that renal efferent nerves in the SrG may control the rostral part of the kidney and innervate the multiple effectors in the cortex. Anat Rec, 300:2263-2272, 2017. © 2017 Wiley Periodicals, Inc.

Neurotrophins and target interactions in the development and regulation of sympathetic neuron electrical and synaptic properties

The electrical and synaptic properties of neurons are essential for determining the function of the nervous system. Thus, understanding the mechanisms that control the appropriate developmental acquisition and maintenance of these properties is a critical problem in neuroscience. A great deal of our understanding of these developmental mechanisms comes from studies of soluble growth factor signaling between cells in the peripheral nervous system. The sympathetic nervous system has provided a model for studying the role of these factors both in early development and in the establishment of mature properties. In particular, neurotrophins produced by the targets of sympathetic innervation regulate the synaptic and electrophysiological properties of postnatal sympathetic neurons. In this review we examine the role of neurotrophin signaling in the regulation of synaptic strength, neurotransmitter phenotype, voltage-gated currents and repetitive firing properties of sympathetic neurons. Together, these properties determine the level of sympathetic drive to target organs such as the heart. Changes in this sympathetic drive, which may be linked to dysfunctions in neurotrophin signaling, are associated with devastating diseases such as high blood pressure, arrhythmias and heart attack. Neurotrophins appear to play similar roles in modulating the synaptic and electrical properties of other peripheral and central neuronal systems, suggesting that information provided from studies in the sympathetic nervous system will be widely applicable for understanding the neurotrophic regulation of neuronal function in other systems.

Dual mechanisms of angiotensin-induced activation of mouse sympathetic neurones

Ang II directly activates neurones in sympathetic ganglia. Our goal was to define the electrophysiological basis of this activation. Neurones from mouse aortic-renal and coeliac ganglia were identified as either 'tonic' or 'phasic'. With injections of depolarizing currents, action potentials (APs) were abundant and sustained in tonic neurones (TNs) and scarce or absent in phasic neurones (PNs). Resting membrane potentials were equivalent in TNs (-48 +/- 2 mV, n = 18) and PNs (-48 +/- 1 mV, n = 23) while membrane resistance was significantly higher in TNs. Ang II depolarized and increased membrane resistance equally in both TNs (n = 8) and PNs (n = 8) but it induced APs only in TNs, and enhanced current-evoked APs much more markedly in TNs (P < 0.05). The AT1 receptor antagonist losartan (2 microm, n = 6) abolished all responses to Ang II, whereas the AT2 receptor blocker PD123,319 had no effect. The transient K+ current (IA), which was more than twice as large in TNs as in PNs, was significantly inhibited by Ang II in TNs only whereas the delayed sustained K+ current (IK), which was comparable in both TNs and PNs, was not inhibited. M currents were more prominent in PNs and were inhibited by Ang II. The IA channel blocker 4-aminopyridine triggered AP generation in TNs and prevented the Ang II-induced APs but not the depolarization. Blockade of M currents by oxotremorine M or linopirdine prevented the depolarizing action of Ang II. The protein kinase C (PKC) inhibitor H7 (10 microm, n = 9) also prevented the Ang II-induced inhibition of IA and the generation APs but not the depolarization nor the inhibition of M currents. Conversely, the PKC agonist phorbol 12-myristate 13-acetate mimicked the Ang II effects by triggering APs. The results indicate that Ang II may increase AP generation in sympathetic neurones by inducing a PKC-dependent inhibition of IA currents, and a PKC-independent depolarization through inhibition of M currents. The differential expression of various K+ channels and their sensitivity to phosphorylation by PKC may determine the degree of activation of sympathetic neurones and hence may influence the severity of the hypertensive response.

Angiotensin selectively activates a subpopulation of postganglionic sympathetic neurons in mice

Angiotensin II (Ang II) increases renal sympathetic nerve activity in anesthetized mice before and after ganglionic blockade, suggesting that Ang II may directly activate postganglionic sympathetic neurons. The present study directly tested this hypothesis in vitro. Neurons were dissociated from aortic-renal and celiac ganglia of C57BL/6J mice. Cytosolic Ca(2+) concentration ([Ca(2+)](i)) was measured with ratio imaging using fura 2. Ang II increased [Ca(2+)](i) in a subpopulation of sympathetic neurons. At a concentration of 200 nmol/L, 14 (67%) of 21 neurons responded with a rise in [Ca(2+)](i). The Ang II type 1 (AT(1)) receptor blocker (losartan, 2 micromol/L) but not the Ang II type 2 (AT(2)) receptor blocker (PD123,319, 4 micromol/L) blocked this effect. The Ang II-induced [Ca(2+)](i) increase was abolished by removal of extracellular Ca(2+) but not altered by depletion of intracellular Ca(2+) stores with thapsigargin. Ang II no longer elicited a [Ca(2+)](i) increase in the presence of lanthanum (25 micromol/L). The specific N-type and L-type Ca(2+) channel blockers, omega-conotoxin GVIA and nifedipine, respectively, significantly inhibited the Ang II-induced [Ca(2+)](i) increase. The protein kinase C inhibitor H7 but not the protein kinase A inhibitor H89 blocked the response to Ang II. These results demonstrate that Ang II selectively activates a subpopulation of postganglionic sympathetic neurons in aortic-renal and celiac ganglia, triggering Ca(2+) influx through voltage-gated Ca(2+) channels. This effect is mediated through AT(1) receptors and requires the activation of protein kinase C. The activation of a subgroup of sympathetic neurons by Ang II may exert unique effects on kidney function in pathological states associated with elevated Ang II.

Electrophysiological properties of lumbosacral preganglionic neurons in the neonatal rat spinal cord

The electrophysiological properties of parasympathetic preganglionic neurons (PGN) in L6 and S1 spinal cord slices from neonatal rats were studied using the patch clamp techniques. PGN were identified by retrograde axonal transport of a fluorescent dye (Fast Blue) injected intraperitoneally before the experiment. PGN in the intermediolateral region of the spinal cord were divided into two classes (tonic PGN and phasic PGN) on the basis of firing properties during prolonged (300 ms) depolarizing current pulses. Tonic neurons exhibited a prolonged discharge (average maximum: 5.6); whereas phasic PGN fired on average only 1.4 spikes during depolarizing pulses. PGN were usually oval in shape. The mean long axis of tonic PGN (20.7+/-0.5 microm) was significantly (P<0.05) larger than that of phasic PGN (16.7+/-0.3 microm). Tonic and phasic PGN had similar resting membrane potentials, thresholds for spike activation, input resistances and action potential durations. The duration of the after-hyperpolarization (AHP) in tonic PGN (200.5+/-11.9 ms) was longer than in phasic PGN (137.6+/-9.8 ms). 4-aminopyridine (4-AP, 0. 5 mM) reduced the threshold for spike activation in tonic and phasic PGN. 4-AP also unmasked tonic firing in phasic PGN (average maximum: 5.5 spikes during 300 ms depolarizing current pulses) and increased firing frequency by 19% in tonic PGN. These data indicate that the different discharge patterns of parasympathetic PGN are dependent in part on differences in the expression of 4-AP-sensitive K(+) channels. The two types of PGN may provide an innervation to different targets in the pelvic viscera.

Potassium currents in rat prevertebral and paravertebral sympathetic neurones: control of firing properties

1. Intracellular recordings were made from rat sympathetic neurones in isolated superior cervical ganglia (SCG), coeliac ganglia (CG) and superior mesenteric ganglia (SMG). 2. Based on their response to a maintained depolarizing current stimulus, neurones were classified as 'phasic' or 'tonic'. All neurones in the SCG were phasic, 85% of the neurones in the SMG and 58% of the neurones in the CG were tonic, and the remainder were phasic. 3. The voltage response of phasic and tonic neurones around threshold to a constant current step was markedly different. The response of phasic neurones was biphasic with an initial depolarizing response followed by significant repolarization of the membrane potential. In contrast, tonic neurones became more depolarized during a prolonged current step. 4. The underlying currents were studied using single-electrode voltage-clamp recording. The M-current was present in all phasic neurones, but was very weak or absent in tonic neurones. 5. An A-current was apparent in both phasic and tonic neurones. The voltage-dependent activation, steady- state inactivation, and current density of the A-current were all similar in phasic and tonic cells. 6. A low- threshold, slowly inactivating outward current (D2-current) was observed exclusively in tonic neurones. The slow inactivation of this current appeared to underlie the slow depolarizing ramp seen in response to a maintained depolarizing current step. 7. Computer simulations, based on the voltage-clamp data, suggested that the different firing properties of phasic and tonic neurones could be accounted for by differential expression of the M-current.

Transitory noradrenergic and peptidergic nerves in the cat kidney

Indirect immunohistochemical methods were used to visualize nerves immunoreactive for tyrosine hydroxylase (TH), dopamine beta hydroxylase (DBH), neuropeptide Y, (NPY) and calcitonin gene-related peptide (CGRP) in sections of the kidneys of cats of different ages. Nerve terminals immunoreactive for TH, DBH and NPY innervated interlobar veins and the renal arterial tree including medullary vascular bundles of cats of each age studied. Most nerve terminals immunoreactive for CGRP innervated interlobar arteries. In kidneys of cats 2 to 10 weeks old, TH- and DBH-immunoreactive axons formed elaborate plexuses that were distributed throughout much of the outer two thirds of the inner medulla. Inner medullary NPY-immunoreactive nerve terminals formed sparse plexuses by comparison, thus suggesting a large population of TH-immunoreactive nerve terminals not immunoreactive for NPY. Plexuses immunoreactive for CGRP also innervated the inner medullae of young cats. Some inner medullary axons appeared degenerate in 8 and 10 week old cats, and no inner medullary nerve terminal plexuses were visualized in 12 week old or adult cats. Cell death or paring of axons resulting from mechanisms intrinsic to the neuronal population or from a change in trophic factors secreted or expressed by cells in the medulla may effect the loss of inner medullary nerve terminals in the kidneys of young cats.

Single-channel and whole-cell recordings from non-dissociated sympathetic neurones in rabbit coeliac ganglia

A procedure is described for performing patch-clamp recordings on mammalian sympathetic neurones within intact ganglia. The plasma membrane of superficial neurones was cleaned by blowing (1.5-3 h) a gentle stream of Ringer saline onto ganglia, the connective sheath of which was previously softened by a short protease treatment. This procedure preserved the intraganglionic connectivity so that the neurones could be activated either synaptically or antidromically by stimulating the appropriate nerves. Depending on the duration of the mechanical cleaning step, recordings were performed on either the neurones or the satellite glial cells covering the neuronal cell bodies. The applicability of the various configurations of the patch-clamp technique to studying sympathetic neurones is illustrated by recordings of whole-cell voltage, whole-cell currents and single-channel currents in cell-attached and excised patches. With these techniques, the resolution of the membrane current recordings is higher than with conventional microelectrodes. The results obtained show that mammalian sympathetic neurones have a very high input resistance (0.5 G omega), are electronically compact and may display pacemaker activity. These techniques provide a useful tool for studying the synaptic transmission and neuromodulation mechanisms operating within the sympathetic ganglia.

Origins of the renal innervation in the primate, Macaca fascicularis

The origins of the renal efferent and afferent nerves in 5 cynomolgus monkeys (Macaca fascicularis) were studied by using the retrograde transport of horseradish peroxidase (HRP) and horseradish peroxidase-wheat germ agglutinin (HRP-WGA). The cut ends of the right renal nerves were soaked for 30-45 min in solutions consisting of 15% HRP and 1% HRP-WGA. Three or four days later the animals were killed and the tissues examined for the presence of retrogradely labeled neurons, HRP-filled cells were observed, with rare exceptions, only in ganglia ipsilateral to the side of tracer application. Renal efferent neurons (4648-14565 cells per animal) were found in relatively equal numbers in prevertebral and paravertebral (sympathetic chain) ganglia. Labeled prevertebral cells were distributed among the renal (52%), aorticorenal (32%) and superior mesenteric (16%) ganglia, whereas labeled paravertebral neurons were mainly located in chain ganglia T11-L3, with 94% of these located in L1-3. Labeled renal sensory neurons (31-543 per animal) constituted less than 5% of all labeled cells and were found in ipsilateral dorsal root ganglia T10-L3, with (80%) in T12 and L1. The labeled sensory neurons ranged from 18-64 microns in diameter (X = 32.4 microns). With the exception of a single cell in one animal, no labeled neurons were observed in the nodose ganglia. Many parallels were observed between the organization of the renal plexuses of macaques and humans, suggesting the utility of the non-human primate as an experimental model for functional studies of renal innervation in humans.

Cardiac sympathetic afferent cell bodies are located in the peripheral nervous system of the cat

Studies of the stellate ganglion and middle cervical ganglion indicate that sympathetic efferent nerve activity can be modified by peripheral excitatory inputs and that these neural connections may function as pathways for a peripheral reflex at the level of the thoracic sympathetic ganglia. This excitatory synaptic input could have a soma in either the central or the peripheral nervous system. A study was designed to determine whether chronic decentralization (3 weeks) of the stellate ganglion in cats would 1) abolish sympathetic cardiac afferent nerve activity recorded at the stellate cardiac nerve and 2) abolish local thoracic reflexes that are generated by stimulation of peripheral nerves. The ansae subclaviae, T3 and T4 rami, and stellate ganglion were also examined by electron microscopy for the extent of Wallerian degeneration. Afferent cardiac activation of the axon collaterals arising from cell bodies located in the dorsal root ganglia was abolished due to degeneration. However, sympathetic afferent nerve activity from the left ventricular receptors was still present and was recorded from the stellate cardiac nerve in all cats. Cardiac receptors were sensitive to mechanical distortion, increases in the left ventricular pressure, and epicardial application of veratrine hydrochloride. These data imply that 1) cardiovascular afferent input to the stellate ganglion persists following chronic decentralization and 2) the sensory neurons are located in the peripheral sympathetic nervous system. Thus, we find that regulation of the heart occurs in part via thoracic ganglia, independently of the central nervous system.

Electrophysiological characteristics of sensory mechanisms in the kidney

R2 chemoreceptors are excited by backflow of urine or isotonic KCl into the renal pelvis, they do not respond to backflow of isotonic saline at the same intrapelvic pressure but are excited during periods of renal ischemia. R1 chemoreceptors are excited during complete renal ischemia but otherwise exhibit no activity. The responses of multiunit afferent renal nerve activity (ARNA) to these stimuli in Sprague Dawley rats follow the same patterns exhibited by R2 chemoreceptors, and the data do not support the presence of mechanoreceptive nerves which are excited by increases in intrapelvic pressure in these animals. The responses of multiunit ARNA in SHR and WKY rats were not different from Sprague Dawley rats. In contrast, while the response of WKY rats during renal ischemia was not different from Sprague Dawley rats the excitation of SHR during ischemia was more than 10 fold greater than that of the normotensive animals.

The extrinsic innervation of the rat kidney

The anatomy of the extrinsic renal nerves is described making use of an in toto staining procedure for acetylcholinesterase activity. Nervous connections between the kidney and the celiac plexus, major and minor splanchnic nerves, the lumbar splanchnic nerves and the intermesenteric nerve plexus have been established. Bundles of nerve fibers enter or leave the kidney in the (peri-) hilar region. The occurrence of inter- and intraindividual variability is emphasized. Implications of these findings for the use of the rat as an experimental animal in denervation experiments are discussed.

Characteristics of phasic and tonic sympathetic ganglion cells of the guinea-pig

Intracellular recording techniques have been used to determine the electrophysiological properties of sympathetic neurones in ganglia of the caudal lumbar sympathetic chain (l.s.c.) and in the distal lobes of inferior mesenteric ganglia (i.m.g.) isolated from guinea-pigs. Passage of suprathreshold depolarizing current initiated transient bursts of action potentials in 97% of l.s.c. neurones, but only 13% of i.m.g. cells ('phasic' neurones). Most i.m.g. neurones fired continuously during prolonged depolarizing pulses ('tonic' neurones). Passive membrane properties varied; mean cell input resistance was similar between groups, but phasic neurones had smaller major input time constants on average than had tonic cells. Current-voltage relations determined under both current clamp and voltage clamp were linear around resting membrane potential (approximately 60 mV), where membrane conductance was lowest. Instantaneous and time-dependent rectification varied in the different neurone types. The current underlying the after-hyperpolarization following the action potential was significantly larger on average in tonic i.m.g. cells than in phasic neurones, although its time course (tau = 100 ms) was similar. Phasic neurones fired tonically when depolarized after adding the muscarinic agonist, bethanechol (10(-5) M to 10(-4) M), to the bathing solution. Bethanechol blocked a proportion of the maintained outward current (presumably the M-current, IM, Adams, Brown & Constanti, 1982) in phasic neurones; this current was small or absent in tonic neurones. Transient outward currents resembling the A-current (IA, Connor & Stevens, 1971 a) were evoked in tonic but not in phasic neurones by depolarization from resting membrane potential. IA could only be demonstrated in phasic neurones after a period of conditioning hyperpolarization. After a step depolarization to approximately --50 mV, IA reached peak amplitude at about 7 ms and then decayed with a time constant of about 25 ms in both neurone types. Activation characteristics of IA were similar for phasic and tonic neurones, but inactivation curves, although having the same shape, were shifted to more depolarized voltages in tonic neurones. That is, IA was largely inactivated at resting membrane potential in phasic, but not tonic neurones. It is concluded that the discharge patterns of the two populations of sympathetic neurones result from differences in the voltage-dependent potassium channels present in their membranes. The anatomical occurrence of the different cell types suggests that phasic neurones are vasoconstrictor and tonic neurones are involved with visceral motility.

Cardiac responses to stimulation of discrete loci within canine sympathetic ganglia following hexamethonium

In dogs which had been administered hexamethonium, localized stimulation with bipolar electrodes in the middle cervical and stellate ganglia as well as ganglia in the caudal pole nerve and ventral ansa elicited alterations in cardiodynamics. The middle cervical and stellate ganglia were divided into 30 discrete regions in order to provide a standardized map of these ganglia. Of all the responses elicited. 5% consisted of increased heart rate while 38% consisted of augmented inotropism without accompanying chronotropic changes. Both rate and force were augmented in 57% of the responses. The sites in which stimulation consistently elicited chronotropic responses were located in right-sided ganglia. Loci which consistently generated force changes on stimulation were in the middle cervical and stellate ganglia bilaterally. Cardiac chronotropism and inotropism were augmented by stimulation of ganglia located in the ventral subclavian ansae or the left caudal pole nerve. No cardiac responses were elicited by stimulation within the superior cervical ganglia. It is concluded that localized stimulation of specific sites in the major thoracic ganglia or smaller mediastinal ganglia can alter cardiac chronotropism and/or inotropism.

Comparison of the distributions of renal and splenic neurons in sympathetic ganglia

Postganglionic neurons in different sympathetic ganglia are innervated selectively by preganglionic neurons originating from different segments of the spinal cord. These selective connections between pre- and postganglionic neurons may determine the specificity with which postganglionic nerves participate in differential reflex reactions. Because specificity of renal and splenic nerve responses to stimulation of visceral afferent nerves may depend on the distribution of postganglionic neurons in sympathetic ganglia, retrograde axonal transport of horseradish peroxidase was employed in this study to identify the ganglionic distribution of cell bodies of postganglionic neurons supplying the kidney and spleen in cats. Superior mesenteric, left and right celiac ganglia usually are fused together into a complex ganglion (solar plexus) in the cat. Most labeled cell bodies of renal nerves were clustered in groups within the solar plexus, but some cell bodies of renal neurons were observed in upper lumbar (L1-L3) and lower thoracic (T12-T13) paravertebral sympathetic ganglia. In contrast, 90% of labeled splenic neurons were scattered randomly throughout the left and right celiac poles of the solar plexus. In conclusion, the disparate distribution of renal and splenic neurons in sympathetic ganglia provides an anatomical basis for differential reflex responses in the two populations of nerves.

An intracellular characterization of neurones and neural connexions within the left coeliac ganglion of cats

Intracellular recordings were made in vitro from neurones located within the left coeliac ganglion of the cat solar plexus. Thirty percent of the neurones within left coeliac ganglia were identified as efferent neurones. Within this neuronal population, splenic-efferent and renal-efferent neurones were identified specifically. Neurones within left coeliac ganglia were characterized as either phasic (fast adapting) neurones or tonic (slowly adapting) neurones depending upon their prolonged firing behaviour. Electrophysiological properties of neurones varied considerably. The wide range of values obtained for both input resistance and input capacitance suggest that sizeable differences in either specific membrane resistance or cell geometry exist within the over-all neurone population. Frequency distributions of input resistance, time constant, input capacitance and current threshold for tonic and phasic neurones were found to be significantly different. Compound excitatory post-synaptic potentials were produced by stimulation of the ipsilateral splanchnic nerves in 69% of the neurones tested and in 3% of the neurones tested upon stimulation of the contralateral splanchnic nerves. Electrical stimulation of nerve fibres located in the coeliac plexus, the superior mesenteric plexus or the left renal nerves generated excitatory synaptic potentials in neurones located within left coeliac ganglia. It is concluded that neurones within the left coeliac ganglion are innervated by splanchnic nerve fibres primarily contained within the left splanchnic nerves, receive excitatory synaptic input from splenic, renal and other peripheral preganglionic fibres and have extremely varied electrophysiological properties.

Single-unit and multiunit analyses of renorenal reflexes elicited by stimulation of renal chemoreceptors in the rat

Stimulation of renal chemoreceptors, induced by backflow of concentrated urine into the renal pelvis, elicited a reflex increase in the rate of multiunit efferent renal nerve activity in rats. Ipsilateral stimulation increased the rate by 17.1 +/- 4.5%; contralateral stimulation increased the rate by 20.8 +/- 4.5%. Efferent activity returned toward control levels within the first minute following removal of the stimulus. The reflex responses were quantitatively similar in rats anesthetized with sodium pentobarbital or alpha-chloralose, but increased following complete transection of the spinal cord at C3. Studies of single efferent units indicated that responses to the stimuli were nonuniform. The majority of single units exhibited an increase in firing frequency during the stimulus; other units showed no change or, less frequently, a reduction in activity during the stimulus. Some postganglionic neurons projecting to the kidney were activated by stimuli applied to either kidney. It was concluded that stimulation of intrarenal chemoreceptors provokes bilateral excitatory renorenal reflexes in the rat. The observation of nonuniform responses among single units suggests that renal nerves contain a mixture of efferent fibers which may contact different targets in the kidney.

Physiological studies of canine sympathetic ganglia and cardiac nerves

Anatomical studies have indicated that the middle cervical ganglion is the primary locus of sympathetic postganglionic neurons which have axons projecting in canine cardiac nerves. Small regions of thoracic sympathetic ganglia were stimulated electrically with bipolar electrodes and the generated compound action potentials were recorded from ipsilateral cardiac nerves. Several procedures were used to distinguish the effects of stimulating preganglionic axons, ganglionic sites and postganglionic axons. Following hexamethonium it was found that only low frequency stimulation (e.g. 1 Hz) of ganglionic sites resulted in the generation of compound action potentials in cardiac nerves. These ganglionic sites were located throughout the middle cervical ganglion, but only in the cranial medial pole of the stellate ganglion. Stimulation of a single locus in a stellate or middle cervical ganglion generated compound action potentials in all major ipsilateral sympathetic cardiac nerves. A ganglion was also identified at the junction of the thoracic vagus and the cranial vagal nerve which may be as important as the right stellate ganglion in cardiac regulation. These results demonstrate that there are discrete loci in thoracic ganglia which contain neurons projecting in several ipsilateral cardiac nerves. Low frequency stimulation of these loci was found to be optimal for generation of compound action potentials.