Repurposing electroencephalogram monitoring of general anaesthesia for building biomarkers of brain ageing: an exploratory study

BackgroundElectroencephalography (EEG) is increasingly used for monitoring the depth of general anaesthesia, but EEG data from anaesthesia are rarely reused research. Here, we explored repurposing brain-age modelling using machine learning. We hypothesised that brain age estimated during associated with perioperative risk.MethodsWe reanalysed four-electrode EEGs 323 patients under stable propofol or sevoflurane to study four signatures (95% power <8–13 Hz) prediction: total power, alpha-band (8–13 Hz), spectrum, and spatial patterns in frequency bands. constructed age-prediction models a healthy reference group (ASA 1 2) anaesthesia. Although all were informative, state-of-the-art performance was unlocked by parsing across electrodes along entire spectrum (mean absolute error=8.2 yr; R2=0.65).ResultsClinical exploration ASA 2 revealed positively correlated intraoperative burst suppression, risk factor complications. Surprisingly, negatively suppression higher scores, suggesting hidden confounders. Secondary analyses age-related specific reflected limited model generalisation maintained sevoflurane.ConclusionsAlthough may enable prediction, differences between anaesthetic drugs can impact effectiveness validity models. To unleash dormant potential clinical research, larger datasets heterogeneous populations precisely documented drug dosage will be essential.