Quantitative relationship between anteriorization of alpha oscillations and level of general anesthesia.

ABSTRACT

A typical electroencephalogram (EEG) change induced by general anesthesia is anteriorization-disappearance of occipital alpha oscillations followed by the development of frontal alpha oscillations. Investigating the quantitative relationship between such a specific EEG change and the level of anesthesia has academic and clinical importance. We quantified the degree of anteriorization and investigated its detailed relationship with the level of anesthesia. We acquired 21-electrode EEG data and bispectral index (BIS) values of 50 patients undergoing surgery from before anesthesia induction until after patient arousal. For each epoch of a 10.24-s window with 1-s offsets, we calculated frontal alpha power [Formula see text], occipital alpha power [Formula see text], and their difference [Formula see text] to quantify anteriorization. We calculated Spearman's rank correlation coefficients between these values and the BIS value. We used locally weighted regression to estimate [Formula see text], [Formula see text], [Formula see text], [Formula see text], [Formula see text], and [Formula see text] at each BIS value. Thirty-six patients (26 females and 10 males aged 24-85 years) were analyzed. The 95% confidence intervals for the mean of Fisher transformations of Spearman's rank correlation coefficients between [Formula see text], [Formula see text], and [Formula see text] and BIS value were [- 0.68, - 0.26], [0.02, 0.62], and [- 1.11, - 0.91], respectively. The change in [Formula see text] and [Formula see text] with BIS value showed different patterns by the type of anesthetic agent, whereas that in [Formula see text] was more consistent with smaller individual variance. Anteriorization, quantified by the difference between frontal and occipital alpha powers, continuously developed in conjunction with general anesthesia. Quantifying anteriorization may provide an objective indicator of the level of anesthesia.