Emergent mechanisms in multiple pattern generation of the lobster pyloric network

By using a hard-wired oscillator network, multiple pattern generation of the lobster pyloric network is simulated. The network model is constructed using a relaxation oscillator representing an oscillatory or quiescent (i.e. steady-state) neuron. Modulatory inputs to the network are hypothesized to cause changes in the dynamical properties of each pyloric neuron: the oscillatory frequency, the postinhibitory rebound property, and the resting membrane potential. Changes in each of these properties are induced by changing appropriate parameters of the oscillator. By changing seven parameters of the network as a whole, modulatory input-dependent patterns are successfully simulated. Received: 13 July 1999 / Accepted in revised form: 17 December 1999