The Guest Speaker: Anna Ipata, Columbia University, New York
Title: Purkinje cells in crus regions facilitate visuomotor association learning task, by accelerating choice for hand selection.
The process of acquiring and generating goal-directed behavior involves a complex interplay of cortical-subcortical networks, encompassing the cerebral cortex, basal ganglia, frontal cerebrum, to understand goal-directed behavior, investigating visuomotor association learning is pivotal, where specific movements are linked to distinct visual cues. Recent research in non-human primates has underscored the role of the cerebro-cerebellum in supporting visuomotor association learning. Transient inactivation of crus II within the cerebellum significantly disrupts the ability of primates to learn new visuomotor associations, delaying motor responses even under overtrained conditions. The simple spikes (SS) of Purkinje cells (PCs) in the regions signal cognitive errors during the monkeys’ acquisition of associating fractal images with left- or right-hand movements. However, the extent to which the SS responses of individual PCs become selective for hand choices on a trial-by-trial basis as learning progresses, and whether such selectivity shifts involve temporal changes in SS encoding, remain unclear. In this study, we investigated the SS activity of PCs in monkeys within the cerebellar crus regions during an arbitrary visuomotor association task. Individual units display learning- dependent activity patterns in response to the same stimuli, exhibiting heterogeneous correlations with upcoming action choices. As learning advanced, the representation of stimulus-movement associations increased among neurons. At the population level, PCs predominantly encoded upcoming action choices during later learning phases, well in advance of movement onset. Utilizing linear decoding method, we confirmed that the PC population developed an early, learning-dependent representation of both symbols and movements. These findings suggest that PCs in this region may accelerate associative SS responses. This acceleration may contribute to a learned forward internal model, predicting the goal or outcome of selected actions, distinct from the predictive signals found in other cerebellar regions involved in sensorimotor processing.