Presentation Information
[P1-06]Entrainment of periodic neural activity for rhythmic temporal prediction may involve cerebellar learning
*Ken-ichi Okada1, Masaki Tanaka1 (1. Hokkaido Univ. (Japan))
Keywords:
nonhuman primate,rhythm perception
The cerebellum plays a pivotal role in rhythmic movement and rhythmic perception. We previously showed that neurons in the cerebellar dentate nucleus gradually synchronize their activity in response to periodically presented visual stimuli in the absence of movement. Given that the dentate nucleus receives GABAergic projections from Purkinje cells (PCs) in the cerebellar cortex, and the interaction between simple spikes (SSs) and complex spikes (CSs) in PCs is central to cerebellar learning, we examined PC activity to understand how rhythmic neuronal activity is generated. Animals were trained to respond to the omission or color change of isochronically presented visual stimulus, depending on the color of the fixation point. Detection of stimulus omission required temporal prediction, whereas that of color change did not. The periodic activity of 112 well-isolated PCs has been recorded from the crus lobules in 3 monkeys. Neurons were classified into 3 groups based on the time course of SS and CS activities in trials with a 400-ms interstimulus interval. Cluster #1 (32%, n = 36) showed a SS peak around 300 ms following each stimulus and a transient CS for repetitive visual stimulus but not for the omission. Cluster #2 (40%, n = 45) showed an early SS peak and exhibited predictive CS around the time of the repetitive visual stimulus, which was sometimes enhanced following stimulus omission. Cluster #3 (28%, n = 31) showed a clear SS peak, but no evident CS response was observed. In all clusters, the magnitude of periodic SS activity was greatly diminished in the color change condition, indicating that neuronal activity reflects temporal prediction. Importantly, CS in Clusters #1 and 2 also decreased during color detection, indicating that CS occurrence is highly context-dependent. As expected, CS-triggered averaging of SS activity revealed a transient pause in SS in all PCs. Clusters #1 and 2 showed two additional decreases in SS activity, one occurring just before the CS and the other after the stimulus cycle. Contrary to the prevailing negative feedback model of the cerebellum, our results suggest the presence of a positive feedback circuit that amplifies a time-specific decrease in SS activity. This cerebellar learning mechanism may contribute to entrain SS activity to rhythm through the context-dependent occurrence of CSs.