An asynchronous pulse neural network model has been proposed. The model uses a dynamical spacio-temporal pattern in a network to process information. In such a network, precise timing of each action potential is important. Therefore, an analog circuit implementation of the model is desired because the analog circuit can handle continuos time and values. In this paper, we integrate a neuron circuit for the asynchronous pulse neural network model. The model parameters of the IC chip can be altered externally, so that the neuron circuit works as either an integrator or a coincidence detector, or in between. Experimental measurements confirm that the integrated neuron circuit qualitatively replicates the behavior of the asynchronous pulse neuron model.
展开▼