In this BRIDGE project, we aim to develop and evaluate the feasibility of a new closed-loop Deep Brain Stimulation (DBS) control system that uses electrophysiological recordings in the brain, together with behavioral and physiological signal recordings, to minimize Parkinson’s Disease (PD) symptoms and maximize patients’ quality of life (QoL). To do this, we will use multimodal sensors and machine learning (ML) algorithms to recognize motor and non-motor symptoms such as emotion (e.g. mania, depression), neuropsychiatric symptoms (e.g. apathy, impulse control disorders), motor symptoms (e.g. tremor, bradykinesia) and cognitive abilities (e.g. memory, processing speed, bradyphrenia). These will, together with the recorded neural oscillations (e.g. excessive/prolonged beta oscillations or lack of gamma oscillations as a biomarker of akinesia and rigidity), serve as input to a heuristic controller of the DBS neurostimulator parameters. The control objective will be to minimize symptoms in the aforementioned four domains. Sensor-based quantification of PD motor symptoms is a well-studied problem, but there are yet no established methods for unobtrusive sensor-based monitoring of emotion, behavior, and cognition in PD. In this project, we propose to analyze continuous multimodal recordings of speech (with a microphone), activities and sleep (with ambient sensors and a bed sensor), body movements and cardiac activity (with a wearable sensor), and discrete measurements of cognitive performance (with a tablet- or smartphone-based computer game).
Temporal fluctuations of non-motor symptoms, including emotional disturbances, neuropsychiatric symptoms, as well as cognitive performances are eminent for the patients’ QoL, but at the same time, these symptoms are difficult to observe and often under-diagnosed in PD. The idea of using emotion detection in spontaneous speech, and leveraging this new item with context information from other sensor modalities (e.g. ambient and wearable sensors), will open a novel approach to better understand the detection and management of non-motor symptoms in PD. This high-risk-high-gain project will, in the best case, result in a working prototype of a closed-loop DBS system that will reduce both motor and non-motor symptoms and will significantly improve the QoL of PD patients. Even if not successful, e.g. patients do not benefit from the new adaptive controller, we will still learn a lot about fluctuations of non-motor symptoms and their biomarkers in PD, which will help us to better treat future PD patients.
Besides PD patients, there is also an enormous potential benefit for psychiatric patients: “The world is poised for a revolution in psychiatry, with psychosurgery increasingly used as an interventional tool. Today, DBS is totally experimental in psychiatry, but eventually, it will become a treatment” (Krack P. in the Lancet Neurology, 2014). With this population, biomarkers for emotional and behavioral aspects, such as those developed in this project, will be of great clinical relevance highlighting the potential benefits of this project beyond PD.
If successful, this BRIDGE project will open new business opportunities for telemonitoring and Neurostimulation companies in Switzerland.