Impulse Activity and Pattern of Large and Small Neurones in the Cardiac Ganglion of the Lobster, Panulirus Japonicus

In Panulirus evidence was provided that two small cells of different types in the repetitive discharge pattern innervate large cells, controlling their activities; one of the two (primary pacemaker neurone) induces small synaptic potentials, and the other (secondary pacemaker neurone) induces large synaptic potentials (Tazaki, 1971 b). Fig. 1 illustrates a typical pattern of burst in a large-cell soma. The burst was mainly composed of large and small synaptic potentials with small-sized spikes (Fig. 1 A). This was readily shown by the experiment of current injection. Separation of large and small synaptic potentials was performed by applying hyperpolarizing current pulses with varying intensity during the burst discharge. Large synaptic potentials remained with a weak current pulse (Fig. 1B), but they were completely eliminated with a strong current pulse (Fig. 1C). On the other hand, small synaptic potentials usually remained during the applied hyperpolarization, and their amplitude increased. This phenomenon has been interpreted as follows: there is electrotonic interaction between large follower and small pacemaker neurones (Watanabe & Bullock, 1960). Apparendy, the electrotonic spread was almost negligible between followers and primary pacemaker, while it was rather effective between followers and secondary pacemaker. The more detailed description of small synaptic potentials has been reported in a previous paper (Tazaki, 1971b). A great deal of information concerning synaptic interaction among neurones has been obtained, but impulse activity of large cells has not been so thoroughly studied because the recording site has been confined to their somata. Furthermore, almost nothing is known about the electrical activity of the small cells. Hartline (1967) has reported impulse identification of the axons of all nine ganglion cells by the axon-mapping experiments in Homarus.

This communication reports the impulse activity and pattern of the component neurones in the lobster cardiac ganglion. Successful impalement of the large-cell axon was performed at the region near the cell soma, and unit impulses could be recorded in the regions of small cells. Single-unit analysis was made by means of simultaneous internal and external recordings, with the result that several kinds of impulses were propagated within the ganglion. The more detailed analysis of the activity and interaction of all nine ganglion cells was attempted.

The cardiac ganglion of the Japanese spiny lobster, Panulirus japonicus, was used. The ganglion-cell somata can be readily identified on the basis of size and position. Nine ganglion cells were designated Cell 1, Cell 2, etc., from anterior to posterior: Cells 1–5 were large neurones, and Cells 6–9 small neurones (Maynard, 1955). Intracellular recordings and current injection were made by means of glass capillary microelectrodes filled with 3 M-KCI. Impulses of a large cell were recorded intracellularly from its axon. In some cases, since impalement of the small cell was very difficult, unit impulses were externally recorded from the regions of small cells in the posterior trunk by microelectrodes. In this case a single unit was identified on the basis of the following criteria: the amplitude of impulses was constant, and no electrical summation occurred.

Cells 1–4 showed almost the same pattern of impulse discharges. Cell 5, which is situated in the middle of the ganglion, however, exhibited electrical activity somewhat different from that of other large cells (Bullock & Terzuolo, 1957). Although all large cells behave as followers, being synaptically controlled by small pacemaker neurones, they come to have pacemaker activity when the posterior part of the trunk which contains four small-cell somata has been removed. The impulse activity of Cell 5 was examined in the sectioned preparation, where a simple pattern of endogenous activity eventually appeared because synaptic inputs from the small cells were excluded.

The time course of repetitive discharges was shown in a graph where impulse intervals calculated as frequencies were plotted on the y-axis against time of onset of impulse on the x-axis. This illustrates well the pattern of repetitive firing in the component neurones.

Other detailed experimental procedures were described in a previous paper (Tazaki, 1971b). The experiment was carried out at room temperatures ranging from 22 to 24°C.

As shown in Fig. 1, intracellular potential changes of a large-cell soma are composed of large and small synaptic potentials with spikes; furthermore, they include electrotonic potentials from other large cells. Consequently, the soma potential changes are rather complex. For this reason the impulse activity of a large cell is more correctly understood by impalement of the axon where the spike originates. Fig. 2 illustrates simultaneous recordings from the soma and axonal region in the large cell of Cell 3. The soma evoked repetitive large synaptic potentials which were preceded by small synaptic potentials (Fig. 2 A). A train of impulses was repeated at regular intervals in the impaled axon. The recording site of the axon was close to the soma (within 200 μm). The soma membrane is inexcitable, originating no proper spike potential, so that the soma spike is an attenuated potential induced by the axon-firing spike (Hagiwara, 1961 ; Tazaki, 1972a). Identification of the axon potential was, therefore, made from a large-sized spike showing overshoot. Soma spikes appeared in correspondence with axon spikes, indicating that they were postsynaptic impulses of Cell 3. In addition, this was confirmed by an experiment in which a brief current pulse of 0·1 msec duration was applied to the soma through an electrode impaled and elicited an orthodromic impulse in the axon (Fig. 2 C).

It was well shown in Fig. 2B that the impulse activity could be more accurately analysed in the axon trace than in the soma trace, where the number of impulses in the initial part of potential deflexions was obscure. The short high-frequency train of impulses was generated in the initial part of repetitive large synaptic potentials in which antifacilitation occurred (Hagiwara & Bullock, 1957; Bullock & Horridge, 1965). When the interval between spikes in the train became shorter, the following spikes decreased in amplitude. Repetitive large synaptic potentials failed to fire a corres-sponding series of spikes. The time courses of repetitive responses in Fig. 2B are shown in Fig. 3. It is presumed that the time course of repetitive large synaptic potentials represents that of repetitive impulses in the secondary pacemaker neurone. The frequency of spikes in the large cell decreased rapidly even while presynaptic impulses arrived almost continuously at a certain frequency. The result shows that the pattern of repetitive firing in large cells is rather phasic.

Apparently, the amplitude of synaptic potentials decreased much more in the axon than that in the soma (Fig. 2B). Thus, the electrotonic spread of synaptic potential took place from the soma to the axonal region. This finding makes it clear that synaptic regions are located far from the locus of spike initiation in the axon and at some distance from the soma. Probably presynaptic nerve fibres from small cells innervate the soma-dendritic regions of large cells.

The above-mentioned result was obtained when large cells were followers and controlled by two kinds of presynaptic nerve fibres of small neurones. Connor (1969) demonstrated that the burst activity of large cells in the sectioned preparation, from which the small-cell somata have been removed, is very similar to that of cells in the intact ganglion. Since the complex pattern of burst discharges took place in the intact ganglion of the present material, the posterior trunk was cut away to exclude the inputs from the small cells. The burst activity of large cells was examined in this sectioned preparation. In such cases the large follower cells came to have pacemaker activity, and their discharge pattern became different from that in the intact ganglion, being considerably simpler. An example is shown in Fig. 4. In the first stage of the experiment after the removal of the posterior trunk, periodic trains of spikes were repeated in Cell 1 (Fig. 4A). The train was preceded by the pacemaker slow depolarization arising continuously during an interburst period, and it was superimposed on the small amount of depolarization. In the later stage of the experiment spikes fired repeatedly on the larger depolarization of about 10 mV amplitude and 200 msec duration, which was called a slow potential (Fig. 4B). The frequency of spikes increased gradually to a peak and thereafter decreased (Fig. 4C). The pattern of repetitive firing differed greatly from that in Fig. 3. This difference is thought to be due to differences between pacemaker and follower activities of large cells. The slow potential was preceded by the pacemaker potential and followed by the after-hyperpolarization, indicating that this potential was the endogenous activity of the large cell and was responsible for repetitive firing. When the amplitude of the slow potential increased and the following after-hyperpolarization became more remarkable during the experiment, the frequency decreased. This fact means that the frequency is determined both by the rate of pacemaker slow depolarization and by the level of resting potential in the same way as in the action potential of the vertebrate heart (Weidmann, 1956). The slow potential of Cell 5 was relatively smaller in amplitude and longer in duration than that of Cell 1. It was also preceded by the pacemaker potential. Therefore, the slow potential seems to result from some intrinsic mechanisms of pacemaker potential in the large-cell membrane.

Four large cells among five, Cells 1–4, generate the same pattern of burst discharges. However, Cell 5 differed somewhat from the others in its discharge pattern. As shown in Fig. 4, a train of small deflexions was evoked, synchronized with the activity of Cell 1. Their amplitude and duration were about 1 mV and 20 msec respectively. There was no correlation between these deflexions and spikes of Cell 1 during spontaneous discharges, suggesting that the former were not synaptic potentials induced by the latter. Fig. 5 shows the result of current application to Cell 1. Spikes were repeated on the applied depolarization in the directly stimulated Cell 1, and small deflexions were similarly repeated in Cell 5. No close relation between activities of two cells was found however (Fig. 5 A). The number of spikes increased with increasing current intensity of the order of 10⁻⁸ A (Fig. 5 B). The slow potential occurred superimposed on the applied larger depolarization in Cell 1. Cell 5 also produced the slow potential. But it was not the electrotonic potential of the applied depolarization in Cell 1 because the electrotonic spread was very weak as shown in Fig. 5 C. This fact means that the slow potential is referred to intrinsic activity of Cell 5. Each of the large cells may generate the endogenous slow potential which leads to repetitive firing. The number of small deflexions increased with the increasing slow potential (Fig. 4). No evidence was obtained showing that Cell 5 received synaptic inputs from other large cells. These facts suggest that small deflexions are attenuated spikes. Probably the spike originates in the axon at some distance from the soma of Cell 5. Thus, the soma spike is of small amplitude and long duration.

Synchronization between the two large cells was strikingly maintained without common synaptic controls from the pacemaker neurones. The same was true when other large cells were impaled. Watanabe (1958) and Hagiwara et al. (1959) described that large cells are electrically interconnected, and that interaction mediated by slow potential changes is more effective for synchronization through this electrical connexion than that mediated by impulses. This agrees well with the foregoing.

Several kinds of impulses different from the postsynaptic impulses of large cells were propagated within the ganglion. The external recordings from the whole ganglionic trunk contained presynaptic and postsynaptic impulses, so that a single unit was externally recorded from the regions of small cells by means of microelectrodes. An example is shown in Fig. 6. The burst, which was composed of several large synaptic potentials superimposed on a number of small synaptic potentials, was repeated in the large cell at regular intervals (Fig. 6 A). A long-lasting train of impulses occurred at the same time as the burst (Fig. 6 B). The number of impulses in a train of about 1 sec duration ranged from 20 to 30. The large cell responded to them with repetitive small synaptic potentials. These synaptic potentials disappeared when the posterior trunk was removed. From this result it is clear that the externally recorded impulses originate in the small cell and that they are presynaptic impulses conducted to large cells. It has been shown that single presynaptic impulses activate all large cells, inducing small synaptic potentials; that this presynaptic nerve cell, which normally plays a role of primary pacemaker, has an acceleratory effect on the large cells; and that it also has a similar effect on the other presynaptic nerve cell of the secondary pacemaker (Tazaki, 1971 b).

The time course of repetitive discharges in this neurone is shown in Fig. 6 C. It was readily observed that the frequency increased gradually to a certain value and decreased slowly. The reliable evidence for the suggestion that the primary pacemaker neurone produces a long-lasting train of spikes (Tazaki, 1971b) was provided by the present experiment. The pattern of repetitive firing in the primary pacemaker neurone was different from that in the secondary, as seen in the graphs of Figs. 3 and 6C. In fact, the activity of the large cell in Fig. 6 B made it clear that the former was extensive, and the latter brief.

Fig. 7 illustrates intracellular spike potentials of the small-cell soma. We could not be sure whether the impaled neurone was Cell 6 or Cell 7 because the positions of small-cell somata were not established in this preparation. The amplitude of the spike was much smaller (approx. 15 mV), showing that the soma membrane of the small cell was also inexcitable. Each spike was preceded by a pacemaker slow depolarization and followed by an after-hyperpolarization. Spike intervals increased in succession. A brief period of repetitive spikes was caused by the relaxation oscillation, which was generally accepted to be characteristic of the pacemaker neurone in various nervous systems (Bullock & Horridge, 1965). This fact suggests that the small cell is a pacemaker which sends presynaptic nerve fibres to large cells. Unfortunately, successful simultaneous recordings from large and small cells could not be made. Accordingly, the relation between these impulses and activity of large cells is not yet exactly known.

A train of impulses having no close correlation with the activity of large cells was recorded in the regions of small cells. An example is given in Fig. 8 A, where external recordings were made in the posterior trunk. The activity of the large cell shows that two pacemaker neurones are active and that the repetitive firing pattern is the same as seen in Fig. 3. The train of impulses appeared almost simultaneously with the burst of the large cell. These impulses were neither presynaptic impulses from small cells nor postsynaptic impulses from large cells; none of potential deflexions corresponding to them was found in the activity of the large cell ; and they did not disappear when the anterior part of the trunk was cut away. Probably they originate from a small cell other than one of the pacemaker neurones. Repetitive firing pattern of this neurone resembled that of the pacemaker neurone such as shown in Fig. 6 though it had no synaptic relation with large cells.

In Fig. 8B the large cell showed a burst which was preceded by a slow depolarization arising gradually during an interburst period. The burst initiation was accelerated by the development of small synaptic potentials, which caused an increase in the rate of slow depolarization (Tazaki, 1971b). Spikes of about 12 mV in amplitude were recorded intracellularly from the small-cell soma of Cell 9. This result also supports the conclusion that the soma membrane of the small cell is inexcitable. The train of spikes was repeated, markedly synchronized with the development of repetitive small synaptic potentials in the large cell. These synaptic potentials were not, however, evoked in the impaled small cell, so that the pacemaker neurone did not innervate this cell. No functional correlation was found between impulse activities of two cells such as seen in Fig. 6; in fact, a second burst of the large cell was initiated, having no relation with the impulses of the small cell. In some preparations small deflexions, which resembled small synaptic potentials in their shape and size, appeared together in a large cell. That they might be also evoked by a small cell was suggested, but the result of Fig. 8 showed that small cells among four other non-pacemaker neurones had no synaptic interaction with large cells. Such small deflexions were abolished by a small amount of hyperpolarization applied to the cell soma. Probably they may be induced by impulses from other large cells.

It was found that the activities of the small cells arose at about the same time as that of the primary pacemaker neurone and were of almost the same duration. This fact suggests that the primary pacemaker controls activity of other small cells as well, with some interaction present among them. Normally, the burst discharges of the ganglion begin with the activity of the primary pacemaker neurone, spreading to five large and other small cells.

There were several kinds of impulses occurring synchronously in the component neurones of the lobster cardiac ganglion. Single-unit analysis supported the conclusion drawn concerning the Panulirus ganglion by many investigators that potential deflexions during the burst in the large-cell somata are a number of repetitive synaptic potentials and attenuated spikes. The burst activity in the sectioned preparation differed sharply from that in the intact ganglion (Fig. 4). In the synchronized ganglion the primary and secondary pacemaker neurones control large cells, producing the integrated pattern of burst with regular rhythm. In the ganglion which is asynchronous or in which spontaneity of the pacemakers fails, the burst activity becomes irregular or is maintained only by large cells. Potential changes during the burst discharges of large-cell somata in Homarus have been referred to the endogenous activity in these neurones (Connor, 1969). Such activity became evident in Panulirus when the presynaptic nerve cells did not show spontaneous activity. The slow oscillatory potential with superimposed spikes was found in large cells when the pacemaker neurones were anaesthetized by procaine-sea water (Watanabe, 1958). In the present study large cells generated the slow potential of their endogenous activity causing repetitive firing when the posterior trunk was cut away. In the cardiac ganglion of the crab Eriocheir the slow potential, which is an endogenous potential arising in the large-cell soma, is very effective in initiating a brief period of repetitive spike discharges (Tazaki, 1971a). Large cells are, however, also innervated by common presynaptic nerve fibres from small cells. The controlling mechanisms of the Eriocheir ganglion closely resemble those of the Panulirus ganglion (Tazaki, 1972b). In the cardiac ganglion of Squilla the pacemaker neurone produces the slow potential for initiation of repetitive spikes during spontaneous discharges, without any synaptic activation from other neurones (Watanabe, Obara & Akiyama, 1967). It is of interest that there are some similarities among these cardiac ganglia. From these facts it may be assumed that analysis of the endogenous slow potential is very important to the understanding of integration in autonomous nervous systems.

Impulse activity of large cells was apparently shown in the axon recording (Fig. 2). Impulse discharges of the axon in the cardiac ganglion have been reported in Squilla (Watanabe et al. 1967), Homarus (Connor, 1969), Eriocheir (Tazaki, 1970, 1971a) and Panulirus (Tazaki, 1972a). In all cases the soma potential changes are very different from those of the axon. Bullock & Terzuolo (1957) suggested that the locus of spike initiation may be separate from the loci of synaptic potentials. Further, Bullock & Horridge (1965) described, on the basis of the detailed anatomical description by Alexandrowicz (1932), how synapses may be located in the soma-dendritic regions of the large neurones. This interpretation was electrophysiologically demonstrated in the present study. The large-cell somata and dendrites are activated by common presynaptic nerve fibres from small cells and their axons initiate synchronous trains of impulses. They generate slow potentials with repetitive impulses by themselves when without synaptic activation from the small cells. The difference of repetitive firing pattern was seen between follower and pacemaker activities of large cells (Figs. 3, 4C). This agrees with the observation of Maynard (1955).

The activity of Cell 5 was characteristically different from that of other large cells (Fig. 4). The synaptic potentials were evoked like those in other cells, but frequently no spike of appreciable size was initiated. Bullock & Terzuolo (1957) found in Cell 5 a brief deflexion followed by a tendency to hyperpolarization, and they identified it as a spike. This was examined with the sectioned preparation in this experiment. In the stimulus-induced activity these deflexions were superimposed on the slowly increasing depolarization of slow potential which was intrinsic to the cell (Fig. 5). The present findings support the above-mentioned identification. The spike, which originates in the axon some distance from the soma, will be greatly attenuated in the soma, where it must be small in amplitude and long in duration.

A brief period of repetitive impulses was seen in the outputs of large cells. Maynard (1955) reported that large cells behave as motoneurones to the cardiac muscles, causing tetanus. It is therefore reasonable to think that the short high-frequency train of impulses such as shown in Figs. 2 and 4 is propagated to muscles and causes a rapid tetanic response.

It has been reported that two presynaptic nerve cells, which are different in the time courses of repetitive discharges, exist among small cells; one tends to produce a brief train of impulses, and the other an extensive train, inducing large and small synaptic potentials in large cells respectively (Tazaki, 1971b). A long-lasting train of presynaptic impulses from the primary pacemaker neurone was plainly demonstrated (Fig. 6). It is, not known, however, whether intracellular spikes in Fig. 7 are presynaptic impulses or not. The membrane of the impaled neurone exhibited relaxation oscillation which fired a brief train of impulses. The pattern of repetitive firing resembled that of repetitive large synaptic potentials. It is presumable, therefore, that this small cell may be the secondary pacemaker neurone. Intracellular spike potentials of the pacemaker neurone have been studied in the cardiac ganglia of Squilla (Watanabe et al. 1967) and of Limulus (Lang, 1971); they are generated, and preceded, by pacemaker potentials. Hartline (1967) reported that the pacemaker neurone in Homarus is either Cell 6 or Cell 7. In this paper the pacemaker neurone somata could not be identified because of the failure of impalement with a microelectrode though the possibility that they might be Cell 6 and Cell 7 was inferred.

The intracellular spike potential in the small-cell soma was very small in amplitude (Figs. 7, 8B), like that in the large-cell soma, indicating that the former also was not invaded by the spike-firing axon. This is in accordance with the observation of Hartline (1967). Two small cells among four are functionally identified; they are pacemaker neurones, their impulses inducing synaptic potentials in large cells (Tazaki, 1971b). In the present experiment there were impulses propagating from small cells, which had no functional relation with the activity of large cells (Fig. 8). Maynard (1955) found that the small cells do not cause a ranid tetanic contraction in cardiac muscles, and suggested that they may induce slow contraction or that they may innervate the ostia. But no reliable evidence has as yet been obtained. It is quite uncertain that the small cells other than pacemaker neurones have any functions.

Impulses in small cells were markedly synchronized with the activity of the pacemakers. Hartline (1967) observed that all small-cell activity arose simultaneously with the burst of large cells. The small cell of Fig. 8B did not receive synaptic inputs from the pacemaker neurone. From this reason it may be presumed that intercellular coordination by electrotrotonic coupling (Watanabe & Bullock, 1960; Tazaki, 1972a) is more effective for synchronization among small cells than synaptic interaction mediated by impulses. The electrotonic effect of slow potential change may occur among small cells in the same way that has been observed between large follower and small pacemaker neurones (Tazaki, 1971b). The burst activity of the ganglion began with that of the primary pacemaker neurone which had the longest duration. The normal ganglionic activity may be accomplished as follow: the small cell of the primary pacemaker controls not only large cells but also other small cells, and its activity determines the pattern of their burst discharges.

As described by Maynard (1966), synchronization is the primary functional aspect of integration in the cardiac ganglion. Synchronous activity among large cells was consistently present even when synaptic inputs from small cells were excluded (Fig. 4). This fact means that synaptic interaction mediated by impulses is effective for coordination between large follower and small pacemaker neurones. Electrotonic spread of slow potential change takes place among all nine neurones, and it is significant for stable synchronous discharges of the component neurones. The slow potential of individual large cells was too high to be explained only by electrotonic spread, indicating that it arose independently by intrinsic mechanisms (Figs. 4, 5). Accordingly, the electrotonic interaction must be important for synchronous development of the slow potential in large cells. Bullock & Horridge (1965) stated that the slow potential is generated in the soma-dendritic regions of the neurone. Recently evidence was provided that the electrical connexion arises between large-cell somata, but not between axons (Tazaki, 1972 a). Probably such connexion may be established among small cells. The electrotonic interaction has been discussed in more detail previously (Watanabe, 1958; Hagiwara et al. 1959; Watanabe & Bullock, 1960). Comparable observations have been reported in Squilla (Watanabe et al. 1967) and in Eriocheir (Tazaki, 1971a, 1972b). It is noteworthy that the activity of all nine ganglion cells is so well co-ordinated by both synaptic and electrotonic interactions.

1. Single-unit analysis was made by means of internal and external recordings in order to observe the impulse activity of the component neurones in the lobster cardiac ganglion.

2. The large cells fired a brief high-frequency train of postsynaptic impulses in the axonal region by repetitive synaptic activation from small cells which was brought about in the soma-dendritic regions. They generated slow potentials with repetitive impulses by themselves when without synaptic controls.

3. A long-lasting train of presynaptic impulses was propagated from the small pacemaker neurone to the large-cell somata, inducing small synaptic potentials. The burst activity of the ganglion was initiated by this neurone.

4. Impulses of different kinds, presynaptic or postsynaptic, were observed in small cells. This activity occurred at about the same time as that of the pacemaker neurone and was of almost the same duration.

5. Synchronizing mechanisms of all nine neurones were discussed with respect to electrotonic interaction mediated by slow potentials, compared to synaptic interaction mediated by impulses.