The role of the cerebello-parietal pathway in state estimation

• Applicant: Professor Dr. Cornelius Schwarz
• Subject Area: Cognitive, Systems and Behavioural Neurobiology
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 464398405

A dominant theory in the field of sensorimotor integration is that the sensory signal flow is integrated with internal predictions about sensory consequences of ongoing movements and other contextual parameters. Such predictions are thought to be generated by an internal model based on an efference copy and other contextual signals. The site of the internal model is assumed to be the cerebellum. The optimal integration of predictive signals with sensory feedback to yield the sensory prediction error has been called state estimation. Although state estimation is a classical concept, its site of computation is not known. One problem has been the anatomical complexities of cerebellar output pathways. Another one is that state estimation may not be unique to cerebellar circuits. For instance, mammalian neocortex is almost certain to provide predicative signals as well. Such non-cerebellar predictive signals may realize different types of behavioral control, perhaps playing out on different time scales and/or related to different learning systems. Little is known about the interplay of different types of predictive signals. In this project, we will focus on cerebellar state estimation, which is most directly related to movement generation. The outputs of the cerebellum massively reach the sensorimotor cortex and we will systematically disentangle cerebellar predictive signals from assumed neocortical ones. We will track prediction signals on the complex cerebello-parietal pathway using multi-neuron extracellular electrophysiology, identify their cerebellar origin using causal, optogenetic analysis, and study their interplay with known parietal predictive signals of presumed cortical provenience. We established a novel behavioral tool to realize such analysis in extensive preliminary work. It translates a classic approach, the opening of the reafferent loop pioneered by Curtis Bell working in weakly electric fish in the 1980ies, to the use in behaviorally trained mice that perform defined whisker movements. This analysis will allow us to disentangle sensory prediction, and state estimation signals and localize them within the branched output stream of signals originating in the cerebellum and projecting to the neocortex. Starting from the cerebellar output we will parse the signals in motor thalamus, sensorimotor higher order thalamus and tactile thalamus. Finally we will monitor signals in primary/secondary somatosensory cortex and neighboring posterior parietal area.

• DFG Programme: Research Grants