Abstract:
The electroencephalogram (EEG) is a measure of brain-activity commonly used during clinical general anaesthesia. During surgery, and during the process of returning to consciousness after surgery has finished (the emergence period), EEG measures can indicate the brain-state of the patient, and may predict clinically relevant outcomes. In this observational study, we recorded single-channel frontal EEG from 305 patients undergoing general anaesthesia for surgery, and also during emergence. We developed different EEG measures in the extended alpha frequency range (7-17 Hz), and assessed when oscillatory alpha activity was present during surgery. During emergence, we fitted sigmoid curves to the alpha and delta (0.5-4 Hz) power when plotted against decreasing anaesthetic concentrations. We also measured the onset of muscle-activity (EMG). In a subset of patients, we estimated synaptic parameter values using Bayesian mapping from the EEG to a cortical model. Following emergence, we used a modified CAM-ICU test to assess delirium, and a painscore to assess high-pain. During surgery, alpha oscillations were inexplicably absent in 4% of patients. When present, these oscillations slowed with increased volatile anaesthetic concentration in 88% of patients, whereas alpha power decreased in only 48%. Alpha frequency also slows with increased age. Clinically, the presence of unexpected transient alpha power losses was associated with bodycavity surgery (p<0.001), but not reported high-pain after waking (p=0.081). During emergence, when volatile anaesthetic drug concentrations are decreasing, some patients maintained their alpha or delta power; we have named these phenomena alpha- and deltainertia. Alpha- and delta-inertia occurred in 13% and 21% of patients respectively, and represent a patient being ‘stuck’ in a slow-wave state. Preliminary model mapping results suggest these patients have unchanging synaptic values during emergence. Alpha- and deltainertia were both associated with delirium at 15 minutes post-waking (p=0.006 & p=0.020 respectively), with delirium more common in older patients (p=0.010), and those undergoing long operations (p=0.003). Early EMG activations during emergence reflect the presence of an endotracheal tube in place (p=0.001), but are unrelated to cortical activation. In conclusion, EEG measures during surgery and emergence are predictive of delirium, but not reports of high-pain in the immediate post-operative period.