Thinking and behaviour during adaptive brain stimulation

Primary outcome: To meet the primary objective of establishing an experimental setup to develop patient specific aDBS with focus on cognitive capabilities, the following endpoints will be acquired: Task performance in the Stroop task under different DBS conditions. The performance is considered as a proxy for cognitive symptom severity. Whether continuous DBS (DBS ON) modulates the reaction time compared to the non-stimulated state (DBS OFF), will be evaluated by comparing reaction times within individual patients. Decoding correlation scores of biomarkers as predictors for Stroop task performance. We will use a machine learning pipeline with task performance as regression labels to extract informative and controllable biomarkers. The quality of these markers will be reported as Pearson*s correlation between predictions based on biomarkers versus the actual observed task performance. Models are trained from data of individual sessions and are evaluated using a chronological cross validation scheme. Correlations will be considered within a single session, i.e. within a single subject and day. Secondary outcome: If biomarkers with significant correlation to the Stroop task performance are found, they will be used to control DBS in a closed-loop setup, i.e., aDBS, leading to behavioural scores in another stimulation condition. Please note that we do not expect informative and controllable biomarkers to be present for every session. This is based on our previous experience with the CopyDraw task (Castaño-Candamil, Piroth, et al., 2020). We will report task performance under the different stimulation conditions on a per session basis, i.e. a comparison between stimulation OFF, ON (open-loop), and closed-loop task repetitions if informative and controllable biomarkers were found for a given session. To establish technical suitability of the setup at the MUMC+, we will replicate the CopyDraw task that was previously used at the UMC Freiburg.