Understanding the genesis and control of movement is a major goal of neuroscience because of movement's role in behavior and the tragic consequences of diseases that impair movement. Rhythmic patterned movements (e.g., walking) have often been used to investigate these issues because of their importance, ubiquity, inducibility, and stereotypy. We now know that the basic rhythmicity and patterning of almost all such movements arise from the activity of relatively small, endogenously rhythmic central neural networks (central pattern generators) that can continue to cycle in the absence of patterned input. These networks thus not only generate patterned rhythmic movements, but are also simple examples of the nervous system's ability to spontaneously create. Their study thus sheds light not only on movement generation, but possibly also on the mechanisms underlying more complex abilities of nervous systems. Rhythmic neural networks have therefore been extensively studied. This work has revealed that these networks generally contain both complex, non-hierarchical, distributed synaptic connectivity patterns, and neurons with complex, non-linear, active cellular properties. This complexity presumably underlies a third surprising result of this research - these networks produce multiple outputs. This research has dramatically increased our understanding of (and appreciation for) the capabilities of these networks, but we still have relatively little understanding of the mechanisms that give rise to these multiple activity modes, or of their functional consequences (i.e., the motor patterns they produce). A common example of this """"""""multiple activity"""""""" capability is that many of these networks produce phase constant outputs as cycle period is altered (e.g., the network proportionally alters all burst durations and inter-neuronal delays so as to produce the """"""""same"""""""" pattern at all cycle periods). This proposal outlines a combination of neuron and muscle experiments on, and computer simulations of, the well described pyloric neuromuscular system of the lobster, Panulirus interruptus, aimed at elucidating how this system maintains phase. The experimental advantages of this preparation allow all the network's neurons to be individually characterized, and the system to be studied on all levels from membrane conductances to muscle contraction.
