ISSN 1064-2269, Journal of Communications Technology and Electronics, 2010, Vol. 55, No. 7, pp. 759–764. © Pleiades Publishing, Inc., 2010. Original Russian Text © V.V. Klin’shov, D.S. Shchapin, V.I. Nekorkin, 2010, published in Radiotekhnika i Elektronika, 2010, Vol. 55, No. 7, pp. 812–817. 759 INTRODUCTION In modern nonlinear physics, much attention is devoted to the problems of investigating ensembles of active elements that simulate the collective dynamics of neural networks [1–7]. Increased interest in this field is caused by both recent progress in neurophysiology and prospects of creating a new generation of artificial elec- tronic systems of information processing in which the principles of neurodynamics are employed. An example of such a problem is the creation of a universal control sys- tem based on olivocerebellar physiology [8–11]. In this study, an electronic circuit, which simulates the basic element of the neurodynamic system of short-term or working memory, is developed. The neurodynamic sys- tem of working memory, an important subsystem of the brain, is intended for temporal storage of information. Data are stored in the form of clusters of periodic electric activity of neurons. The most interesting property of the working memory system is its ability to ensure simulta- neous remembering of several images (pieces of informa- tion) in the same structural elements, which is not achiev- able in existing artificial data storage systems. Below, basic principles of system operation are described with the use of its model proposed and investi- gated in [12–18]. The neural network intended for per- forming the functions of working memory was described in [12–14]. The mathematical model of this network was proposed and analyzed [15–18]. It has been shown that neurons need to be fed with a mixture containing a peri- odic subthreshold stimulus (an oscillatory signal) and a short excitation pulse (an information signal) to imple- ment information storage. Before the arrival of informa- tion signals, neurons are in the mode of subthreshold oscillations. After the arrival of information signals, neu- rons begin demonstrating periodic excitations, which cor- respond to the appearance of neuron activity clusters and remembering of loaded data. In this paper, it is shown that our electronic circuit is described by the equations similar to the dynamic model of the neuron element of the work- ing memory system [15]. The experimentally studied cir- cuit dynamics also demonstrates properties resembling those of the theoretical model proposed in [15]. 1. CIRCUIT OF THE BASIC ELEMENT The working memory system is a network of neurons with a specific property of so-called afterdepolarization. This property manifests itself as a temporary decrease in the excitation threshold of a neuron membrane after the action potential is generated (Fig. 1a). The electronic cir- cuit proposed in this study simulates the dynamics of one neuron with afterdepolarization. Such a neuron is the basic element of the working memory system, which is capable of storing data on the arrival of a single informa- tion pulse in the form of periodic electric activity. This pulse forces the system to change over from the mode of subthreshold oscillations to the mode of periodic genera- tion of excitation pulses (Fig. 1b). The block diagram of the basic element comprises two main blocks (Fig. 2a). Excitable block 1 is responsible for generation of electric excitation pulses. Block 2 varies the excitation threshold of the basic element. The electronic circuit operates as follows. Excitable block 1 is stimulated by an oscillatory (sinusoidal) signal and continuously gen- erates periodic subthreshold oscillations. After the arrival of a short information pulse, block 1 is excited and acti- vates block 2 (intended to generate exponential pulses). The exponential pulse of block 2 is transmitted through a feedback loop to block 1 and raises its excitability. This allows an oscillatory signal to implement subsequent peri- odic excitations of block 1. As a result, the system stores data on the arrival of an information signal in the form of periodic electric activity. The schematic diagram of the basic element of the neurodynamic memory system is depicted in Fig. 2b. Block 1 is a FitzHugh–Nagumo pulse oscillator with cubic nonlinear component N built around an opera- tional amplifier. The output signal of block 1 is fed to the input of block 2 and, thereafter, passes through a compar- ator based on an operational amplifier, a low-pass filter, and a divider. The output signal of block 2 passes through THEORY AND METHODS OF SIGNAL PROCESSING Simulation of the Neurodynamic System of Working Memory V. V. Klin’shov, D. S. Shchapin, and V. I. Nekorkin Received March 16, 2009 Abstract—The electronic circuit of the basic element of a neurodynamic memory system has been developed and implemented. The dynamic modes of the circuit and its responses to different external signals have been investigated experimentally. The circuit under study is shown to have all dynamic properties required for the memory system. The experimental results agree with those obtained in the previous theoretical studies of the neurodynamic memory system. DOI: 10.1134/S1064226910070065