Computational Investigation of Adaptive Deep Brain Stimulation

Abstract

Deep Brain Stimulation of the sub-thalamic nucleus (STN) has been proven to be effective at reducing symptoms of patients with Parkinson’s disease (PD). Currently an implanted pulse generator provides chronic electrical stimulation to the STN via an electrode and the stimulation parameters are chosen heuristically. Closed-loop Deep Brain Stimulation (DBS) has been proposed as an improvement to this, utilising neural signal feedback to select stimulation parameters, adjust the duration of stimulation and achieve better patient outcomes more efficiently. In this research, potential neural feedback signals were investigated using a computational simulation of the basal ganglia. It was found that the interspike-interval in the globus pallidus externus provided a possible metric for ‘on’ and ‘off’ states in Parkinson’s disease. This parameter was subsequently implemented as neural feedback in an adaptive closed-loop DBS simulation and was shown to be effective. In particular, the thalamic relaying capability was evaluated using an Error Index (EI) and the adaptive DBS was found to reduce the EI to 2%, which compared with 20% for the PD case without DBS. This was achieved using 58% of the stimulation time used during continuous DBS, indicating a large reduction in DBS energy requirements. This selection and implementation of a potential neural feedback parameter will assist in developing improved implanted DBS pulse generators.