ISTC Seminar-Closed -Loop Simulation for Parkinson's Disease

ISTC Seminar

Dr. Sridevi Sarma, Dept of EECS, MIT

Title:  Towards a Closed Loop Deep Brain Simulation for Parkinson's Disease

Abstract:  Parkinson’s disease (PD) is a chronic progressive neurological disorder that affects millions of people worldwide causing movement disorders. There is no treatment to stop disease progression, however, a highly promising therapy is deep brain stimulation(DBS). Many neuropathologies are now being treated using DBS. However, to date, DBS has relied on open loop stimulation, with parameters of the stimulus pulse train (eg. pulse width, frequency, amplitude) set through an extensive blind search for each patient. It takes several months to find these parameters from millions of alternatives, and when found, they are likely sub-optimal and remain constant for years. In order to eliminate calibration and increase effectiveness, we propose to develop a low-powered closed-loop DBS system. Such a project entails 1) developing electrodes and electronics to sense discharge patterns of local neurons and to actuate neurons with stimulation, 2) computing mathematical models of diseased and healthy neuronal activity from neurophysiological data collected from PD patients and healthy primates (healthy human surrogates); and, 3) designing a dynamic feedback controller that, based on model predictions, automatically chooses appropriate stimulation patterns to generate healthy neuronal spiking activity.

In this talk, we present preliminary results on data-driven modeling of diseased and healthy neuronal activity in the sub-thalamic nucleus (typical DBS target for PD) and a simple control algorithm based on these models. When we applied our controller on test data for 7 PD patients, we were able to generate comparable motor performance to typical programmed DBS settings with signals that generate 4 times less power.