P 43 Optimizing beta-burst driven adaptive deep brain stimulation for Parkinson's disease

Introduction Adaptive deep brain stimulation (DBS) aims at improving DBS therapy by adjusting stimulation amplitude to patient specific biomarkers tracked in real-time. In Parkinson’s disease (PD), a promising closed-loop approach exploits fast fluctuations of beta power (beta bursts) in subthalamic local field potentials (LFP). With this method, stimulation is applied as soon as beta bursts of a given magnitude and duration are detected. However, closed-loop algorithms offer clinicians a wide range of possible settings. This will further increase the parameter space and thus the complexity of DBS programming, ultimately limiting ease of clinical usability. To increase the translational potential of adaptive DBS, this constraint must be overcome by narrowing down the parameter space towards a clinically optimal combination of stimulation settings. Objectives To explore the parameter space of a beta burst driven adaptive DBS algorithm and to characterize its neurophysiological and behavioral correlates in order to identify a clinically optimal parameter combination. Patients & methods A research platform for adaptive stimulation has been designed comprising a stimulation artifact suppressing amplifier, a custom-made software interface and an external neurostimulator. On this platform, a configurable LFP beta burst driven closed-loop algorithm has been implemented. This setup was used to conduct closed-loop stimulation in PD patients undergoing two-stage-surgery for DBS. Rest recordings were combined with a standardized motor task to quantify stimulation-induced behavioral effects. By systematically changing the minimum beta burst duration upon which stimulation is triggered and the amount of smoothing applied to real-time beta power, the parameter space of this algorithm was explored. Results Stimulation patterns were governed by parameter combination with higher minimum burst duration leading to less frequent stimulation. Beta dynamics varied depending on parameter choice. Clinically optimal settings are yet to be explored. Conclusion The choice of adaptive DBS parameter combination strongly influences stimulation patterns and stimulation-induced neurophysiological responses. Combined recordings of neurophysiological and behavioral adaptive DBS correlates for different parameter combinations may pave the way for guiding parameter choice based on electrophysiological evaluation. This may facilitate the clinical translation of adaptive DBS.