Towards A Cerebellar Neuroprosthesis: A Simulation and In Vivo Experimental Study

Cerebellum is an important brain region which coordinates muscular control for smooth and accurate motor movement. Damage to the cerebellum, as a result of various medical conditions, impacts everyday life of the patients and there is still no complete cure to the neurological disorders. It has also imposed a significant cost to the healthcare system. With advances in fundamental neuroscience and computational technology, implantable biomimetic neuroprosthesis that are capable of restoring brain function has become possible as a therapeutic tool to neurological disorder. In this project, we propose to develop a realistic cerebellum neuroprosthesis which can operate in real-time. We will also address issues that may contribute to more efficient neuroprosthesis in the future.The ultimate goal of this proposed study is to develop a realistic cerebellar prosthetics based on the cerebellum-mediated eyeblink conditioning. This proposal will first address the system configuration appropriate for developing an efficient neuroprosthesis implant by using a biologically plausible cerebellar computational model. We will test if a neural controller, which adaptively adjusts the output of the neuroprosthesis, can lead to more flexible and more efficient configuration for constructing such neuroprosthetic device. The model will then be programmed into an implantable prosthetic device for real-time in vivo testing. To achieve high-speed computation as well as fast and flexible prototyping of a large-scale neural system, we will implement the model into a FPGA-based chip to be interfaced with a multichannel recording system. The device will be tested on awake rats using traditional eyeblink conditioning protocol. We will examine if cerebellar-dependent motor learning function can be restored by the implant.Furthermore, another important question to address in the current study is to investigate the chronic interaction between the biological neural network and the prosthetic implant. We postulate that the potential neural plasticity induced through chronic interaction with the prosthetic device, particularly in an impaired neural network, has important implication for developing future rehabilitation therapy.This proposed study will demonstrate the feasibility of a realistic cerebellar neuroprosthetics for real-time simulation and in vivo operation. The device will feature its biological plausibility as well as generality for future extension to incorporate other motor tasks. This study will also provide new insight to empower future neuroprosthesis development for rehabilitation purpose.