PROJECT SUMMARY A hallmark of the nervous system is its ability to transform and combine multiple sources of information to guide behavior, and to tune this signal processing through learning, for more accurate control of behavior. How this adaptive signal processing is implemented by neural circuits remains largely unknown, due in part to the presence of feedback loops that mix together and temporally shape neural signals in complex ways. This project uses an interdisciplinary experimental-computational approach to dissect the signal processing operations supporting the control of eye movements by vestibular and visual sensory inputs. Rigorous experimental protocols, computational models, and statistical fitting approaches will be developed to characterize the signal transformations occurring at multiple sites in the relevant neural circuitry, to disentangle the roles of feedforward and feedback signals in the control of eye movement behavior. These approaches will be used to analyze how specific changes in the way vestibular and visual signals are processed by the circuit support adaptive changes in the amplitude and timing of the eye movement responses to these sensory inputs. By advancing our understanding of how learning alters the way a circuit computes, this work bridges the key conceptual gap between neural plasticity at the synaptic level and learning at the behavioral level, and provides a scientific foundation for the development of improved clinical approaches for promoting recovery or replacement of vestibular function.