PNRR PNC Fit4MedRob
“Dual-Cereb-Control” aims at developing an efficient neuro-inspired controller for sensorimotor learning and control. The computational architecture will embed a realistic spiking model of the cerebellum, with advanced physiologically-inspired plasticity rules, into a network of interconnected functional blocks, associated to functionally-related brain areas. Specifically, two cerebellar modules will define a forward and an inverse model (tandem configuration). This system controller will be embodied into a human-like robotic platform, targeting task-dependent degrees of freedom. It will be used to explore pathological states in patients with movement disorders involving the cerebellar loops, monitoring both internal parameters and external behavior. Indeed, patients with cerebellar diseases show behavioral impairment consistent with tandemly arranged internal models. These models can be used to guide therapeutics and rehabilitation.