Single threshold adaptive deep brain stimulation in Parkinson's disease depends on parameter selection, movement state and controllability of subthalamic beta activity.

BACKGROUND: Deep brain stimulation (DBS) is an invasive treatment option for patients with Parkinson's disease. Recently, adaptive DBS (aDBS) systems have been developed, which adjust stimulation timing and amplitude in real-time. However, it is unknown how changes in parameters, movement states and the controllability of subthalamic beta activity affect aDBS performance.

OBJECTIVE: To characterize how parameter choice, movement state and controllability interactively affect the electrophysiological and behavioral response to single threshold aDBS.

METHODS: We recorded subthalamic local field potentials in 12 patients with Parkinson's disease receiving single threshold aDBS in the acute post-operative state. We investigated changes in two aDBS parameters: the onset time and the smoothing of real-time beta power. Electrophysiological patterns and motor performance were assessed while patients were at rest and during a simple motor task. We further studied the impact of controllability on aDBS performance by comparing patients with and without beta power modulation during continuous stimulation.

RESULTS: Our findings reveal that changes in the onset time control the extent of beta power suppression achievable with single threshold adaptive stimulation during rest. Behavioral data indicate that only specific parameter combinations yield a beneficial effect of single threshold aDBS. During movement, action induced beta power suppression reduces the responsivity of the closed loop algorithm. We further demonstrate that controllability of beta power is a prerequisite for effective parameter dependent modulation of subthalamic beta activity.

CONCLUSION: Our results highlight the interaction between single threshold aDBS parameter selection, movement state and controllability in driving subthalamic beta activity and motor performance. By this means, we identify directions for the further development of closed-loop DBS algorithms.