Electroencephalogram spectral power changes signalling the loss of volitional control induced by propofol: a secondary analysis of an observational study in spinal surgery patients.

PURPOSE: There are uncertainties about the timing of electroencephalogram (EEG) changes relative to loss of consciousness during induction of anesthesia. We sought to evaluate EEG spectral power changes with segmented linear regression to objectively identify those that most reliably signal loss of volitional control (LVC), a reliable marker of unconsciousness.

METHODS

In this secondary analysis of a previous observational study, we analysed EEG data from 12 spinal surgery patients and 11 controls. The EEG was recorded during induction with propofol (1.0 mg·kg-1·min-1 iv) while the study participants were holding a 0.5-kg cylinder in their hand. The drop of the cylinder marked LVC. To objectively identify meaningful EEG changes for each patient, we used structural breakpoint analysis. We then used linear regression to assess the relationship between the breakpoints and LVC. We characterized dynamics of the frequency band changes with the use of coherence analysis (Rayleigh test).

RESULTS

After an initial increase in beta (18-25 Hz) power 145 sec before LVC, the increase in delta (1-4 Hz) power and the decrease in beta power were linked, with all patients showing an inverse relationship between these power bands (P < 0.001). There was also a reliable correlation between the time of LVC and the breakpoints for the delta power plateau (R2 = 0.85; P < 0.001) plus beta (R2 = 0.66; P < 0.05) and gamma (R2 = 0.79; P < 0.01) power nadirs.

CONCLUSIONS

The present observations with high temporal resolution show that an inverse relationship between beta and delta EEG power bands signals LVC induced by propofol and that a robust association exists between LVC and the delta power maximum. These findings likely reflect changes in thalamocortical network dynamics that could account for anesthetic-induced loss of consciousness.