Analysis of neuronal networks in the visual system of the cat using statistical signals

Abstract

If the input signals of the visual system in the cat are statistical patterns in space and time, a complete system analysis can be carried out. What counts here as a system are the neuronal networks between retina and recording site. In the case of linearity, one obtains the temporal impulse response functions at every point in the receptive field with the aid of correlation methods. The measuring time is about one minute. Some aspects of the procedure are explained in terms of examples. The method of measurement also makes it possible to determine the characteristic function of the system in time and space between different recording sites within the cortex. It is possible to specialize the procedure for analysing the stationary space dependent behaviour of neuronal networks. The extension of the analysis procedure allows a concise, relatively simple description for nonlinear systems in which linear and nonlinear subsystems can be separated. Besides this, there are no restrictions concerning the kind of nonlinearity. A second paper will present the detailed experimental application of these methods.