Approaches to understanding the role of the cerebellum in sensorimotor control and learning: the big picture of the little brain

Abstract

It is well-established that the damage or removal of the cerebellar cortex not only leads to severe deficits in fine control and coordination of movements but also impairs the ability to correct the motor behavior, thus pointing towards the role of the cerebellum in sensorimotor control and learning. However, the understanding of how this is achieved by the cerebellum remains unclear. This dissertation offers different approaches that try to elucidate the cerebellar mechanisms underlying motor control and learning, while also tackling an important methodological barrier associated with these questions. The first approach (Appendix 1) thoroughly investigates different kinematic variables of fast and precise index finger movements in cerebellar patients and healthy controls. Here I demonstrate that the increased end-point variability—"motor noise”—observed in cerebellar patients is a direct consequence of the loss of a cerebellum-based velocity-duration trade-off mechanism that consistently adjusts movement duration by utilizing the information on the expected velocity of the upcoming movement. Understanding the neural underpinnings of these cerebellar mechanisms requires direct access to the core element of the canonical cerebellar microcircuitry—the Purkinje cell. One of the most intriguing and debated features of the Purkinje cell discharge, the only output of the cerebellar cortex, is the complex spikes. However, owing to their complex wave morphology accompanied by extremely low firing rates, detecting these events is an immense challenge. The second approach (Appendix 2) in this dissertation deals with the long-standing problem of complex spike detection, by offering a fast and reliable, fully automated deep-learning based algorithm that not only detects these events with impeccable accuracy but also provides useful information on their duration. The direct benefits of this approach form the basis of the third approach (Appendix 3), which specifically deals with the role of complex spikes recorded in a repetitive saccade paradigm. Different findings, based on highly specific paradigms, have led to divergent views on the role of these events. In this approach, I demonstrate that in addition to conveying error-related information, as well as information on the metrics of both primary and corrective saccades, complex spike activity also seems to predict the upcoming events. Furthermore, complex spikes convey this information in a time-specific manner, with changes in complex spike firing probability paralleled by changes in their duration. Hence, complex spikes are fully capable of conveying a vast spectrum of behaviorally relevant information in a multiplexed manner, all together in one task. These findings are compatible with both classical and non-classical roles of complex spikes, thereby proposing a broader framework for their roles. Lastly, I present an overall picture of the role of the cerebellum in the form of a review (Appendix 4) that focuses on the oculomotor system as a model for motor control and learning. In conclusion, the thesis discusses the implications of the results obtained for future work on our view of the cerebellum.