The Nature of Reinforcement in Cerebellar Learning

Abstract

In a now classic study, W. J. Brogden and W. H. Gantt (1942,American Journal of Physiology,119,277–278) demonstrated that movements (limbs, head, eyelid) elicited by direct electrical stimulation of certain regions of the cerebellum (particularly the ansiform lobe) could be trained to respond to neutral auditory or visual conditioned stimuli with appropriate pairing. In recent work we have replicated these results in detail and presented considerable evidence that the reinforcing or “teaching” pathway so activated for the learning of discrete movements is the inferior olive–climbing fiber projection system to the cerebellum. Very strong evidence now indicates that the memory traces for this “skilled response” learning are formed and stored in the cerebellum. The climbing fiber system and inhibitory pathway from the interpositus nucleus to the inferior olive appear to form a neural instantiation of the Resorla–Wagner formulation of classical conditioning and accounts for the “cognitive” phenomenon of blocking. It is concluded that reinforcement in this form of learning is not due simply to contiguity/contingency or to unconditioned stimulus aversiveness, per se, but rather to activation of a particular brain circuit, here the climbing fiber system, a circumstance that may apply to other forms of learning, with other reinforcement circuits, as well.