Progressive Decrements in the Activity of Aplysia Neurones Following Repeated Intracellular Stimulation: Implications for Habituation

Behaviourally, habituation can be defined as a centrally mediated, progressive and reversible decrement of a response to a constant, repeated stimulus. It is possible to demonstrate habituation in very simple forms such as single-cell organisms (Wood, 1970; Applewhite & Gardner, 1971), multicellular invertebrates (Bruner & Tauc, 1966a, b;Krasne & Woodsmall, 1969; Pinsker, Kupfermann, Castellucci & Kandel, 1970; Roberts, 1962) and in reduced neural systems of vertebrates (Buchwald, Halas & Schramm, 1965; Griffin & Pearson, 1967; Thompson & Spencer, 1966; Wickelgren, 1967a, b).

The physiological basics of habituation is not known, although in some cases a progressive decrement of the excitatory post-synaptic potential (EPSP) in neurones receiving repeated trans-synaptic stimulation shows a temporal course similar to behavioural habituation (Krasne, 1969; Kupfermann, Castellucci, Pinsker & Kandel, 1970). Partly because of the correlation between these EPSP decrements and behavioural habituation, there has been considerable interest in studying the phenomenon of synaptic depression following low-frequency stimulation, and several current reviews deal with synaptic depression in invertebrates and its hypothesized role in habituation (Bruner & Kehoe, 1970; Kandel, Castellucci, Pinsker & Kupfermann, 1970; Horn, 1970).

Some of the studies of synaptic depression in invertebrates have suggested that there is no change in the membrane properties of the post-synaptic cell during repeated trans-synaptic stimulation (Bruner & Kennedy, 1970; Bruner & Tauc, 1966 a, b;Castellucci, Pinsker, Kupfermann & Kandel, 1970). However, the possibility that small changes in post-synaptic neurone properties may contribute to the habituation phenomenon has never been critically studied. Because modulation of excitability in the post-synaptic cell per se could potentially make a significant contribution to phenomena of response decrements, it seemed important to examine this issue in more detail. Using the probability of action potential discharge to a repeated intracellular stimulus as a measure of excitability, the work reported here will show that neurones of Aplysia can exhibit the parametric characteristics of habituation independent of any synaptic modification.

The animals used were Aplysia california ranging in size from about 50 to over 500 g. The connectives and nerves of the abdominal ganglion were cut and the ganglion was removed and placed in a chamber filled with artificial sea water. The connective tissue surrounding the ganglion was pinned down into resin (Sylgard 182 Encapsulating resin, Dow Chemical) which lined the bottom of the chamber. Neurones were exposed by using very fine forceps and iridectomy scissors to dissect away the connective tissue above the cell bodies of interest.

Glass pipettes filled with 3 M‐KCl were used for intracellular recording and stimulation. The electrodes were pulled so as to have a resistance of about 8 megohms in sea water. For most experiments stimulation and recording were carried out with the same electrode using a WPI amplifier (W‐P Instruments, Inc.). An Ag‐AgCl wire electrode served as ground. Current was delivered to the WPI by a Grass S44 stimulator (Grass Instruments) through a Grass SIV5 stimulus-isolation unit. The preamplifier-stimulator circuit balance was checked repeatedly during each experiment by observing the initial rapid displacement of potential at the start of a stimulus. No adjustment of the balance was required during most experiments.

To ensure that a constant current was being delivered, several experiments were performed using separate electrodes for stimulating and recording. In these cases the recording electrode and ground electrode (a sea water-Ag‐AgCl agar bridge) were connected to high-impedance pre-amplifiers (Elsa‐2, Electronics for Life Sciences) and the difference in pre-amplifier outputs was displayed on the oscilloscope. The stimulus was delivered through a 20 megohm resistor to the stimulating electrode. During stimulation current was monitored across a resistor between a silver wire in the bath and ground.

The intracellular stimulus used to demonstrate response decrements consisted of a current pulse intensity and duration sufficient to initiate an action potential, but close to the firing threshold of the cell (about 15–20 mV more positive than resting). The current utilized ranged from 6 to 800 nA with pulse durations from 50 to 200 msec, depending upon the responsiveness of the particular cell and experimental procedure being utilized. The individual stimulus pulses were delivered at intervals which in different experimental sequences ranged from 2 to 10 sec.

In some experiments repeated extracellular stimulation was delivered to the right connective. The connective was placed across a bipolar silver-wire stimulating electrode positioned just above the surface of the sea water. The connective was stimulated at intervals which were the same as those used in the intracellular stimulation experiments, i.e. 2–10 sec.

Forty‐two abdominal ganglion cells were studied. Each was naturally silent, had a resting potential of at least 50 mV and an action potential of 90–110 mV. Nineteen of these cells were identified as the right dorsal giant cell (Frazier, Kandel, Kupfermann, Waziro & Coggeshall, 1967); the others were located on the surface of the dorsal side of the ganglion. All experiments were performed at approximately 23 °C.

The initial depolarizing stimulus was adjusted so as to produce one action potential per pulse. In all of the cells studied the probability of action potential discharge progressively decreased as the same stimulus was repeatedly presented. In a typical experiment at least eight to ten action potentials occurred during the first block of ten stimulations, and the number of action potentials in successive blocks always progressively decreased. A typical example of this response decrement is illustrated in Fig. 1. During the first block of ten stimuli each stimulus produced an action potential, while during the ninth block of ten stimuli (or the 81–90th trials) only three action potentials were produced. The decremental process took place while the current remained constant throughout the experimental tests; thus changes in the resistance at the tip of the electrode did not account for the effects.

In all cases the response decrement which developed during repeated intracellular stimulation was reversible if the cell was allowed to rest. With sufficient rest periods there was complete recovery and the ‘rehabituation’ curve had essentially the same form as the original ‘habituation’ curve (Fig. 1, 20 min rest). In general, recovery was complete only after 10–20 min.

In order to test the time course of recovery, multiple experiments of ‘habituation’ and ‘rehabituation’ were conducted. Fig. 1 shows the results of one experiment in which the time course of recovery was sampled at different periods of time after ‘habituation The cell was’ habituated tested for recovery by ‘rehabituation’ after a rest period of a particular length, and then the neurone was allowed to rest for 30 min. This rest period was followed, again, by a ‘habituation’ run with a ‘rehabituation’ test for recovery after a different time lapse from that previously used. The initial ‘habituation’ sequences produced response curves which essentially overlapped, indicating that the neurone had not deteriorated during the experiment. Fig. 1 indicates the amount of spontaneous recovery after different time periods.

At any point during the ‘habituation’ or recovery period a stimulus somewhat stronger than the ‘habituating’ stimulus could initiate action potentials. The decreased responsiveness was, therefore, not due to any total inactivation of the mechanisms responsible for the production of the action potential.

Parametric studies of habituation typically show greater habituation with short’ rather than long, interstimulus intervals (Buchwald & Humphrey, 1971 ; Thompson & Spencer, 1966). To determine whether this feature of habituation also resulted from intracellular stimulation a variety of interstimulus intervals was used, e.g. 2, 3, and 5 sec. The neurones clearly showed the greatest decrements when the shortest interstimulus intervals were used (Fig. 2).

Another parametric characteristic of habituation is the greater response decrement which develops during repeated presentations of a weak, rather than a strong, stimulus (Buchwald & Humphrey, 1971; Thompson & Spencer, 1966). Aplysia neurones subjected to repeated intracellular stimulation typically showed this inverse relationship between stimulus intensity and degree of ‘habituation’, i.e. the decrements appeared greatest with the weakest stimulation (Fig. 3).

In examining the decrement described above it is important to consider whether the injection of current per se through the intracellular electrode could cause the decreased responsiveness. A positive intracellular pulse, such as that used in the present experiments to elicit an action potential, might inject some potassium into the cell; if a significant amount were released, more than just the slow passive leakage of potassium chloride out of the electrode, this could conceivably affect the production of action potentials during repeated intracellular stimulation. Any increase in internal potassium concentration should increase outward potassium current, which in turn would raise the threshold for action potential production, since threshold is defined as the point at which inward current just equals outward current. However, the stimulus per se was apparently not causing the effects, as was demonstrated by the two control experiments described below.

If injection of potassium into the cell body with just-threshold stimuli was causing the decreased responsiveness, then a slightly subthreshold stimulus should have a similar effect. That this was not the case is shown in Fig. 4. Fifteen presentations of a subthreshold stimulus made no difference in the decrement of response probability to an ‘habituating’ stimulus. Thus, either (1) the production of action potentials is necessary to the decremental process, or (2) stimuli which are subthreshold for action potential production are also below the threshold for the decremental mechanism to be effective.

A second confirmation of the lack of any spurious effects of direct intracellular stimulation was provided by stimulating the giant cell antidromically through the right connective. The resultant action potentials decreased in firing probability during repeated extracellular stimulation of the axon just as they did with repeated intracellular stimulation (Fig. 5). The amplitude of all antidromic spikes recorded from the dorsal giant cell remained constant, which indicated that spike transmission along the axon was not blocked ; had such a block occurred, one would expect to record a greatly diminished spike in the soma (Tauc, 1962). Thus the spikes recorded in the soma probably were elicited by the spikes occurring in the axon. However, the decremental process could have originated in the axon at the site of stimulation, rather than in the neurone soma. These experiments indicate that the form of intracellular stimulation used did not cause ‘habituation’ if applied at just-subthreshold levels, and was not necessary to produce the phenomenon.

Three principal mechanisms might cause the ‘habituation’ observed with postsynaptic stimulation of Aplysia neurones: (1) the membrane potential might be hyperpolarized progressively, so that a larger stimulus was required to reach the threshold depolarization; (2) the membrane conductance might increase, producing a small depolarization for a given outward-current stimulus ; (3) the critical depolarization, or theshold, might rise to more positive levels.

The membrane resting potential was measured at the start and end of each block of ten stimuli. In all of the cells examined action potential decrements occurred without a measurable concurrent change in the resting potential. Sometimes a gradual hyperpolarization developed during the course of several hours of recording, but such drifts in the resting potential were independent of decremental and recovery processes.

Brodwick & Junge (1972) and Connor & Stevens (1971) have reported a long-lasting (order of seconds) increase in potassium conductance in molluscan nerve cells following depolarizations sufficient to produce action potentials. Such an increase in membrane conductance could account for the decremental phenomenon with very small changes in conductance, if the stimulus applied is close to the threshold level. In order to determine whether any change in membrane conductance occurred during the process of ‘habituation’, hyperpolarizing pulses were delivered immediately before and after each decremental sequence (Fig. 6). The current intensity was the same for each hyperpolarizing test pulse, so any increase in membrane conductance should appear as a decreased amount of hyperpolarization. With the accuracy afforded by this method no change in conductance could be seen as a result of the decremental process. However, even a very small increase in membrane conductance might cause a decrease in effectiveness of the stimulus sufficient to produce the observed decrements.

Alternatively, the decremental process could result from an accumulation of sodium inactivation (Hodgkin & Huxley, 1952; Geduldig & Gruener, 1970; Chandler & Meves, 1970) resulting in an increase in the threshold depolarization for action potential production. It is possible to obtain an indication as to which of these mechanisms is operating by imposing hyperpolarizing stimuli after the decremental process has developed and examining the effect on the subsequence response level. Brodwick & Junge (1972) reported that hyperpolarizing pulses had no effect on the post-stimulus conductance, while it is well known that hyperpolarization can remove sodium inactivation (Hodgkin & Huxley, 1952).

To determine whether a build-up of sodium inactivation might be the mechanism underlying the decrement in firing probability, a hyperpolarizing pulse was delivered intracellularly after a significant degree of ‘habituation’ had developed. The hyperpolarization extended for about 4 sec during the 5 sec interval which intervened between two depolarizing, ‘habituating’ stimuli. As demonstrated by Fig. 7, such intracellular hyperpolarization resulted in a partial reversal of the decreased action potential discharge to the depolarizing ‘habituation’ stimuli. These experiments suggest that sodium inactivation might be a primary cause of the decreased firing probability resulting from repeated stimulation.

The present research indicates that repeated intracellular stimulation of Aplysia neurones induces a decremental process which has the parametric characteristics of habituation. When Aplysia nerve cells were stimulated intracellularly close to their firing threshold the probability that an action potential would occur decreased with repeated stimulation. After a period of rest the neurones showed spontaneous recovery such that a subsequent series of depolarizing stimuli resulted in a decremental process with the same general time course as the initial series. Other parametric characteristics of habituation displayed in these experiments were the relationship of the strength of the stimulus and interstimulus interval to the amount of ‘habituation’. Thus, a stronger stimulus or a longer interstimulus interval produced less decrement than a weaker stimulus or a shorter interstimulus interval.

The fact that the excitability of an individual neurone can decrease in response to a constant, repeated stimulus indicates that the possibility of changes in excitability of neurones independently of synaptic modifications should be considered in any physiological analysis of habituation. It should be emphasized that the neuronal excitability changes demonstrated in this report have not as yet been related to behaviour. However, previous studies have shown that this decremental process is of significant importance during trans-synaptic depression (Stephens, 1972). ^*

The mechanism responsible for these changes is not obvious in the present experiments; there was no experimentally related change in the resting potential and the passive membrane response to stimulation was constant. This does not, however, completely rule out steady-state conductance changes. Since the stimulus is near threshold, a small effect, such as a light increase in potassium conductance, could decrease firing probability. Such a small change might not be detected by either a change in the resting potentials or in the conductance as measured by a hyperpolarizing pulse.

Previous studies (Bruner & Tauc, 1966a, b;Castellucci et al. 1970) of habituationlike phenomena in Aplysia neurones have typically measured the decrements of EPSP following repeated stimulation. These studies have focused upon synaptic alternations in an attempt to develop an intercellular model of habituation. Since an EPSP decrement could be caused either by conductance changes in the membrane of the post-synaptic cell or by changes taking place at the synaptic site, the response of the post-synaptic cell to a hyperpolarizing pulse before the start of stimulation was compared with the response to a hyperpolarizing pulse after the EPSP had decreased. Because there was no change in the response to the hyperpolarizing pulse, it was concluded that the EPSP decrement resulted from changes at the synaptic site. Although small changes in conductance might not be detected by a hyperpolarizing pulse, changes which were large enough to significantly affect EPSP amplitude would probably have been detected. Therefore, it is reasonable to conclude that the EPSP was probably decreasing because of changes at the synapse, e.g. transmitter release. However, the results of the experiments reported here demonstrate that the post-synaptic neurone could also undergo a decrease in excitability independent of synaptic modulation. These results suggest that, in considering possible mechanisms of habituation, it is important not only to look at alterations of the synaptic potential but also at changes of the membrane properties that develop only as a result of repeated stimulation and independent of any synaptic activity.

This work was supported by NS 05434 and GM 00448. It is a pleasure to thank Dr Jennifer Buchwald for her encouragement and support during all phases of this work. I would also like to thank Dr Douglas Junge for clarifying various theoretical aspects of neurophysiology and for his helpful comments on the manuscript. In addition, I wish to thank Dr Susumu Hagiwara for his critical reading of a preliminary draft of the manuscript and for his helpful suggestions.