A three-variable silicon neuron circuit

The silicon neuron is a type of artificial neuron implemented with electronic circuit. Previously a design approach based on mathematical structures under neuronal dynamics was proposed. It is based on the mathematical techniques such as phase plane and bifurcation analysis. These methods allow us to implement silicon neuron with smaller circuit area and to strategically adjust the bias parameter voltages without losing variety of output patterns. In this study we demonstrate a mathematical-structure-based silicon neuron, which operates a three-dimensional system. This silicon neuron can generate a firing pattern called square-wave bursting. In this report we show the experimental results of this silicon neuron. We are planning to make pattern generating network using this silicon neuron and silicon synapses.