The neural networks that underlie locomotion are complex and require integration of sensory input and physiological state. However, how these networks function during active locomotion to incorporate sensory input from the environment and the internal state of the animal remains poorly understand. The zebrafish larva is an ideal vertebrate to study these questions thanks to its simple locomotor repertoire, transparency, and amenability to genetic manipulation. In Chapter 1, I describe a program to track behavior at high speeds and automatically characterize locomotor patterns in a high-throughput manner. V2a interneurons are excitatory interneurons in the spinal cord and hindbrain identified by the chx10 transcription factor. In Chapter 2, I validated the use of a genetically-encoded botulinum toxin to silence the chx10 population in vivo. Using fictive locomotor recordings and calcium imaging, I demonstrated that silencing V2as leads to decreased activity in primary motor neurons during fast swimming, corresponding to a lower swimming frequency in V2a-silenced larvae. Cerebrospinal fluid-contacting neurons (CSF-cNs) are intraspinal neurons that relay sensory information to motor circuits. CSF-cNs in diverse species express GABA and the transient receptor potential channel PKD2L1. In Chapter 3, I used genetic targeting, calcium imaging, pharmacology, and electrophysiology to investigate the role of spontaneous activity in CSF-cNs. I showed that single channel opening of PKD2L1 represents an intrinsic source of spontaneous activity in CSF-cNs. These tools and results will allow a more complete picture of how dynamic interactions shape locomotor output in vivo.