A rule-based expert system for real-time feedback-control in deep brain stimulation

Abstract

Programming in deep brain stimulation (DBS) is often a labour-intensive process. Although automatic closed-loop stimulation has recently been receiving considerable attention, it is still far from clinical settings. Testing in-loop stimulation in a clinical setting is extremely challenging due to manual programming and the lack of synchronisation between stimulation and monitoring devices. In this work, we present a simple rulebased expert system to test feedback-controlled DBS in a clinical setting. The new application operates in closed-loop with the physician as acting person and real-time feedback from an accelerometer. Patients with movement disorders such as in essential tremor announce an individually acceptable level of tremor as a boundary condition for control. As a proof-of-concept, the expert system provides continuous recommendations of stimulation parameters and guides the physician to increase or decrease DBS amplitude by capturing tremor acceleration power on the patients’ forearms. The introduced application considers the technical and practical aspects in a clinical setting. Data obtained from test subjects provide insight into tremor dynamics. We demonstrate the clinical applicability of the rule-based control system for future research focusing on tremor dynamics and inloop stimulation. Finally, a telemetry streaming system could provide the interface for the application of automatic tremor control without the physician as acting person.

Keywords

• movement disorders
• essential tremor
• wearable motion sensors
• rule-based expert systems
• deep brain stimulation