Project Summary The cerebellum plays a significant role in cognitive, emotional, and social behaviors, but we lack a basic circuit- level understanding of how it contributes to these non-motor functions. Classical cerebellar learning models postulate that cerebellar Purkinje neurons use error feedback to shape future actions, but recent recordings of Purkinje activity in associative learning tasks revealed that Purkinje neurons encode sensory, cognitive, and reward-related responses, not motor errors. These remarkable observations challenge the error-learning model and suggest novel cerebellar functions during associative learning. However, this has never been causally tested, and it is also unclear how these non-traditional responses are acquired during learning. This gap in knowledge is largely due to limitations of past technologies, since linking Purkinje activities with behavior would require simultaneous recording and perturbation of Purkinje activity, ideally with single-cell and sub-second precision in awake behaving animals. Therefore, to define non-motor computations in the cerebellum and to understand how they contribute to behavior, I will employ two-photon calcium imaging and holographic stimulation of novel excitatory and inhibitory opsins to record, track, and modify Purkinje activities throughout learning. My preliminary data suggests that complex spike responses in Purkinje neurons encode reward-related signals, whereas simple spike responses in Purkinje neurons encode timing signals that precede motor output. Based on this data, I hypothesize that reward-related signals present in complex spike responses are necessary for the acquisition of timing responses in simple spikes in Purkinje neurons and are therefore crucial for generating well-timed behaviors during associative learning. In the first Aim (K99), I will characterize simple and complex responses within individual Purkinje neurons during associative learning. Then, in Aim 2 (K99), I will determine how simple spikes in functionally defined groups of PNs influence behavioral output. These initial studies will allow me to determine, in Aim 3 (R00), how reward-related signals in complex spike responses shape timing signals in simple spike responses and therefore modulate behavior during learning. Together, this proposal will define cerebellar computations using non-motor signals and could profoundly enrich our understanding of cerebellar function. In the K99 phase, I will be mentored by Dr. Karl Deisseroth, co-mentored by Dr. Liqun Luo, and will be advised by an exceptional advisory team composed of Dr. Sean Quirin, Dr. Scott Linderman, and Dr. Reza Shadmehr. With their support and the tremendous scientific environment at Stanford University, I will gain technical training in two-photon holographic stimulation, extracellular electrophysiology, probabilistic modeling, and conceptual training in cerebellar neuroscience. This training will prepare me well for my long-term goal ...