Ruolo del cervelletto nella distonia: studio attraverso un modello di rete neurale cerebellare spiking durante l'apprendimento associativo

Although traditionally associated with dysfunctional basal ganglia, dystonia is now commonly regarded as a motor network disorder involving multiple parts of the brain. One of them is the cerebellum. Animal and human studies of dystonia have reported cerebellar abnormalities, however, it is unclear how this brain area could participate in the disorder. In the present work, functional role of the cerebellum in dystonia was explored using a realistic cerebellar spiking neural network model to simulate eyeblink classical conditioning. The model included multiple direct and indirect loops involving the cerebellar cortex, the inferior olive and the deep cerebellar nuclei, with synaptic plasticity at parallel fiber to Purkinje cell synapses driving cerebellar learning. The model, which produced typical learning behaviors of conditioned eyeblink response acquisition, was modified aiming to induce cerebellar abnormalities reported in animal models of dystonia. The results showed that reduced inferior olivary input to Purkinje cells led to impaired learning with insufficient modulation of Purkinje cell and deep cerebellar nuclei activity as well as fewer conditioned responses. Purkinje cell aberrant burst-firing, induced by directly stimulating Purkinje cells within the model, also showed compromised conditioned eyeblink response acquisition due to improper timing of deep cerebellar nuclei responses. No acquisition of conditioned responses was present when Purkinje cell burst-firing was induced by applying a repetitive intermittent stimulation to the inferior olive, with the model producing excessive output and, therefore, not a time-locked activity modulation. On the contrary, partial imbalances in Purkinje cell afferent connections from molecular layer interneurons, parallel fibers and climbing fibers did not compromise the learning process significantly, highlighting a certain robustness to structural alterations. Overall, the results suggest that cerebellar abnormalities reported in animal models of dystonia, namely, reduced olivocerebellar input and Purkinje cell aberrant burst-firing, could result in impaired eyeblink classical conditioning. The model predicts abnormal cerebellar motor control and learning in dystonia. The current work shows that realistic modeling could be a powerful tool in predicting causal mechanisms in different types of dystonia, illustrating how certain levels of alterations in specific cerebellar functional or structural aspects can lead to quantified impairments in cerebellar learning. Keywords: dystonia, cerebellum, motor control, motor learning, eyeblink classical conditioning