A multiarea multiscale neurorobotic controller for Parkinson’s disease study

Parkinson’s disease (PD) is the second most common neurodegenerative disorder and has been widely investigated in the last decades. However, an exhaustive knowledge of PD pathogenesis is still lacking. For instance, there is growing evidence of the involvement of both basal ganglia (BGs) and cerebellum in the altered Parkinsonian brain activity. To investigate this interplay, we developed a model of rat BGs which includes a parameter to vary dopamine depletion levels. We then added a spatially detailed scaffold model of the cerebellar microcircuit. Both models are composed by spiking neural networks embedding point neuron models with high physiological resemblance. We connected them by including a mass model of the cortex and the thalamus. To handle this multi-scale level of resolution, we developed and tested a custom software tool, written in Python. Finally, we investigated if the cortical drive obtained from the model could be applied to perform a motor task with a cerebellar controller. The resulting model can replicate physiological firing rates observed experimentally in healthy rats. Further, reducing the amount of dopamine level in the BGs model increases the inhibitory output to the other areas, causing an overall reduction of activity in the functional loops and the emergence of beta-bands synchronous oscillations. The cerebellar controller succeeded in accomplishing the task only if case of limited dopamine depletion level