A Computational Model of the Cerebellum to Simulate Cortical Degeneration During a Pavlovian Associative Paradigm

Abstract

Thanks to the development of realistic neural models, computational neuroscience is becoming fundamental to understand the complex functioning of the human brain. Several studies have demonstrated the role of computational models in revealing the links between the low-level activity at multiple neural scales and human-like behaviors, under physiological conditions. However, limited resources have been employed in applying computational neuroscience to the understanding of brain diseases. In this work, we focused on the cerebellum that has a crucial role in driving learning and timing of motor commands during associative tasks. We provided a preliminary analysis on the use of a computational model to investigate a type of cerebellar ataxia that involves cortical degeneration. Starting from an optimized realistic Spiking Neural Network representing the cerebellar microcircuit, we applied specific local modifications to mimic the same damage as in patients. Then, we analyzed the behavior generated by the low-level neural activity of the network during a Pavlovian associative task. We were able to reproduce the same impaired behavior shown in human tests, characterized by a decreased capability to associate the provided stimuli. Moreover, the changes in the simulated neural activity agreed with the findings from neural recordings in animals affected by the same pathological condition. Therefore, we demonstrated that our model could be used also to represent cerebellar pathologies, with the possibility to link damages at neural level to compromised behaviors, which is one of the main actual challenges in neuroscience.