A reliable neuromime representation of fatigue based mutual inhibition

In this study a pair of VLSI-based artificial neurons with built in fatigue are used to realize mutual inhibition. Each artificial neuron is nominally self-excitatory with a single inhibitory input driven from the activation of the other neuron. The fatigue feature of each cell weakens its response over the duration of the complementary cell's activation. This allows the inactive cell to recover and thus inhibit its counterpart. Using this representation of mutual inhibition results in stable oscillations with a wide range of flexibility and control. Parametric analysis of network frequency indicates that this type of reciprocal inhibitory network is more profoundly affected by fatigue properties than the rate of individual cell self-excitation. >