Dimitar Kostadinov: Organization of climbing fiber inputs to Purkinje cells during behavior

The cerebellum facilitates smooth motor execution and learning by constructing internal models that link sensation to action. In addition to this canonical function, there is increasing evidence for cerebellar contributions to cognitive processes, such as reward-related computations. My aim is to understand how sensory, motor, and higher-order parameters are encoded by populations of cerebellar neurons, and how these signals emerge and change over the course of learning. To address these questions, I have used chronic two-photon calcium imaging and Neuropixels probes to record activity from populations of cerebellar neurons in mice learning and executing a virtual-reality based sensorimotor integration task to obtain a delayed reward, with a specific focus on climbing fiber inputs to Purkinje cells, the primary drivers of learning in the cerebellum. Clustering of spontaneous climbing fiber signals revealed the microzonal organization of these inputs to Purkinje cells. In trained mice, microzones can be activated by sensory stimuli, by movements, and after reward delivery, and individual microzones were generally modulated by combinations of these parameters. Furthermore, climbing fiber inputs to Purkinje cells are dynamic across learning and contextually modulated, demonstrating that the instructive signals themselves are refined with learning. Finally, the prominence of reward-related activity in climbing fibers extends the role of these inputs beyond sensorimotor error corrections to reward-based associative learning. Thus, climbing fiber inputs convey multimodal signals to Purkinje cell populations that are well-suited to drive learning of all variables that may assist animals in executing goal-directed behaviors.

Dimitar Kostadinov received his PhD from Harvard University, where he studied the molecular mechanisms and functions of dendritic self-avoidance in the retina under Josh Sanes. He then moved to Michael Häusser’s lab at UCL for postdoctoral training, where he has been studying cerebellar dynamics during learning and execution of goal-directed behaviour. His graduate and postdoctoral work was funded by NIH and EMBO fellowships. This summer, he will be opening his own lab at King’s College London in the Centre for Developmental Neurobiology funded by a Wellcome Trust Career Development Award.