To understand the neural basis of behaviour, identified cellular, synaptic, and molecular mechanisms must be examined in the neuronal networks that generate specific behaviours. The lamprey, a lower vertebrate, allows activity in the locomotor network that generates swimming to be examined in the intact spinal cord in vitro. A single 10min application of the neuropeptide substance P evokes a long-term (>24 h) increase in the frequency of segmental network activity in adult, but not larval lampreys (Parker 2000). The burst frequency modulation has three phases: an induction phase (~2 h) that is associated with the protein kinase C-mediated potentiation of NMDA receptors; an intermediate phase (~2-15 h) that requires protein synthesis, but not de novo RNA synthesis; and a final phase (>15-20 h) that requires de novo RNA synthesis. At the synaptic level, substance P converts the activity-dependent depression of excitatory network interneuron (EIN) inputs to motor neurons during physiologically relevant spike trains into facilitation, an example of metaplasticity. This effect is associated with a reduction of the transmitter release probability (to allow facilitation), but an increase in the number of docked synaptic vesicles (which ensures that the facilitation does not develop from a reduced initial EPSP amplitude; Bevan And Parker 2004. The synaptic metaplasticity shares the same induction and maintenance mechanisms as the burst frequency modulation, suggesting that the two effects are linked. While interneuron inputs to motor neurons provide some insight into the plasticity of the network output, it is the connections between network interneurons that probably pattern network activity. The effects of substance P on these connections between network interneurons are now being examined. The analysis is complicated by the uncertainty over the network organisation (i.e. what types of neurons and synapses are involved in generating the network output; Parker 2000). It seems likely that interneurons involved in generating segmental network activity are relatively small, the larger interneurons (lateral and crossed caudal interneurons) that were originally examined having a role in intersegmental coordination. This analysis shows that substance P has synapse specific effects on connections between putative network interneurons: it reduces the strength of connections between the EINs, but increases the strength of feedback inhibition onto the EINs. The connections between identified network interneurons must now be examined in detail, and the changes at these synapses related to effects on the network output. The molecular mechanisms underlying the protein synthesis-dependent synaptic and network effects are being examined, and should facilitate these analyses.