Search

Studies the interaction of tonic impulsation, carrying information on the structure of external images (letter symbols) and rhythmic forms of activity, fixing this information, on the model of a network of neurosimilar elements. A mathematical model of a neural network, realizd by the Nord-100 computer in program form, was based on neurophysiological data characterizing the properties of the exciting neurons, and their contacts and conditions of their interaction at learning. The model was used to solve a system of equations at every clock cycle, describing the state dynamics of each element and all contacts between them. Learning in the network was realized by Hebb's rule. >

Microelectrode analysis of red nucleus (RN) neurons was used to study the regulatory influences on an axiosomatic input from a cerebellar nucleus interpositus and on an axodendritic input from a sensorimotor area of a cortex. To bring together these experimental data and previous theoretical results on detector properties of RN neurons, an improved model neuron with double-barrier synaptic elements has been developed. Based on results of simulation experiments with the model, several quantitative generalizations have been developed as aids to an extended understanding of the computational properties of the RN neuron. >

A model is proposed that is based on an attempt to unify the informational duties of a biosystem and to describe a canonical way of solving any problem that can appear. A language is introduced to make a correspondence between neuron-molecular and behavioral levels that, on the one hand, could be realized at the molecular level and, on the other hand, could be formal basis for solving artificial intelligence problems. The physical solver operates with the physical symbols' predicates as symbol space elements. It solves equations expressed in predicates, but not numerical or functional equations. An experiment proving the existence of a physical solver is described. >

A novel approach to the design of devices demonstrating the behavior of biological organisms that can be based on the associative memory mechanism is presented. A model of associative memory is proposed in which the dependent number of memory elements linearly increases with the number of medium elements, and every element of the medium is impressed with the activities of all the other elements. The authors focus more on the technical realization of parallel calculations on neural-like elements than on the modeling of actual neural activity. >

General features of an asymptotical method for an analyzing complex system like neural networks are presented. The method is analogous to the mean-field approach and allows treatment not only of steady states but also of dynamical properties of networks. It may also be interpreted as a Galerkin procedure for the master equation. The types of neural networks and related problems to which the method can be applied are discussed. It is shown that the method can treat synchronization processes, networks of excitable neurons, nonidentical neurons, and nonidentical synapses. >

Grounds and principles for the construction of a clonal neural network model are considered. The structure of the clonal network is then considered. Algorithms are presented which form the basis for the construction of a neuroelement model for differentiating and integrating networks of the clonal neuronetwork model. >

The authors consider motor control neural networks as specific dynamic systems and assume the choice of the particular movement among the different implementations of the motor plan to be a consequence of their organization and operation. They attempt to develop a model of multijoint movement control and learning in the context of the whole neuro-musculo-skeletal system, whose observed movement is the consequence of mechanical properties of the musculo-skeletal system, nervous regulation and neural mechanisms of motor learning. In accordance with these natural levels of description three inherently connected components of the model are considered. >

An attempt is being made to create an experimental model of the information process of nerve cells which will make it possible to investigate the mechanisms of the electron modification of the separate inputs, molecular organization, and dynamics of this process. The possibility is investigated of using the electron plasticity of the separate regions of the cells to experimentally verify a hypothesis about how the weights of the separate inputs change. The experimental data obtained show that local effects on the inputs of the molecular processor bring about, at the first stage, uniform rearrangements in the functioning of output molecules of its effector systems. Such rearrangements take place in a limited zone and do not depend on the direction that any further change of the resulting response may take. An experimental methodology has been developed making it possible to obtain data about the mechanisms by which molecular complexes function in the cell in the processes of perceiving, processing, and fixing information. >