Developing spinal motor networks produce a diverse array of outputs, including episodic and continuous patterns of rhythmic activity. Variation in excitability state and neuromodulatory tone can facilitate transitions between episodic and continuous rhythms; however, the intrinsic mechanisms that govern these rhythms and transitions are poorly understood. Here, we tested the capacity of a single central pattern generator (CPG) circuit with tunable properties to generate multiple outputs. To address this, we deployed a computational model composed of an inhibitory half-centre oscillator. We tested the contributions of key properties predicted by the model to the generation of an episodic rhythm produced by isolated spinal cords of the newborn mouse. The model was capable of reproducing the diverse state-dependent rhythms evoked by dopamine in the neonatal mouse spinal cord. In the model, episodic bursting depended predominantly on the endogenous oscillatory properties of neurons, with persistent Na+(INaP), Na+-K+ ATPase pump (IPump), and hyperpolarization-activated currents (Ih) playing key roles. Modulation of all three currents produced transitions between episodic and continuous rhythms and silence. Pharmacological manipulation of these properties in vitro led to consistent changes in spinally generated rhythmic outputs elicited by dopamine. The model also showed multistable zones within a narrow range of parameter space for IPump and Ih, where switches between rhythms were rapidly triggered by brief but specific perturbations. Outside of those zones, brief perturbations could reset episodic and continuous rhythmicity generated by the model. Our modelling and experimental results provide insight into mechanisms that govern the generation of multiple patterns of rhythmicity by a single CPG. We propose that neuromodulators alter circuit properties to position the network within regions of state-space that favour stable outputs or, in the case of multistable zones, facilitate rapid transitions between states.Significance statementThe ability of a single CPG to produce and transition between multiple rhythmic patterns of activity is poorly understood. We deployed a complementary computational half-centre oscillator model and an isolated spinal cord experimental model to identify key currents whose interaction produced episodic and continuous rhythmic activity. Combined, our experimental and modelling approaches suggest mechanisms that govern generating and transitioning between diverse rhythms in mammalian spinal networks. This work sheds light on the ability of a single CPG to produce episodic bouts often observed in behavioural contexts.