Bionics: estimation the number of code states of data exchange in the structure of microconsuming, high-dimensional, nature-like neurocomputers

Background. At present, the interest of developers in neuromorphic calculations and the construction of artificial neurons that at least partially repeat the work of natural neurons of the simplest living organisms has increased. The objectives of the article are to evaluate the previously obtained data and to study the easily observable characteristics of microconsuming, high-dimensional, nature-like neurocomputers with their subsequent simplified projection into the structure of artificial neurons. Materials and methods. Methods of continuous-discrete information processing by natural and artificial neurons are used using neural network analogues of statistical criteria, as well as real data obtained in the study of neural network activity of living beings from the middle of the last century to the present. In the process of information processing analysis, special attention is paid to the transfer of data between natural neurons along long axons. It is proposed to consider natural neurons as a kind of impulse modem that encodes data transmitted via the axon to other natural neurons in the Q-ar system of neurocalculus. Results and conclusions. It was obtained that in all the previously recorded bursts of impulses in the axons of natural neurons, the maximum number of impulses is significantly greater than two. This is interpreted as the absence of binary natural neurons in living organisms. This thesis is supported by numerical simulations of binary and ternary artificial neurons. Binary artificial neurons are much worse than ternary artificial neurons in terms of information enrichment during its extraction and with an increase in entropy with neural network protection of knowledge. It is proposed to approximate the Q-arity of the neuronumber system used by the maximum number of impulses contained in the transmitted data pack of specific neurons, both natural and artificial neurons.