Rhythmic movementssuch as swimming, flying, and walkingare ubiquitous in nature. Intrinsically active neural networks called central pattern generators (CPGs) provide the feedforward signals to actuate these movements, but the preferred movement frequency is often equivalent to the resonant frequency of the musculoskeletal system. Sensory feedback is essential to synchronize the neural and musculoskeletal systems to the mechanical resonant frequency, a phenomenon called resonance tuning. In this dissertation, we use a simple computational model of rhythmic movement to understand how the configuration of sensory feedback affects both the sensitivity of resonance tuning to parameter variation and the resiliency of resonance tuning to perturbation. Although previous studies have shown that resonance tuning is limited to frequencies that are above the intrinsic CPG frequency, we demonstrate that this limitation is only valid with negative feedback and with endogenously bursting CPG neurons. Specifically, we show that with positive feedback, resonance tuning occurs at frequencies that are below the intrinsic CPG frequency. Moreover, when the synaptic connections within the CPG are required for bursting activity, resonance tuning occurs both above and below the intrinsic CPG frequency with negative feedback and does not occur with positive feedback. Using Floquet analysis, we then demonstrate that perturbations decay more quickly when resonance tuning is realized with positive than with negative proportional feedback. Finally, we evaluate how the intrinsic CPG frequency, feedback gain, and mechanical damping affect the stability and range of resonance tuning with negative and positive feedback. Overall, these results indicate that the configuration of sensory feedback dramatically affects both the parameter space in which resonance tuning occurs and the stability of the resultant periodic motion.