The EEG as a monitor of midazolam amnesia: changes in power and topography as a function of amnesic state.

In order to identify EEG parameters that might be specific for identifying amnesia during midazolam infusion, we examined changes in the EEG power spectrum associated with a period of amnesia, determined by inability to recall a sequence of numbers and objects presented verbally, after intravenous midazolam 0.07 mg/kg in ten normal volunteers. Measurements were taken at baseline, during infusion immediately before and after the onset of amnesia, immediately at end of infusion, and 0.5 and 1.5 h after infusion. All subjects had onset of amnesia during infusion, were completely amnesic by the end of infusion, partially amnesic 0.5 h after infusion, and had complete recall by 1.5 h after infusion. The EEG beta power increased and alpha power decreased during amnesic periods. The beta 1/alpha power ratio was the parameter most specific for amnesia. From a baseline value of 0.20 +/- 0.05 (standard error of the mean [SEM]), it increased to 0.96 +/- 0.26 at the end of infusion and decreased to 0.61 +/- 0.15 0.5 h after infusion. By 1.5 h after infusion, all EEG parameters had returned to baseline values. Beta power changes associated with midazolam amnesia were most pronounced in the Fz and Cz lead positions, and alpha power changes were most pronounced in the Oz position. We conclude that 1) EEG power values, particularly the beta 1/alpha ratio, can identify periods of amnesia after midazolam infusion; 2) specific EEG changes and the presence of amnesia vary with the probable serum concentration of midazolam; and 3) the characteristic EEG pattern during partial or complete amnesia varies as one moves across the cerebral cortex.