ESTIMATING ALERTNESS FROM THE EEG POWER SPECTRUM

In tasks requiring sustained attention, human alertness varies on a mi nute time scale. This can have serious consequences in occupations ran ging from air traffic control to monitoring of nuclear power plants. C hanges in the electroencephalographic (EEG) power spectrum accompany t hese fluctuations in the level of alertness, as assessed by measuring simultaneous changes in EEG and performance on an auditory monitoring task. By combining power spectrum estimation, principal component anal ysis and artificial neural networks, we show that continuous, accurate , noninvasive, and near real-time estimation of an operator's global l evel of alertness is feasible using EEG measures recorded from as few as two central scalp sites. This demonstration could lead to a practic al system for noninvasive monitoring of the cognitive state of human o perators in attention-critical settings.