Synaptic basis of predictive coding in the cerebellar cortex – SynPredict

An essential feature of the brain, for its interaction with the world, is its ability to predict immediate future and continuously update internal representations. The cerebellum is often viewed as a structure in which predictions can be efficiently generated and updated for use by other brain regions. Nevertheless, the synaptic, neuronal and circuit mechanisms underlying the ability of the cerebellum to perform multimodal and temporal integration required for well timed sensory prediction is not known. Deficits in such computations are thought to contribute to symptoms of psychiatric diseases such as schizophrenia and autism spectrum disorders. The input layer of the cerebellar cortex is thought to perform sparsification of information arising from internal and external sensory stimuli in order to elaborate a representation of the rich temporal features of contextual stimuli. Recently Partner 1 identified a novel cellular mechanism that further enhances feature representation in GCs, through multisensory innervation of single granule cells and temporal firing signatures of different sensory inputs. The proposed research consortium will combine cellular, computational, systems and behavioral neuroscience approaches to examine how synaptic diversity and dynamics are used by the cerebellum to predict sensory consequences of behavior. We hypothesize that the diversity of synaptic function at the input layer of the cerebellum is critical for encoding temporal features of auditory, somatosensory stimuli, cortical and higher cortical information necessary for the prediction of expected sensory stimuli, and thus allows the cerebellum to detect mismatched internal and external sensory signals.

This study is ground breaking as few studies have identified quantitatively a role for synaptic dynamics in circuit processing and behavior. Moreover, we will provide a novel mechanism to resolve an age-old question of how the cerebellum encodes precise timing. Finally, as deficits in prediction within the cerebellum may underlie psychiatric diseases, we will provide the basic research groundwork for understanding how deficits in multimodal and temporal integration might contribute to autism spectrum disorders and schizophrenia.