Recovery Course of Excitability in A Single Neurone of Onchidium Verruculatum

It is well known that a neurone with all-or-none property acts as a threshold element in the nervous system and that the neurone produces nerve impulses as a result of its activity. Up to the present time we have had many excellent works by various authors since Hodgkin, Huxley and Katz et al. about the mechanism of impulse production, especially regarding the physico-chemical and electrical properties of a membrane of a nerve cell during spike discharge. However, almost all of these studies have been concerned with the physico-chemical aspects of a nerve cell and do not touch upon the aspect of information processing by the nerve cell. It has been generally accepted that, in the nervous system, information is carried by nerve impulses. Then, firstly, the informational mechanism of impulse initiation must be considered.

By some authors (Junge & Moore, 1966; Frazier, Kandel & Kupfermann, 1967; Chalazonitis, 1968), auto-excitability of a nerve cell had been reported in some molluscan giant nerve cells. According to them, a neurone has its own excitability and this excitability changes, if change occurs, by some intrinsic endogenous mechanism within the nerve cell. Especially when it oscillates rhythmically, it exceeds a critical level without any synaptic input, and discharges impulses regularly. We have found some nerve cells with auto-excitability in Onchidium verruculatum.

It is clear that the effect of inputs on the nerve cell may be modified by its own excitability, and moreover that the firing pattern may be affected considerably by this. Therefore we thought that it might be important to investigate the excitability of a nerve cell in order to make a study of the mechanism for the coding and decoding of neural information. Above all, the excitability after each spike was believed to be most important, because immediately after the spike, the excitability might change greatly, and this change of excitability obviously might affect the information processing by a single neurone.

In the present paper, using a marine pulmonate mollusc, Onchidium verruculatum, because of the large size of its nerve cell, we report the excitability (that is to say the electrical excitability) of nerve cells after the spike as well as during the resting state. These large nerve cells, so-called giant neurones, offered the advantages that we could identify them morphologically and functionally.

Marine pulmonate molluscs, Onchidium verruculatum, were used in our experiments. Their body weights varied between 12 and 17 g. They have some giant neurones (with a diameter of 100–400 μ) in the peri-oesophageal ganglionic ring. These large neurones were convenient for the insertion of one or two microelectrodes into their somata. We could easily record the behaviour of their intracellular potentials and stimulate them by transmembrane current.

Our investigations were made in the peri-oesophageal ganglionic ring isolated from the body of Onchidium verruculatum. The isolated ganglionic ring with the longest possible peripheral nerve fibres was immersed in sea water. The natural sea water which was filtrated and maintained at 15–17°C and at about 1·025 specific gravity, was circulated quietly (about 5 ml/min). The ganglionic ring was covered with thin, white, rough connective tissue, and when this tissue was removed, several giant neurones appeared. Some of these giant neurones were used in our experiments.

Usually we employed two intracellular microelectrodes. With one microelectrode the transmembrane potential of the neurone was recorded, and with the other this same neurone was stimulated by trans-membrane current pulses. However, with the Wheatstone bridge circuit we were able to record the membrane potential and stimulate the neurone simultaneously with a single microelectrode. The microelectrode which was used in the experiments was a glass microcapillary electrode filled with 2 M potassium-citrate, which had a resistance of 20–40 MΩ in sea water. An outline of the arrangement of the experimental instruments used is shown in Fig. 1.

One intracellular microelectrode for recording was connected with a d.c. amplifier with 40 dB gain, through a high input-impedance pre-amplifier with a field effect transistor (FET). The signal thus obtained was displayed on an oscilloscope (NIHON KOHDEN; VC-7) and recorded on magnetic tape by a data recorder (TEAC; R-400). To stimulate the neurone intracellularly, the other microelectrode was connected to a stimulator driven by a pulse generator, with which we could get arbitrary pulses whose intensities and durations were independently controlled. Stimulating transmembrane current used in our experiment was estimated from the electric current through a resistance (100 KΩ or 1 MΩ, shown as Ri in Fig. 1).

In order to avoid the drift of the recording system throughout a series of experiments, we used an underpolarized calomel electrode for the indifferent electrode and for the junction between capillary microelectrode and pre-amplifier. Room temperature was maintained at 15–18°C.

The transmembrane potential was differentiated by the unit labelled dV/dt in Fig. 1.

Once microelectrodes were inserted into one of the giant neurones the experiment continued according to our programme and lasted for 2 h duration for one giant neurone without any remarkable impairment.

The peri-oesophageal ganglionic ring consists of seven ganglia, as shown in Fig. 2; they are right and left cerebral ganglion, right and left pleuro-parietal ganglion, right and left pedal ganglion, and one visceral ganglion. The pedal ganglion cannot be seen from the dorsal side. We could distinguish about 30 giant neurones by their morphological features, their location, size and pigments, and by their electro-physiological characters, spontaneous firing patterns and input-output relations.

It seems that each neurone has a characteristic spontaneous firing pattern (Fig. 3). For example, some neurones seldom produced impulses, some fired very frequently and regularly, and some discharged apparently at random, as shown in Fig. 3, upper, middle and lower, respectively. It is suggested that each neurone may have some characteristic input-output relations.

There arose one question, whether or not the synaptic inputs wholly determined the output impulses. Here we conceived of the possibility that a neurone might excite itself without any synaptic input, as a result of the intrinsic excitability of the neurone, and that this excitability interacted with synaptic inputs. Both the innate excitability and synaptic inputs are probably individually characteristic of each neurone, and these interactions determine a characteristic output pattern.

Many studies which have been concerned with the excitability of the nervous system have reported the effects of stimulation by extracellular or intracellular electric current. One of these concerns the strength-duration relation or the strength-latency relation.

Our observations are here reported on the strength-latency relation, in the silent type of neurone which rarely discharges impulses. Individual stimulating current pulses used in this experiment had a width of 100–300 ms, which was long enough to initiate a spike during this current pulse, and stimulus was delivered in the resting state.

Fig. 4 shows some examples illustrating the behaviour of a neurone in response to the stimulus of transmembrane current pulse. It can easily be seen that, with a strong stimulus, spike discharges occur repeatedly throughout stimulation, and that interspike intervals are very brief near the beginning of stimulus but gradually become longer and longer. Here we noted the interval from the onset of stimulus to the initiation of the first spike. We termed this interval a latency, and measured it at various strengths of stimulus. As shown in Fig. 4, the stronger the stimulating current, the briefer the latency.

We plotted these relations between strength and latency (Fig. 5,a). In Fig. 5 (b) the abscissa is the reciprocal of the latency. From Fig. 5 (a) and (b) it is clear that a strength-latency relation is approximately represented by a hyperbolic curve. This suggested that, in the resting state, a minimum quantity of electricity (that is, the product of the strength and the duration of stimulus; Katz, 1966) is required to produce impulses. But when the strength was very weak, no spike could be discharged even with the greatest duration of stimulation. From this, we can estimate rheobasic electric currents (5 × 10⁻⁹ A) of Onchidium neurones in the resting state.

By stimulating the same silent type of neurone by transmembrane current pulses of various strengths and durations, we investigated whether they could produce nerve impulses following the cessation of the stimulating current pulse. The form of their responses is shown in Fig. 6. We also measured the relation between the strength and the duration of a stimulus which could initiate impulses (Fig. 7a, b). A curve thus obtained was a strength-duration curve of an Onchidium giant neurone for intracellular stimulation applied during the resting state. In Fig. 7 it is seen that electrical stimuli of various strengths and durations (belonging to the upper right-hand area of the curve) were effective in evoking nerve impulses.

The strength–duration curve was approximately hyperbolic (Fig. 7a, b), from which it is clear that a certain minimum quantity of electricity (that is, stimulus strength multiplied by its duration) is required for impulse initiation, because strength × duration = constant. A very weak stimulus could not evoke a nerve impulse, probably because of accommodation.

Considering both relations, that is, strength-latency relation and strength-duration relation mentioned above, the electrical excitability of a giant nerve cell of Onchidium verruculatum might be represented by the reciprocal of the product of the strength and the latency (or duration) required for impulse discharge. This value might be used as an indicator of the excitability of the nerve cell in the resting state.

A nerve cell was stimulated by long current pulses which began immediately after the peak of a conditioning impulse and lasted for 300 ms, as shown in Fig. 8. During the stimulus the next spike was initiated after a specific delay. This delay decreased with an increase of the strength of the stimulus. This relation between the strength of stimulus and the delay is illustrated in Fig. 9; and membrane potential changes are also shown (from spike initiation to resting state through after-hyperpolarization). Thus we measured a strength-latency relation immediately after spike discharge.

It was found from Fig. 9 that the strength of stimulus required to evoke a second impulse decreased exponentially with increase of the interval between the first and second spikes. And it was also clear that on semi-logarithmic coordinates this strengthlatency relation was represented by two straight lines (broken lines in Fig. 9). Moreover, it was noteworthy that this relation was not hyperbolic, but exponential. As mentioned previously, the strength-latency relation during the resting state was hyperbolic. This difference probably arose from the difference in the electrical excitability of the neurone between the resting state and the recovery state after the spike. Thus this strength-latency relation after the spike might show one of the aspects of the recovery process after spike discharge, this recovery process being described by two exponential terms.

However, since the stimulus began always at a fixed time; that is, at 5–10 ms after the peak of the impulse, and lasted for 300 ms as mentioned above, we could not observe the effect of the stimulus onset. Therefore, in studying the process of the recovery of excitability after a spike, we had to consider the time of onset of the current pulse as well as its duration. In the next section we shall consider stimulus onset.

In the previous section the strength-duration relation of silent-type neurones and the electrical excitability during the resting state was mentioned. In this section we shall take up the change of the electrical excitability after the spike; that is, the time course of the recovery of excitability, as follows. We measured the strength-duration relations at various time intervals after the conditioning spike. We selected a regularly-firing or frequently-firing neurone which was to be stimulated by a current pulse through a microelectrode inserted into its soma.

Stimuli were begun at various time intervals after a preceding conditioning spike whose rising phase triggered the sweep of a CRT. The time interval was measured from the trace on the CRT. Thus we stimulated a specified neurone with various strengths and durations after arbitrary intervals. Fig. 10 (a–f) shows the relations between the strength and duration of stimulus at different times after a preceding spike. After a long interval (1·2 s) the relation was approximately hyperbolic (Fig. 10 a); however, immediately after the spike (0·05 s) it differed slightly (Fig. 10f). In part this probably was caused by the side effects which long-lasting stimulating current pulses might introduce. Since the influence of a long pulse lasted even after the excitability had fully recovered, even a weak stimulus might have some effects. The strength latency relations with parameters of pulse durations are plotted on semilogarithmic co-ordinates in Fig. 11. These relations appear to be of the same form and to apply to all neurones which have been observed so far. Two representative samples of these relations are shown in Fig. 11 (a, b) and (bottom) membrane potential changes are also shown (from spike to after-hyperpolarization and gradually occuring depolarization). As is clear from Fig. 11, the results, both in (a) and in (b) are represented mathematically by two exponential terms with the change from one to the other at about 20–50 ms after the initiation of the preceding spike.

From the above relation the quantity of electricity was estimated as aforementioned, and plotted on semilogarithmic co-ordinates (Fig. 12). The value thus obtained represents the minimum quantity of electricity required to evoke a spike at various times after a conditioning spike and, as mentioned before, the reciprocal of this value might represent the electrical excitability. Both Fig. 12 (a) and (b) show that the change of this value with time after a preceding spike is described by two exponential terms. It is therefore concluded that the excitability of a nerve cell may be regained exponentially through two stages after a preceding spike.

These results were obtained for regularly firing or frequently firing neurones. However, it is assumed that this recovery process may be a process common to every nerve cell, because, in the case of silent-type neurones, the electrical excitability was regained exponentially even after an evoked spike discharge. This exponential recovery process is of interest, considering that the course of depolarization from after-hyperpolarization is expressed approximately by an exponential curve.

The membrane of a nerve cell is depolarized by excitatory synaptic input, and when this depolarization exceeds a certain level (threshold), nerve impulses are discharged. Hagiwara & Saito (1959) reported that an Onchidium neurone could readily initiate impulses during stimulation by depolarizing currents applied through an intracellular microelectrode. Therefore, transmembrane outward current that could depolarize the membrane potential could be experimentally regarded as the model of excitatory synaptic inputs which produce EPSP in a postsynaptic cell. In the present work it was confirmed that impulses could be evoked by transmembrane rectangular current pulses with an intracellular microelectrode, and an attempt was made to define the excitability of a neurone; in other words, the capability of the nerve cell to initiate impulses in response to electrical stimulation. It was considered that a neurone responding to a weak current was more excitable than a neurone for which a strong current was necessary. We employed a rectangular current pulse of varying strength and duration, and estimated the capability of the neurone to initiate an impulse in response to each pulse. In this way the concept of the excitability of the nerve cell is defined, at least as regards electrical excitability.

The strength-duration relation of stimuli applied to the nervous system has been reported by some authors, and is commonly used as an indicator of the excitability of active tissues; and it is known that this relation is expressed approximately by a hyperbolic curve (Katz, 1966). Frank & Fuortes (1956) and Coombs, Curtis & Eccles (1959) reported the strength-latency relation of intracellular stimuli in the spinal neurone of a cat. We have now confirmed this relation by stimuli applied through an intracellular microelectrode in the giant neurones of Onchidium. Specifically, during the resting state of a neurone, the strength-latency curve was approximately hyperbolic, and the strength-duration curve was also hyperbolic. This means that a constant quantity of electricity (intensity multiplied by duration of stimulating current pulse) is required for the initiation of an impulse, and that a minimum strength of current is needed, below which no nerve impulse can be initiated at any duration or latency. From this, we can also derive a rheobase and a chronaxie in the case of the giant neurone of Onchidium using stimuli applied through an intracellular electrode. Here arises the problem whether there is a difference between the strength-duration relation and the strength-latency relation and what this difference means, if anything. However, in the case of former relation, since the spike which was induced by the stimulus had to start promptly at the end of the stimulating current pulse, so both relations of one neurone might be approximately the same. The reciprocal of the quantity of electricity (intensity × duration) might represent the capability of a neurone for impulse initiation as mentioned above. Thus the strength and the duration of an electrical current pulse necessary for spike initiation had to be known in order to estimate the electrical excitability of a neurone. That is, if we know the relations between the strength and the duration.of a current pulse required for other spikes at other times; we can calculate the quantity of electricity, and therefore can know the course of change of excitability. Especially, this relation after the spike might disclose the recovery course of excitability after a discharge.

It has been well known that there is an absolute and a relative refractory period immediately after a spike. In the former period the excitability of a neurone is zero, and so a nerve impulse cannot be initiated. But after this period excitability is gradually restored. It was reported that in neurones of Aplysia californien, the threshold for spike initiation changed after the spike, and immediately after the spike a strong stimulus was required (Junge & Moore, 1966). Recently, Enger, Jansen & Walloe (1969) have shown the recovery course of the excitability of the stretch receptor of a crayfish. They used a short current pulse as stimulus, and measured the intensity of stimulus and the interval between conditioning and test firing, and obtained the recovery time course as a hyperbolic curve. However, detailed conditions of stimulation such as duration of stimulating current were not clear, and the beginning of stimulus was also not clear. Since we determined strength-duration curves at different times after a preceding spike, the onset and the duration of stimulus were clearly defined. On the basis of results thus obtained, the refractory period can be estimated quantitatively.

As mentioned before, the time course of the quantity of electricity indicates the course of the recovery process of electrical excitability after a spike. This is expressed by two intersecting straight lines on a semilogarithmic plot. The course of recovery is herefore divided into two processes which take place exponentially. Our experiments partly support the report by Enger et al. (1969) that the course of recovery is not well described by a single curve, but we cannot agree that it is hyperbolic. Finally, it is concluded that the excitability of a nerve cell after a spike recovers exponentially with time, and this recovery process takes place in two stages; also, that this process may occur automatically by a mechanism intrinsic within each nerve cell. The excitability, especially after a spike, would play an important role in eliciting nerve impulses; that is, in determing output pattern.

I wish to thank Professor R. Suzuki and Professor K. Murata for their advice and encouragement.