Isolated Crayfish Stretch Receptor Neuron Electrophysiology May Explain a Longstanding Mystery of Human Brain Functioning: Eureka Moment

Abstract

Neural network of our brain is complex, but single-neuron physiology is still important to understand the higher brain function. While conducting electrophysiological experiments using the isolated crayfish stretch receptor neuron, a phenomenon which may explain a longstanding mystery of human brain functioning, Eureka moment, was found. In this article, we demonstrate electro-physiologically GABAergic inhibitory synapses contribute for “switching” and propose a novel idea that can explain how sudden switching occurs in the brain.

1. Introduction

“It’s no secret that the brain’s complexity is vast, and although scientists have come closer to understanding how it functions, there is still a long way to go [1].” Historically speaking [2], observed detailed architecture of the brain cells with silver staining (Golgi’s method). Eccles [3] and Kuffler [4] illuminated the working of the nervous system on a large scale. Buzsáki [5] explained why the rhythms of the brain are important, and Stuart and Spruston [6] emphasized the critical role of dendrites in information processing in the brain.

Buzsáki and da Silva [7] mention that “high frequency oscillations constitute a novel trend in neurophysiology.” Thus, oscillations, including fluctuation, synchronization, and coupling, of assembled neurons are distinct physiological events occurring in the brain. However, there seems to be no satisfactory explanation for the science of the mind: “Eureka moment,” and “Rubin’s vase” or the “rabbit–duck illusion” which are bi-stable switching with ambiguous images or reversible figures. The switching is subjective, autonomic, or voluntary transformation occurring in adult brain.

We should go back again to the basics, observing a single-isolated neuron, asking whether “switching” occurs. Neuron is physical machinery composed of three parts: the dendrite, the cell body, and the axon. The commonality must be evolutionary consequence of convergent adaptation, from crustacean to human.

It is known that gamma-amino butyric acid (GABA) is the most important and dominant inhibitory neurotransmitter substance in the brain of human and crustacean too. However, its biochemical characteristics, inhibitory nature, and evidences for its synaptic release have been not established in human brain. But rather, crustacean neurons were used to show a significance of GABA molecule as the most important inhibitory neurotransmitter. Using crustacean nerve cells in 1950s, a German neurobiologist Florey, and Kuffler and his lab researchers; e.g., Otsuka and Kravicz [8, 9], contributed for identifying GABA as an inhibitory substance [10].

In mammalian brain, isolation of a single principal cell such as “pyramidal cell” is possible but not practical idea. In the brain slice preparation, it is inevitable that undetectable actions might affect neuron’s response when artificially stimulating target neuron(s) in the slice. Dissociated and isolated culture neuron are not adequate specimen because it is not matured naturally.

We would like to use the crustacean stretch receptor neuron, which is the mechano-sensory receptor neuron (RN) of the muscle receptor organ (MRO). The RN is one of the most thoroughly understood neurons studied by many researchers [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31]. Using this material as a model of the principal neuron (such as the pyramidal cells in the cortex), we examine whether “switching” is observable at a single-cell level. We consider that sudden change of electric current flows in the neurons might be occurring at “Eureka moment.”

It is known that, in mammalian brain, the principal neurons receive abundant inhibitory inputs from the interneurons such as Ramón y Cajal’s “basket” neurons. This inhibition is mediated by chloride ions [4]. MRO neurons (i.e., RNs) also receive GABAergic inhibition which is also mediated by chloride ions [4].

In the present study, we focus on the inhibitory transmission around the principal neurons. More precisely, we examine the role of GABAergic inhibition in the microcircuit. We will discuss the functional significance of GABAergic neurons for explaining neurophysiological mechanisms of “Eureka moment.”

2. Single-neuron experiments

2.1 Cell body matters

Figure 1 shows five examples for the loci of soma within the architecture of neuron. A shows the soma of RN in the MRO. This neuron extends its axon to the brain (abdominal ganglia). In the ganglia, the axon connects to the inhibitory interneuron (black neuron in Figure 1A). RN receives recurrent inhibitory influence (pink arrow in Figure 1A). B shows the soma of the principal neuron of vertebrate. Pyramidal cell is known to send axon collateral to the inhibitory interneuron (i.e., Cajal’s basket cell, black neuron in Figure 1B). C shows the spinal sensory cell. D and E show representative neurons of crustaceans, mollusks, and insects. There are distinct two types in the brain: one is A and B, the other is C, D, and E. The latter case (C, D, and E) is understandable, but the former case is not. The soma of A and B sit on the main route. It is seemingly an obstacle. It might be disturbing the information flow. The soma loci (A and B) might have undiscovered physico-physiological functions contributing for “neural computation” or “information-processing” in the highly cognitive brain.

Figure 1.

Diagrammatic representation of neurons demonstrating the soma loci. A, crustacean MRO neuron. B, vertebrate principal neuron. C, vertebrate sensory neuron in the spinal cord. D and E, neurons of invertebrates such as crustaceans, mollusks, and insects. Arrow heads, afferent inputs. Arrows, direction of electrical signal flow. Pink arrows, recurrent inhibitory microcircuit.

Intriguingly and surprisingly, the soma A and B are covered by highly packed synaptic knobs on surface (see Figures 1–4, page 7 of ref. [3]). Most of synapses on the soma are inhibitory [3]. What is the function of the soma? Is it not obstacle? What is the function of inhibitory synapses densely distributing on the somal and axonal membranes? We found answers in the present study. We found an important role of soma-axonal inhibitory synapses. It might be working everywhere across our brain.

2.2 Cell isolation for electrophysiological experiments

2.2.1 Nerve dissection: Microscopic technique

Figure 2 shows dissection procedure of abdominal MRO in the present study. Figure 2A shows the locus of MRO, at the fourth segment (blue), Figure 2B shows MRO from rostral segments, and Figure 2C shows MRO from caudal segments. Note that caudal MROs are stick together. Figure 2D–F illustrates dissection procedure, separating two MROs off using a forceps and a fine needle. In Figure 2E, RNs may not seriously be injured at this state. But in Figure 2F, desheathing manipulation of the soma-axon portion of RNs (red) is conducted although it is venturous procedure. As a consequence, caudal MRO is suitable for obtaining a successful desheathed preparation.

Figure 2.

Dissection procedure of abdominal MRO used in the present study. A, crayfish abdomen. B and C, isolated MROs from the abdomen. D–F, further dissection procedures after isolation.

Figure 3 shows light-microscopic view of MRO in the present study. The photographs were taken by a handy camera through an eyepiece. Left picture in Figure 3 corresponds to the diagrammatic representation shown in Figure 2C. Arrows point axons of two RNs. F denotes one of the clamp forceps clamping at the cut ends of the receptor muscles. Right picture in Figure 3 corresponds to the diagrammatic representation shown in Figure 2F. Axon diameter is less than 10 μm.

Figure 3.

Light-microscopic view of MRO. Left, an isolated MRO-RNs. Two axons of the slowly adapting and fast adapting RNs are pointed by arrows. Right, the slowly adapting neuron alone is isolated. One can see that an axon is exposed.

2.2.2 Electrophysiology: Methods

The fast and slowly adapting stretch receptor organs with their receptor neurons, muscles, and efferent nerves attached were dissected out from the abdomen (Figure 2). Fundamental electrophysiological methods are previously mentioned [32]. Briefly, A perfusion chamber contains 0.6 ml of van Harreveld solution [33], flowing speed 0.1 ml/s, flowing direction from dendrite to axon (see Figure 4 inset), at room temperature, 18–22°C. Glass microelectrode is filled with 4 M potassium acetate, for recording and passing current through the membrane. For the iontophoretic application of GABA, a double-barreled electrode (see Figure 4 inset) system was applied, one barrel filled with 1 M GABA, pH adjusted to 4.0 with HCl, and the other filled with 1 M NaCl. GABA was ejected as cations when a positive pulse was applied. For the prevention of possible leakage of GABA, negative current of 5 nA was continuously applied to the GABA electrode. The ejection current was superimposed on this breaking current. Stimulation of nerve was done by a suction electrode. The membrane potential and the iontophoretic current for GABA application were displayed on an oscilloscope and photographed. Facilities at the laboratory of Kazuo Ikeda (deceased 2016), Neuroscience, City Of Hope, Duarte, CA, USA, were used in 1985–1986. Crayfish Procambarus clarkii specimens were captured locally at Ontario, CA, USA.

Figure 4.

GABA responses recorded intracellularly in the present study. Inset shows diagrammatically our experimental methods in general for the present article. GABA was applied to a spot on the soma iontophoretically by a 30 nA, 50 ms pulse. Uppermost trace shows the GABA current for A, B, and C. In A, the tip of GABA, electrode rested just at the surface of soma. In B, the GABA electrode was pulled back by 10 μm, and in C, by 20 μm.

2.2.3 Intracellular recording: Technical difficulty

We tried to reproduce recordings appeared in Kuffler’s reports [15, 16, 21, 22, 23]. As shown in Figure 5, we successfully reproduced action potentials (APs) which were recorded from the soma membrane (Figure 5B). Inhibitory postsynaptic potentials (IPSPs) were also reproducible (Figure 5B). However, success rate is not high. Penetration of glass microelectrodes into the soma was an obstacle. For example, we obtained valuable data from only three neurons out of 37 neurons, during a two-week five-day experiment, for example (see Table 1). Table 1 summarizes what neuron we used and what end result we obtained, hit, or miss.

Figure 5.

We confirmed action potentials and inhibitory postsynaptic potentials in our preparation in the present study. Unpublished data.

Wed 27Aug

Thu 28Aug

Fri 29Aug

Somite

Side

Cell

End

Somite

Side

Cell

End

Somite

Side

Cell

End

1

4th

Right

Fast

Miss

4th

Right

Fast

Miss

2nd

Right

Fast

Miss

2

4th

Left

Fast

Miss

4th

Left

Fast

Hit

2nd

Right

Fast

Miss

3

3rd

Right

Fast

Miss

4th

Left

Fast

Miss

2nd

Left

Slow

Miss

4

3rd

Right

Slow

Miss

2nd

Left

Fast

Miss

5

3rd

Left

Fast

Miss

1st

Left

Fast

Miss

6

2nd

Right

Fast

Miss

1st

Left

Slow

Miss

7

2nd

Left

Fast

Miss

3rd

Left

Fast

Miss

8

3rd

Right

Slow

Miss

9

3rd

Right

Fast

Miss

10

3rd

Right

Slow

Hit

Tue 2Sept

Wed 3Sept

Thu 4Sept

Fri 5Sept

Somite

Side

Cell

End

Somite

Side

Cell

End

Somite

Side

Cell

End

Somite

Side

Cell

End

1

2nd

Right

Fast

Miss

2nd

Right

Fast

Miss

2nd

Right

Fast

Miss

2nd

Right

Fast

Miss

2

2nd

Left

Fast

Miss

2nd

Right

Slow

Miss

2nd

Right

Slow

Miss

2nd

Right

Slow

Miss

3

3rd

Right

Fast

Miss

2nd

Left

Slow

Miss

3rd

Left

Slow

Miss

3rd

Left

Slow

Hit

4

3rd

Right

Slow

Miss

3rd

Right

Slow

Miss

2nd

Left

Slow

Miss

5

3rd

Left

Slow

Miss

Table 1.

Low success rate: A two-week five-day experiment.

Figure 5A shows diagrammatical illustration of stimulation and recording methods in the present study. A glass microelectrode (Rec) is penetrated to the soma of neuron to observe membrane potential of slowly adapting neuron. S-Ax (arrow-1) denotes the point where antidromic stimulation of axon of slowly adapting receptor neuron (shown in blue-purple) was applied. Electrical pulse 0.05 ms duration was used at just above threshold intensity. S-Ax (arrow-2) denotes the point where stimulation of the inhibitory nerve (black) was applied. This inhibitory stimulation was done following the “axon reflex method” described by Kuffler and Eyzaguirre [23]. Using the just above threshold intensity is important, for preventing that excessive electric currents could stimulate other unrelated neuronal elements in the tissue.

Figure 5B shows antidromically initiated APs recorded at soma evoked by stimulation at S-Ax (arrow-1 in Figure 5A), and IPSPs evoked by stimulation at S-Ax (arrow-2 in Figure 5A). Time difference between arrow-1 and arrow-2 is fixed, approximately 20 ms, and several oscilloscope sweeps are superimposed.

Figure 5C shows superimposed weeps of APs and IPSPs. The time of antidromic stimulation was gradually shifted but IPSP stimulation time was fixed. IPSP blocks invasion of axonal spikes (failure of initiation of somal active spikes), but passive axonal spike is seen. B and C are different experiments with different cells thus, our experiments showing AP and IPSP are reproducible in the present study.

3. GABAergic inhibition

3.1 GABA iontophoretic application

Dendrites, soma, and axon areas receive a number of inhibitory synapse contacts, in both the MRO and the vertebrate principal neurons [3, 4, 5, 6, 18, 19, 20]. This issue has not been well examined by physiological methods, because vertebrates, including human, neurons are impossible to isolate at a healthy-normal-operating state. Therefore, we considered that crustacean MRO neurons are beneficial. We examined synaptic function by iontophoretically applying small dose of GABA to localized area of the dendrites, soma, or axon (Figures 4 and 6). To do this experiment properly, it is imperative that the applied GABA is localized to a small spot so that there is no chance that other area will be stimulated by the application. Thus, the smallest detectable dose was determined, and the size of the affected area was estimated (Figure 4).

Figure 6.

Response to GABA applied locally on the various spots of the dendrite (right), and soma and axon (left). A 27 nA, 50 ms pulse was used for all GABA applications (uppermost trace). Application site was indicated in the diagram.

In Figure 4, we found that a single 30 nA current pulse of 50 ms duration produces a 1 mV response when the GABA electrode is located just on the surface of the somal membrane of the slowly adapting receptor. When the electrode was pulled back from the membrane, the latency of the response increased (see A and B in Figure 4) and the amplitude of the response diminished until it had disappeared completely with a 20 μm withdrawal from the membrane. This suggests that the GABA ejected by this current pulse is probably effective only in an area about 20 μm from the electrode.

When GABA current was increased, larger response with a similar time course was obtained (not shown). For the maximum GABA current employed in this experiment, the distance from the neuron surface to the tip of GABA electrode was 50 ± 10 μm to make the GABA response that disappears. Thus, even when the strongest ejection current (50 nA) was applied, synapses located more than 50 μm from the tip of the electrode were unaffected by the applied GABAs in this experiment.

It is known that somal and axonal areas receive synapse contacts [18, 19, 20]. In Figure 6, we conducted the experiment surveying the GABA-sensitive area on the somal and axonal surfaces of the slowly adapting receptor neuron (Figure 6 left panel). Hyperpolarizing responses can be seen when a small dose of GABA was iontophoretically applied to the soma and axon. The numbers in a diagram denote the distance from axon hillock. The experiment was repeated with 14 different preparations. Somal and axonal responses were obtained in every case.

A similar experiment was also done with an undesheathed preparation (likewise Figure 2E). When GABA was applied from the outside of the axon sheath, the response was almost null, even with the highest dose (50 nA, 1 sec).

With the same technique, the responses to GABA applied to the dendritic area were investigated with undesheathed preparations in the slowly adapting receptor neuron (Figure 6 right panel). The amplitudes of the responses in the dendritic area were variable, depending on the location of the GABA electrode. The variability may well be related to the structure of the dendrite, but the fine structure of the dendrite in this recording condition was not clearly observable under a low-magnification microscope.

The electric currents using GABA application might cause direct effects on neuronal membrane. Thus, the possibility that the observed response was caused by the ejecting current, but not by GABA, was tested using a double-barreled electrode (see inset Figure 4) in which one barrel was filled with 1 M GABA and the other with 1 M NaCl (Figure 7). In this experiment, GABA was first applied to the dendrite, and then, the same current was passed through the NaCl barrel. As can be seen in Figure 7A and B, the neuron responded only to GABA.

Figure 7.

A, response to GABA and B, that to NaCl applied to the same site in the dendritic area of the slowly adapting receptor neuron of the 3rd abdominal segment using double-barreled electrode (see inset in Figure 4). The same 12.5 nA iontophoretic current of 50 ms duration was used. C, D, response to GABA in a saline containing 3 mM CoCl₂.C, soma and D, axon (70 μm away from axon hillock).

It is known that neurotransmitter substances affect both pre- and postsynaptic terminals [4, 25]. There is a possibility that the response could be mediated through a presynaptic terminal. If the iontophoretically applied GABA was affecting inhibitory terminals, which in turn released transmitter to the receptor neurons, an inhibitory potential would result. This possibility was tested by applying GABA in the presence of 3 mM CoCl₂ (see a standard text book, such as ref. [4]). The receptor neuron is responded to iontophoretically applied GABA in a 3 mM CoCl₂-containing solution in a similar manner to that in the standard van Harreveld solution (Figure 7C and D). It was, thus, confirmed that the applied GABA directly affects the postsynaptic membrane of the receptor neuron.

3.2 IPSP at the axon

The above results show that the GABA receptor exists not only on the dendrites, but also on the soma and the axon. Further evidence can be obtained by the focal recording of extracellular IPSPs upon stimulation of the inhibitory neuron. As shown in Figure 8, the local IPSP was recorded from the axon of slowly adapting receptor neuron at a portion of the axon 90–180 μm away from the soma. In the record (Figure 8B), the impulse of the inhibitory axon preceding the IPSP can be seen.

Figure 8.

IPSPs recorded extracellularly from a site on the axon between 90 μm and 180 μm from the soma of the slowly adapting receptor neuron of the 3rd abdominal segment. Inset shows diagrammatically experimental method (see also Figure 5 inset for the method of inhibitory stimulation). In both A and B, two CRO traces were superimposed. In the record with faster sweep speed (in B), an impulse of the inhibitory axon preceding the IPSP is seen (B is 2.5 times faster than A).

4. Function of the somal and axonal inhibitions

The stretch receptor neurons in crustaceans are one of the most extensively studied neurons [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31]. The input signal, i.e., sensory-muscle extension, stimulates the dendrites and induces the generator potential at the dendritic membrane [4]. The depolarizing current conducts to the axon causing the initiation of the action potential [4]. The greater of the generator potential, the receptor neuron produces the higher frequency of discharge of action potential [4]. The inhibitory neuron can reduce the amount of the generator depolarization. This is the long-established understanding of neuronal function [4].

However, the above results in the present study revealed that crayfish stretch receptor neuron is innervated by the inhibitory neuron not only on its dendrite but also on its “soma” and part of its “axon.” The presence of GABA-mediated inhibitory synapse is conclusively proven in the present study.

In our brain, the principal neuron (e.g., the pyramidal cell in the cerebral cortex) is innervated by the inhibitory neuron [3]. The most notable inhibitory interneurons that innervate the principal neuron are two cells: the basket cell and axo-axonic (or chandelier) cell [3, 4, 5]. The former innervates on its soma, and the latter innervates on its axon (i.e., initial segment or axon hillock). It is believed that the principal neuron plays important roles in advanced cognitive functions [6, 7].

Questions arise. What is the function of these inhibitory synapses distributing on the somal and axonal membranes? Do they work to attenuate the excitatory input? The functional role of the inhibitory synapses on the axon, axon hillock, or initial segment, should be reexamined.

4.1 Novel inhibitory function

4.1.1 Blockade of antidromic invasion of impulses into the soma by GABA applied to the axon

Axonal GABA receptors are demonstrated in Figure 6 (see also [14]). However, the functional significance of axonal GABAergic synapse is not demonstrated physiologically [4, 5]. It is understandable that dendritic and somal inhibitory GABAergic synapses can inhibit neuronal activity. But, what is the functional significance of axonal GABAergic synapses? Why the axon receives GABAergic synapses?

We investigated the function of inhibitory synapses on the axon by observing nerve impulses with an intracellular electrode inserted into the soma (Figure 9). The cut end of the axon was stimulated while applying GABA iontophoretically onto a part of the axon 40 μm away from the soma. In the example shown in Figure 9, while the axon was being stimulated with a frequency of 2 Hz, GABA was being applied with a single ejecting current of 50 nA and 500 ms, a portion of the axon 40 μm away from the soma. Impulses designated A to F were recorded successively. GABA was applied for 50 msec beginning slightly before record B and ending prior to record C. In C and D, the somal regenerative impulses are blocked, leaving the axonal impulses, which electrically spread into the soma. The GABA effect lasted for about 1 sec. Thus, it is demonstrated that the invasion of the impulse into the soma was blocked, so that only the axonal impulse which is occurring prior to the point of inhibition by GABA is observed. We also observed that the blockage of the somal invasion required a large dose of GABA when applied to a single spot on the axon. In comparison to the dose of GABA which causes a detectable membrane hyperpolarization or depolarization (see Figure 4), the dose necessary for the blockage was more than 10 times higher. However, it can be assumed that this dose does not mean that the conduction block requires a higher dose of GABA to induce a larger conductance increase at a spot on the axon but rather indicates that a larger area of the axon must be inhibited to block an impulse, i.e., the effectiveness is related to the spatial coverage rather than the concentration at a particular spot. Blockage of the conducting impulse may require a certain length of axon to be inactivated (i.e., shunt); otherwise, the impulse may jump the inactivated area. This possibility was tested below (see Figure 10).

Figure 9.

Blockade of antidromic invasion of impulses into the soma by iontophoretic application of GABA onto the axon. Inset shows diagrammatically present experimental method. Antidromic impulses were elicited by stimulating (0.05 ms duration square pulse, blue-color suction electrode) the cut end of the axon of the slowly adapting stretch receptor neuron of the 3rd abdominal segment: Every 500 ms. GABA was applied with a pulse of 50 nA and 500 ms at a portion of the axon 40 nm away from the soma at the time indicated in the figure. Impulses were recorded successively on moving film. A shows the control before application of GABA. The GABA pulse was applied just before the impulse in B, and ended before C. One can see that somal invasion was blocked in C and D and recovered in E and F. In E, the conductance increase caused by GABA is seen as a faster time course of the impulse.

Figure 10.

Spatial effect of inhibition on the blockage of antidromic impulses. Inset diagrammatically presents experimental method. A: 9 nA, 500 ms pulse at 0 μm, No blockage. B: 9 nA, 500 ms pulse at 60 μm, No blockage. C: 9 nA, 500 ms simultaneous pulses at 0, 60 μm. Three impulses failed to invade. D: 12 nA, 500 ms simultaneous pulses at 0, 60 μm, five impulses failed to invade.

In Figure 10, antidromic impulses were elicited with a frequency of 3 Hz by stimulating the axon of a slowly adapting receptor neuron of the 3rd abdominal segment. GABA was applied with two electrodes, one placed at the axon hillock (0 μm), the other 60 μm down the axon (see inset of Figure 10). Two glass capillary microelectrodes of the same resistance when filled with 1 M GABA were selected. Two electrodes in parallel were connected to a 1000 M ohm resister through which the ejecting current was applied.

Figure 10 shows that single-GABA electrode did not block antidromic spikes (see Figure 10A and B). But, two GABA electrodes, 60 μm apart each other, blocked antidromic spikes (see Figure 10C and D). Experiments shown in Figure 10 demonstrate that blockage of the conducting impulse requires a certain length of axon to be inactivated (i.e., shunt); otherwise, the impulse may jump the inactivated area.

4.1.2 Effect of invading impulses on the stretch-induced impulse

Experiments shown in Figure 10 demonstrated that antidromic-impulse invasion can be blocked when axonal GABAergic synapses are activated. Action potentials of neuron arise generally at the axon initial segment of neuron [4, 6]. Then, the action potentials travel toward two directions: one for orthodromic direction and the other for antidromic direction [4, 6]. It is obvious that orthodromic impulses convey neural information to the next cells. But functional significance of antidromic impulses is still controversial (see ref. [6, 34]).

In order to test the possible interaction between invading antidromic impulses and stretch-induced impulses, antidromic stimulation was given while the slowly adapting receptor neuron was stretched and held at constant tension. In Figure 11A, the impulse above the horizontal bar is antidromic impulses invading the soma. The stretch-induced impulses are seen to be accompanied by the generator potentials. As can be seen here, the frequency of the stretch-induced impulses becomes reduced for a certain period after the antidromic stimulation. In Figure 11B, the somal invasion was blocked for the entire period. Thus, all of the impulses seen here are axonal impulses recorded at the soma electrically. As can be seen, the antidromic stimulation failed to affect the frequency of the stretch-induced impulses. This set of experiments demonstrates that when the somal-dendritic area is invaded by regenerative impulses, the frequency of the stretch-induced impulses is reduced. However, if the invasion is blocked, the frequency remains unchanged.

Figure 11.

Effect of antidromic stimulation on the stretch-induced impulses. A. with antidromic invasion. While a slowly adapting receptor neuron of the 3rd abdominal ganglion is firing with constant frequency with somal impulses unblocked, repetitive impulses caused by antidromic stimulation invaded the soma (12 impulses at 7 Hz above the bar). Note the frequency of stretch-induced impulses decreased after the antidromic stimulation, even though the tension was held constant. Time scale of 1 sec applies to A. B. without antidromic invasion. After blocking invasion, the effect was observed similarly. Note the frequency of stretch-induced impulses is unchanged before and after the antidromic stimulation (5 impulses above the bar). Time scale of 0.1 sec applies to B.

4.1.3 Effect of antidromic invasion of stretch-induced impulses on the stretch-induced frequency

When orthodromic impulses are set up by stretch, the excitation occurring at the spike initiation site also sends impulses back to the soma [4, 15, 22, 23]. This antidromic invasion must have an effect on the stretch-induced frequency. In other words, unless antidromic invasion is blocked, the output frequency will be lower than the frequency dictated by the generator potential, which is directly coupled with tension.

In order to demonstrate the effect of these returning impulses (antidromic invasion) on their own generator potential, constant tension was given to a slowly adapting receptor to elicit orthodromic impulses of constant frequency, and then GABA was applied to the axon in order to block antidromic invasion.

As can be seen in Figure 12, while the receptor neuron was firing at a constant frequency, GABA was applied to a spot on the axon 180 μm away from the soma. Antidromic invasion was blocked during the effective period indicated by the horizontal bar (16 axonal spikes without active somal membrane activities). It is clearly demonstrated that the frequency of firing during this period increased and stayed at a higher constant level. The frequency is returned to the original level after the cessation of the GABA action (Figure 12).

Figure 12.

Effect of blocking antidromic invasion of the stretch-induced impulses on stretch-induced frequency. While a slowly adapting receptor neuron of the 3rd abdominal segment was firing at constant frequency with its somal impulses unblocked, GABA was applied with a 50 nA, 2.5 sec pulse (top trace). During the period of GABA action (above the bar, colored yellow), antidromic invasion was blocked. Note the increased frequency of orthodromic impulses even though the tension was held constant. After GABA action, the frequency returned to the original level. Inset (right), a photograph from different specimens with a slower sweep speed. The receptor muscle was kept at a constant tension. Note constant frequency of impulse firing. A brief (30 ms duration, strong current) GABA application to the axon induced the blockage of the antidromic impulses. During the blockage period, it is clearly seen that hyperpolarizing after potential, which is mediated by potassium ions, is disappeared due to disappearance of the somal spikes.

This result indicates that when antidromic invasion is allowed, the output frequency is lowered from the level originally set by the fixed tension. Only when antidromic invasion is blocked, does the output frequency precisely reflect the level of tension.

A question arises. How much difference is there between the precise level and the lowered (distorted) level? The frequencies of “higher constant level” and “original constant level” are measurable from the oscilloscope-photographed data (9 cells). Seventeen measurements show following results, but it seems that data should not be statistically taken account because technical condition (dissection technique, electrode penetration technique, GABA electrode variation, etc.) varies.

Even though it is measurable in the present studies (all unpublished results). In Figure 12, high-level spiking discharge was at 2.47 Hz and low level at 1.64 Hz; thus, the invasion lowered the true information down to a 60 percent level, which is “false-distorted” information but meaningful information (see discussion). The most lowered case was 39 percent. The next significantly lowered case was 48 percent (3.33 Hz vs. 1.61 Hz). In Figure 12 inset (right), the antidromic invasion lowers the information quality to the accuracy of 66 percent.

All other data obtained from the oscilloscope photographs are as follows: In the present studies, we obtained following calculated numbers: 60, 70, 71, 71, 59, 68, 57, 64, 57, 68, 61, 70, 77, and 78(in percent). In conclusion, the blockage can push the output frequency up to approximately 1.5 times.

5. Discussion

5.1 Crustacean stretch receptor cell

When the stretch receptor is uninhibited, antidromic impulse invades the soma. The present results demonstrate that the invasion is blocked when a part of axon is inhibited by the iontophoretic application of GABA. The effect of inhibitory postsynaptic potentials on the antidromically invading impulse was studied by Kuffler and Eyzaguire [23] by stimulating the inhibitory nerve while observing the response at the soma. They reported only a 2 percent decrease in the amplitude of the impulse and no blockage (see Figure 14 of ref. [23]). The inhibitory nerve synapses are not only on the soma and the axon, but also on the dendrites. Thus, stimulating the inhibitory nerve cannot differentiate the function of synapses localized on a particular part of the neuron. This is the reason why iontophoretic application of GABA was used in the present study in place of inhibitory nerve stimulation.

The failure to block the invasion of impulse to the soma in the experiment reported by Kuffler and Eyzaguirre [23] is understandable by looking at the record shown in that paper. The soma is depolarized by a large IPSP (see Figure 14 of ref. [23]), thus, providing a facilitative effect on the invading impulse. The effect of increased conductance must be overcome by this depolarizing potential change. When GABA was applied to a spot on the axon, the dose required for the blockage was as high as the equivalent of 2.5 × 10⁻⁸ coulombs in the example shown in Figure 4. When GABA was applied with two electrodes 60 μm apart along the axon, the necessary dose for blockage was lowered to the equivalent of 4.5 × 10⁻⁹ coulombs. In this case, the applied GABA had to diffuse to cover many synapses along the axon. It seems reasonable that the dose required for blockage by applying GABA at one spot was about five times higher than that required for the application with two electrodes separated from each other by a reasonable distance to effectively cover a certain length of axon. The length of axon necessary to effectively block the conducting impulse was, thus, estimated to be about 100 μm. In Figure 4, it was demonstrated that the dose of iontophoretically applied GABA on the axon necessary to cause a hyperpolarizing response of 1 mV was about 7.5 × 10⁻¹⁰ coulombs. If the transport number is assumed to be 0.5, this would correspond to 3.7 × 10⁻¹⁵ moles of GABA. In the neuromuscular junction of the same species of crayfish, Takeuchi and Takeuchi [35] reported that iontophoretic application of 4.2 × 10⁻¹⁵ moles of GABA elicited a depolarizing response of 0.5 mV. Of course, the polarity and the amplitude of the GABA response is dependent on the membrane potential relative to the equilibrium potential of the carrier ion, and deviation of these potentials from preparation to preparation but suggest that the sensitivity of axonal GABA receptors is similar to that of the muscle.

From the present experiment, the estimation of the axon length to be kept under inhibition in order to effectively block the conducting impulses was about 100 μm. The present experiment (Figure 6) showed that a part of the axon about 200–300 μm from the soma was found covered with inhibitory synapses and sensitive to GABA. It indicates that the distribution of inhibitory synapses along this length of axon provides a safety factor of 2–3 times for the blockage of conducting impulses.

The requirement that a certain length of axon must be under inhibition in order to block conducting impulses may explain why the orthodromic impulse initiation site of the neuron is at a location on the axon about 300 μm away from the soma. When constant tension is applied to the dendrites of the slowly adapting neuron by stretching the receptor muscle, the receptor neuron fires repetitively with a constant frequency determined by a steady generator potential level which is linearly related to the tension. These orthodromic impulses are initiated at a spike initiation site on the axon about 300 μm away from the soma, which is outside of the axonal area covered by inhibitory synapses. In the stretch receptor neuron of the lobster, the impulse initiation site is reported to be 500 μm away from the soma [16]. The impulses not only propagate toward the CNS (orthodromic direction), but also propagate back to the soma (antidromic). Thus, the antidromic impulse invading the soma is recordable with an intracellular electrode inserted into the soma.

In the experiment shown in Figure 12, the receptor neuron was kept under constant tension to induce firing of constant frequency. When GABA was applied to the axon at a spot 180 μm away from the soma, the invasion of impulses to the soma was blocked, leaving axonal impulse only. During this blocking period, the frequency of stretch-induced impulses increased, although the stretch had been kept constant. This implies that the higher frequency under the blockage of antidromic invasion of impulses is the frequency directly related to the generator potential level set by the steady tension. The frequency without blockage of antidromic invasion, on the other hand, is a frequency affected by the antidromic impulses. In other words, when antidromic invasion occurs, the frequency of impulses becomes lower than that directly set by the tension.

When the neuron is firing as a result of a given tension, the persisting generator potential keeps the membrane at a depolarized level set by the tension. If an impulse invades the soma in this condition, the repolarizing phase of the impulse creates a hyperpolarizing after potential which is superimposed on the generator potential—resulting in a lowering of the level of the generator potential. This will naturally reduce the frequency of firing. The effect of lowering the generator potential level by invading impulses must be dependent on the frequency, because the mechanism is apparently due to the summation of hyperpolarizing after potentials. Thus, the higher the level of the generator potential, the larger the hyperpolarizing after potential. Therefore, the effect of the invading impulse on the frequency is more stressed when the receptor is firing at a higher frequency. It is predicted that the frequency-tension relationship will be skewed from linearity at the high-frequency end.

Only if antidromic invasion of its own impulses is blocked then the frequency-tension relationship can become linear. This must be the role of the inhibitory synapse distributing on the soma and on the axon, between the soma and the spike initiation site. Orthodromic impulses traveling toward the CNS are not blocked by GABA applied to the soma or axon. Thus, under this inhibition, the CNS receives information without being modulated by antidromic invasion—information of high fidelity.

5.2 Vertebrate brain cell

Inhibitory synapses on the soma and axon hillock have been found in many other neurons (e.g., pyramidal cell). Most of physiological paper has interpreted the function of these synapses as being to attenuate the excitatory input—classical postsynaptic inhibition—the dendrites were originally thought to act as simple receivers (p. 1718 in ref. [6]). The present study has demonstrated that the function of the inhibitory synapse on the soma and the initial segment of the axon is not the classical one, but rather is to protect the dendritic area where input is received from disturbance coming from the backfiring of the soma caused by antidromic invasion of its own impulses. This thereby provides high fidelity for the input-output relationship of the neuron.

5.2.1 Axo-axonic cell or chandelier cell

As shown in the review [36], GABAergic chandelier cells (axo-axonic cells first documented by Ramón y Cajal) have a unique arborization (see Figure 2 of ref. [36]). One axo-axonic cell innervates 26 pyramidal cell’s axon initial segment. Schneider-Menzel et al. [37] reported that 153 pyramidal cells receive chandelier input, although the synapse number is highly variable across the population and is correlated with structure features of the target neuron. This indicates that an interneuron connects to multiple principal cells. Furthermore, pyramidal cells are known to be electrically coupled. Single neurons and populations work together [38]. It is understandable that biophysically homogenous population [39] of neurons, assembly of neurons, can synchronously function as the switching device. We assume that this switch works as powerful (i.e., working simultaneously across the large brain area) converter of specialized cognitive machinery. “Eureka moment” and “rabbit-duck illusion” might use this switch, because we sometimes behave momentary extremely concentrated mode to think about just one thing like thoughtful Archimedes did.

5.2.2 GABA in the brain neural network

A review [40] described: “neural networks in the brain include principal neurons and GABAergic interneurons (e.g., basket cell). The latter is vital for normal brain function because they regulate the activity of principal neurons. PV (parvalbumin) interneurons (i.e., basket cell) are involved in gamma-frequency oscillations, and they also play a role in complex network operations, including expansion of dynamic activity range, pattern separation, modulation of place and grid field shapes, phase perception and gain modulation of sensory response. PV interneurons also play key role in numerous brain diseases. These include epilepsy and also complex psychiatric diseases such as schizophrenia. Thus, PV interneurons may become important therapeutic targets in the future.”

The review continues: “However, much needs to be learned about the basic function of these interneurons before clinical neuroscientists will have a chance to successfully use PV neurons for therapeutic purpose.” The present finding on the crustacean stretch receptor neurons provides a novel insight how PV interneurons shape complex network function.

Electrical stimulation to a special point of brain slice is usually not feasible because electric current cannot precisely stimulate the target, producing non-specific influences on many other unrelated neurons. Using GABA application to a point, Gómez et al. [41] studied somatodendritic inhibitory effects of GABAergic transmission. The GABAergic shunt mechanism supporting cognition could be proved in the brain one day. “The blockage of backfire at the axon initial segment” requires verification in the intact vertebrate brain.

5.2.3 GABA and chloride ion

Inhibitory postsynaptic potential (IPSP) at GABAergic synapse is mediated by chloride ionic currents. The reversal potential of IPSP (Ei) is close to the resting membrane potential (Em). Em was −60 mV with a good condition in the present study. If Ei is assumed to be −65 mV for example, IPSP must show a hyperpolarizing potential. As shown in Figure 5C, hyperpolarizing IPSP blocked antidromic invasion (cf. Figure 14 in ref. [23]). If Em is close to Ei, GABA causes little change in membrane potential but a significant decrease in input resistance because chloride channels are open. This shunt inhibition is the mechanism of blockade of antidromic invasion. If Ei is kept near Em, inhibitory nerve can induce only shunting effects without altering the membrane potential.

GABAergic signaling is controlled by the intracellular concentration of chloride ions. The chloride-extruding K⁺-Cl⁻ cotransporter (KCC2) and the Na⁺-K⁺-2Cl⁻ cotransporter (NKCC1) which facilitates the accumulation of Cl⁻ in neurons, maintain intracellular concentration of chloride ions. Thus, GABAergic signaling is controlled by both KCC2 and NKCC1. It is understandable that functional defect of them leads to the alteration in chloride homeostasis in CNS cells. GABA-chloride relationship is important for brain healthiness. Blockage of backfiring under the chloride ionic mechanism could be a therapeutic target in pathophysiology and pharmacology: the Renshaw cell in the spinal cord (this is not GABAergic but chloride dependent inhibitory), the basket cell, and chandelier cell in the brain might be candidate targets.

Abnormal inhibitory neurotransmission causes brain disorders [42, 43]. Gonzalez-Burgos et al. reported that cognitive deficits in schizophrenia may result from a GABA synapse dysfunction that disturbs neural synchrony [43]. The blockage of the antidromic invasion by GABA is not a very scientific fiction idea but physiological phenomena because it is measureable. Legitimacy of this phenomenon requires investigation in the intact brain.

Our brain can quickly shift between different conditional schemes, executing complex cognitive processes. Perhaps our brain has the switching tools to represent concepts, inner world of thought and desire, images and idea, self and consciousness, and so on. “Some interneurons might be to determine the timing of action-potential firing during rhythmic activity” (see the review by Spruston, in page 212 of ref. [44]).

5.2.4 Insight accompanied by “Aha” experience or “Eureka moment”

People solve problems with a unique process called insight, accompanied by “Aha” experience [45] and revealed a sudden burst of high-frequency gamma-band neural activity at a time of insight solution (see Figures 4–6 of ref. [45]. Both alpha-band and gamma-band power (scalp electro-encephalogram recording with power spectrum analysis) quickly shifted from a low state to a high state or vice versa at an insight moment. It took approximately 500 ms for this switching. Thus, the switching speed is ca. 500 ms.

In a study where impulse firing rate was studied [46], the switching speed of 260–610 ms was calculated (Pair #1 in Figure 5 in ref. [46] was used for this calculation), although the authors focus was not the speed of switching. In a computer model of neuron and circuit [47], the switching speed of 1.4 s was calculated (Figure 7B in ref. [47] was used for this calculation).

In Figure 12, the firing frequency before GABA is 1.64 Hz (antidromic spikes are invading) and that during GABA is 2.47 Hz (during the blockage). It took 500 ms for the switching. If the neuron fires much high frequency, this switching speed would be shorter.

This consideration indicates that brain wave recording in human [45], firing rate recording in rat [46], and crayfish recording (the present study Figure 12) do not show inconsistency.

5.2.5 Why principal neuron has a large oval soma?

Figure 13 diagrammatically illustrates principal neuron and inhibitory nerve. The dendrites integrate afferent inputs [6, 34]. Integrated currents initiate action potential at the axon. Action potential conducts both orthodromic and antidromic directions (green arrows). Electric currents necessary to initiate antidromic spikes are shown in red arrows. Inhibitory interneuron (black) makes synaptic contacts on the axon (Ax-Ax), on the soma (Soma), on proximal site of dendritic tree (p Dendrite), and on the distant site away from the soma (d Dendrite). Only one dendrite trunk is shown but it represents all dendrites, such as apical and basal dendrites of the pyramidal cell.

Figure 13.

Diagram to illustrate one principal neuron and one interneuron.

The diameter of the axon is 5 μm and the diameter of the oval soma is roughly 50 μm: 10 times difference in diameter between the axon and the soma. Safety factor for the action-potential generation at somal membrane significantly decreases due to less current per area. Consequently, active spike invasion tends to fail. On the other hand, “shunting switch” works efficiently at the soma area and at wider dendrite area too (see ref. [34] for dendritic integration). This consideration is the answer to the question “why principal neuron has a large oval soma?” Once again, axonal firing is relatively uneasy to stop it, because it jumps like the salutatory conduction due to a high safety factor of the axon. On the other hand, somal firing is relatively easy to stop, because less electric current per area flows into.

Actually, sodium channels distribute not evenly (see ref. [34]). Dendrite soma can generate Ca⁺⁺ spikes [34]. At the dendrite, EPSP, IPSP, and presynaptic IPSP occur [34]. The dendrite is like the battle field of various currents shooting. Thus, things are not simple, but “shunting switch” scenario is untested but potential hypothesis in vertebrate brain.

Individual inhibitory interneuron has own target [3, 34]. Their target to execute shunting is genetically and behaviorally determined during maturation. Each neuron “decides” where to settle, whom to connect, what to do, finally contributes forming LTP synapses, Hebbian synapse, and/or working memory synapses, all for the sake of neuroplasticity.

Each interneuron, selectively and independently, synapses on the soma, on the “p Dendrite” or “d Dendrite” and so on. Among them, axo-axonic synapse (Ax-Ax in Figure 13) seems to be most powerful. It can stop all antidromic invasions. The axo-axonic synapse alignment can afford precise output control (i.e., bi-stable, on/off or 0/1 digital control).

In conclusion, in the present study, we revealed powerful switching mechanism from crustacean experiments. The switching phenomenon might be possible to contribute to various mind-mental functioning such as “Eureka,” “Rubin’s vase,” “rabbit–duck illusion,” and more such as “selective attention” [48] or “cocktail-party effect.” “The science of the brain” and “the science of the mental” [49] move closer to each other. The present study was partly appeared at the IUPS international congress [50].

5.2.6 GABA in brain: Concluding remarks

Crayfish are great commercial importance as a human food delicacy [51]. The US east coast (Maine) fisheries supply 80% of the world lobster [52]. Its abundance and common have probably contributed to neurobiology progress, as Bullock [53] noted that the crustacean animals were unquestionably a giant stride forward in the understanding of substratum of nervous system function. Historically, Alexandrovicz [11, 12] first introduced lobster neuron for neurobiology research. Kuffler [21], who worked at Harvard and at Woods Hall Marine Laboratory (both closer to the US east coast), used lobster stretch receptor to bring the brain science forward, and then neurobiology has indeed advanced [54].

A small molecule, GABA, was first identified in plant extracts [55]. GABA was then found in mammalian brain tissues [56]. The report [56] showed that GABA is formed naturally from glutamic acid in the brain. However, it takes many years before GABA is recognized as an important inhibitory neurotransmitter in brain [10, 57]. At first, unknown inhibitory substance “an inhibitory factor (factor I),” was described [58], and then isolated and identified as GABA [59]. Finally, GABA formation in nerve cells and GABA release from nerve cells were experimentally demonstrated using “crustacean” nerve cells [8, 9, 22, 23].

The present study might strengthen the brain’s GABA history aforementioned, that is, there is a similarity of GABAergic nervous system between invertebrate (i.e., inhibitory neurons innervating stretch receptor neurons) and vertebrate (e.g., basket interneurons innervating pyramidal neurons). In the future, “GABAergic neural switching function” that we report here, could be confirmed by electrophysiology experiments, for example, the multi-electrode recording while studying intact brain [60]. We believe that the major brain function, e.g., “how to select and match preexisting memory with events in the world [60]” could be executed by GABAergic “switching” mechanisms, which is revealed by the present study. This has worked out with heading along the way paved by Kuffler.

To conclude, a small molecule naturally occurred evolutionally very earlier (in the cells of microorganisms, plants and animals) as a non-protein amino acid [55, 61]. Meanwhile, the same molecule was utilized as a key signaling molecule in the brain of both lower and higher animals. It is beyond controversy that life on earth uses a lot of common basic molecules, like DNA, ATP, etc. to survive. In analogy with these molecules, evolutionally speaking, GABA is a fundamental substance too. Mostly, GABA works inhibitory, and thus, it helps very quietly healthy life everywhere in brain. If alterations in GABA function occur, sickness surely occurs, like epilepsy [57, 62]. This report reports the significance of such a humble substance working in our life. Healthy life is sweet indeed.

Acknowledgments

I am thankful to the Editor Dr. Noor Saher for her revisions and to Dr. G. Diarte-Plata for helpful comments. I am also thankful to T. Tsuruta, Nippon Unisoft Co., Ltd. Tokyo, for decades-long encouragements for my neurobiology study. This work is supported from the JSPS grant 21 K12688 to TY.