Effect of Sedation on EEG During Deep Brain Stimulation Surgery in Parkinson’s Patients

Mousavi, Mahta; Paulsen, Lisa; Blankertz, Benjamin; Kübler-Weller, Dorothee; Kühn, Andrea A.; Spies, Claudia; Borchers, Friedrich

doi:10.3390/anesthres3020010

Journal Article

DZNE-2026-00780

Effect of Sedation on EEG During Deep Brain Stimulation Surgery in Parkinson’s Patients

Mousavi, M. ; Kübler-Weller, D. ; Paulsen, L. ; Borchers, F. ; Spies, C. ; Kühn, A. A.DZNE* ; Blankertz, B.

2026
MDPI Basel

Anesthesia research 3(2), 10 (2026) [10.3390/anesthres3020010]

This record in other databases:

Please use a persistent id in citations: doi:10.3390/anesthres3020010

Abstract: Background: While providing enough sedatives to avoid pain and trauma during surgery is important, studies show a link between the received sedatives and the development of postoperative delirium (POD). Therefore, predicting POD from clinical or physiological data before or during surgery is highly advantageous. This capability enables healthcare providers to proactively implement necessary measures, thereby mitigating or preventing potential complications. Methods: In this study, we focus on patients with Parkinson’s disease undergoing deep brain stimulation surgery who are particularly susceptible to POD. We investigate what aspects of EEG’s power, functional connectivity and complexity during the course of the surgery are influenced by the amount of sedative. Furthermore, we aim to determine whether and to what extent the recorded brain activity during surgery can serve as a reliable means for the prediction of POD in this group of patients. Results and Conclusions: Our results show significant correlations between various power, connectivity and complexity features of EEG and the amount of sedatives. Even though single EEG features are not significantly different between the two groups who either developed or did not develop POD, we show that a classifier based on support vector machines using the selected EEG features could predict POD. Furthermore, our results provide evidence that a classifier trained only on the amount of sedatives is unable to predict POD. Accompanying this paper, our code is published as an open-source toolbox for the analysis of the EEG signal recorded with the four-channel SEDLine Root system, which is among the widely used EEG systems in operation rooms and its recorded data come with challenges that are addressed in our toolbox.

Classification: