The predictive cerebellum: a universal operating principle across both cognitive and social domains

This doctoral dissertation investigates the cerebellum’s role as a supramodal predictive hub, with particular emphasis on its contributions to cognitive and social processing. Over the last decades, cerebellar involvement in cognitive and social functions has been increasingly recognized, but the precise computational mechanisms underlying its contributions across these diverse domains remain incompletely understood. The sequence detection theory refers to a proposed computational mechanism of the cerebellum, namely the detection and simulation of repetitive temporal and spatial patterns across motor, perceptual, and cognitive domains. This mechanism enables the cerebellum to construct internal models of sequenced events, which can be used to generate predictions about future outcomes and to signal errors when these predictions are violated. Accordingly, predictive processing can be seen as a continuation and extension of the cerebellum’s operational mode: it is tightly grounded in sequence processing capacities, yet it represents a further functional step that builds upon them. However, the specific neural substrates supporting these predictive operations, their reorganization in pathological conditions, and their developmental trajectories across cognitive and social domains require systematic investigation. The central hypothesis of this thesis is that the cerebellum integrates prior experience with incoming sensory information to generate predictive models that optimise behaviour across multiple domains, including motor planning, cognitive task execution, and social interaction. In line with this unified predictive processing framework, this work employs an integrated multidisciplinary approach combining structural, microstructural, and functional neuroimaging methodologies with advanced experimental paradigms to investigate cerebellar contributions to both cognitive efficiency and social cognition, and how these mechanisms are maintained or reorganized across development and in pathological conditions. Following two introductory chapters establishing the anatomical, functional, and theoretical foundations supporting this thesis, the experimental work is organized into two principal sections. The first section focuses on the cerebellar role in cognition and comprises two complementary studies. The first characterizes the macrostructural and microstructural neural substrates of cognitive efficiency in healthy individuals, whereas the second examines compensatory cerebello-cerebral network reorganization in patients with spinocerebellar ataxia type 2, elucidating mechanisms of cognitive reserve that sustain function despite progressive structural degeneration. The second experimental section presents three interconnected studies exploring predictive mechanisms in social processing. The first study investigates sequencing and prediction across social and non-social domains in remitted bipolar disorder, together with a concurrent assessment of cerebellar structural integrity. In bipolar disorder, both social cognition impairments and cerebellar structural and functional abnormalities have been consistently reported. Building on this evidence, the study aims to determine whether social cognition deficits in this population stem from disrupted predictive processing mechanisms resulting from cerebellar alterations. The second study employs immersive virtual reality and computerized social predictive paradigms to examine how contextual priors modulate social prediction across developmental stages in healthy participants. This approach provides a normative framework of how predictive mechanisms mature and operate across the lifespan in social contexts, establishing the necessary baseline against which pathological alterations can be compared. The third study applies analogous context-based predictive tasks alongside comprehensive social cognition assessments in cerebellar degenerative patients, investigating parallel potential cerebellar grey matter volumetric reductions. By assessing social cognition at different levels of predictive complexity using tailored tests, this study determines whether cerebellar degeneration impairs predictive processing in social cognition and whether contextual information influences performance as it does in real-world social interactions. Collectively, these investigations aim to elucidate the cerebellum’s function as a domain-general predictive processor, clarifying shared principles underlying cognitive and social operations, and demonstrating how this predictive architecture supports adaptive behaviour across typical development, healthy aging, and neurological or psychiatric pathology.