Characteristics of Random Nets of Analog Neuron-Like Elements

摘要: The dynamic behavior of randomly connected analog neuron-like elements that process pulse-frequency modulated signals is investigated from the macroscopic point view. By extracting two statistical parameters, state equations are derived in terms these parameters under some hypotheses on stochastics microscopic states. It shown a random net statistically symmetric structure monostable or bistable, and stability criteria explicitly given. Random nets consisting many different classes also analyzed. Special attention paid to excitatory inhibitory elements. stable oscillation exists such net?in contrast with fact no oscillations exist even if negative as well positive synaptic weights permitted at time. results checked by computer-simulated experiments.