We have demonstrated earlier that excitatory influences in some cases may exert a protective action. Therefore, a one-dimensional interpretation of beneficial influences as inhibitory cannot be a general rule. Similarly, when we have considered cellular damage and protection in topic 2, we certified that homeostasis can support a lively state of cells even if some vital parameter, for instance excitability, has deviated from the norm.
"Protective" influence of the hyperpolarization during habituation and classical conditioning was postsynaptic in nature, since it was connected with reorganization of excitable membrane. This was forestalling perfection of acquiring during learning adjustment and it was evidently directed in the right direction and supported the adequate change in responses. Inhibition retarded ineffective response to habitual stimulus and accelerated augmented response to the conditioned stimulus. During instrumental conditioning, inhibition and excitation of neurons lacked dependence from the level of membrane potential and reorganization of the instrumental reaction had another reason.
We compared in our discussion the behaviors of membrane potential and neuronal responses during training and have revealed inconsistency in the manners of their behaviors. Linear correlation between these characteristics during habituation and classical conditioning also differed from normal regularities. In addition, at the beginning of extinctions, hyperpolarization coincided in time with the augmentation of responses to the CS+ (Fig. 3.4). Nevertheless, these are too indirect observations. Therefore, it would be important additionally to reveal how the efficiency to generate an AP depends on the level of membrane potential at the moment of spike generation. Taking into account that implementation of direct intracellular stimulation for evaluation of excitability may give an improper result, we compared how the level of the membrane potential in effective responses (when an AP was generated) differs from the level of membrane potential during failure of AP generation. Fig. 3.9 illustrates this comparison for the control neuron.
Fig. 3.9. Reorganization of properties of excitable membrane in the control neuron during a learning session (instrumental conditioning). Ordinate the value of membrane potential, on a per-unit basis. Filled symbols: membrane potential in those trials where the control neuron generated APs. Light symbols: membrane potential in the trials where the control neuron failed to generate APs. Standard errors are indicated.
At the beginning of training, the control neuron more often generated AP in the trials when it was spontaneously depolarized, in correspondence with the classical notion for dependence of excitability from membrane potential level. However, just in those trials where the control neuron generated an erroneous instrumental reaction (between 20 and 30 trials), the control neuron was hyperpolarized, whereas it was inhibited when the membrane potential decreased. Classical dependence of excitability to the level of membrane potential recovered only to the end of training (Fig. 3.9). This result is in agreement with the assumption that a protective role of inhibition endorses recovery of excitability after motivation-induced neuronal damage and this promotes the generation of output reaction. Taking this into consideration, we may evaluate the hyperpolarized state as a protection which improves neurons and thus paradoxically increases their excitability. It is most interesting to note here that between the 20th and 30th trials, the membrane potential of the control neuron, as an average, does not decrease. On the contrary, this particular neuron was hyperpolarized during this period of training. Hence, hyperpolar-ization can recover excitability of neurons not only as a counteraction to the excitotoxic depolarization, but also in the cases when distortion of excitability is not connected with a large fall in the membrane potential.
Such paradoxical regularities are present also for responses to the unconditioned stimulus. When the neuron tried to generate an instrumental reaction (both correct and erroneous), response to the unconditioned stimulus was larger during neuronal hyperpolarization [1257]. This phenomenon corresponds to recovery of the responses during classical conditioning, when compensatory hyperpolarization arises. We can suppose that the motivational excitation in our experiments induced non-excitotoxic damage in the control neuron. Compensational hyperpolarization decreased the damage and recovered spike generation. A similar phenomenon was discovered in the motor neuron of the Aplysia after the delivery of strong sensitizing stimuli: sensiti-zation resulted in hyperpolarization of the resting membrane potential and a simultaneous decrease in the spike threshold [246]. Also a paradoxical reaction is described, consisting of the facilitation of spike responses of neurons to the conditioned stimulus evoked by antagonists of ionotropic glutamate transmission, which was blocked by GABA [1184]. Application of the antagonists increased baseline neuron activity, often by factors of 23 compared with the baseline, while the latency of the response to the conditioned signal decreased and its duration increased. There is constant tonic inhibitory control of the activatory spike responses of pyramidal neurons to conditioned stimuli in conscious animals. Inhibition actively involves organizing the excitatory responses of neurons in the instrumental conditioned reflex [1184]. In addition, during thirst, when neurons of the mammalian subfornical organ depolarized in response to an increase in osmolarity, the amplitude of the AP increased [35], indicating an alteration in AP-generating mechanisms. Certainly, paradoxical augmentation of responses after amelioration of damage may be connected with recuperation not only of electric characteristics of membrane, but, too, other important life factors.
This phenomenon was discovered by N.E. Vvedensky more than one hundred yeas ago [1311]. Injury of neuromuscular junction leads to development of the paradoxical state of the tissue. While responses of tissue to stimulus after injury decrease, the higher the irritation the smaller becomes the response. Thus, small damage excites, whereas large damage inhibits and when damage is large, responses increase, if strength of stimulus decreases. Nevertheless it is necessary to emphasize that paradoxical properties of neuronal damage cannot explain the selectivity of reorganization of neuronal responses during learning, if damage to a cell and especially its homeostatic recovery is an unspecific phenomenon. For instance, change in membrane potential usually spreads throughout the cell and this parameter may selectively affect only remote parts of a cell, such as different dendrite branches.
When we implemented the same method for investigation of a trained neuron, we revealed a more complex picture (Fig. 3.10).
Fig. 3.10. Reorganization of properties of excitable membrane in the trained neuron during a learning session. Filled symbols: membrane potential in those trials where the trained neuron generated APs. Light symbols: membrane potential in the trials where the trained neuron failed to generate APs.
Only at the beginning of training before the influence of learning and during a short period of maximal depolarization was the trained neuron inhibited while being hyperpolarized, in correspondence with the classical rule. Spontaneous depolarizations of the membrane potential correlated with the increase in the current response to the CS+ only in the period of trials 28-35. Here the membrane potential declines from that found in the steady state condition. Outside of these short periods, we did not reveal any dependence of excitability from the membrane potential. Although absence of correlation between membrane potential and spike generation indicates a disturbance of excitability in the trained neuron during active avoidance of punishment, we did not reveal the paradoxical phase, when inhibition protects neurons from excitotoxic damage. Moreover, although the trained neuron was depolarized just in this period of training (trials 28-35) (Fig. 3.7,bottom), its excitability was recovered through other means (Fig. 3.10) as, for instance, happens after activation of cyclic AMP system. The instrumental AP in the trained neuron arises during trials 28-35, when abundant depolarization of membrane potential began to decrease (Fig. 3.10). Just as depolarization is not the only single means for induction of neuronal damage, hyperpolarization is not the only single means for protective action of homeostatic compensation. As we have indicated, this result could be predicted, since hyperpolarization of membrane cannot control a selective augmentation of the reaction evoked by the CS+, in opposition to the CS~. A trained neuron, probably, generates an instrumental reaction not because of the protective influence of hyperpolarization to injury. Although the membrane potential of neurons changed during instrumental conditioning, this is not the only reason for the reorganization of the neuronal activity, if we proceed from classical neuronal properties: activation of a neuron during depolarization and its inhibition during hyperpolarization. A neuron, evidently, uses additional means (transient modification of channels, pH, second and retrograde messengers) for protection.
