ABSTRACT
EEG spectral power has been shown to correlate with level of arousal and alertness in humans. In this paper, we assess its usefulness in the detection of behavioral microsleeps (BMs). Eight non-sleep-deprived normal subjects performed two 1-hour sessions of a continuous tracking task while EEG and facial video were recorded. BMs were identified independent of tracking performance by a human rater by viewing the video recordings. Spectral power, normalized spectral power, and power ratios in the standard EEG bands were calculated using the Burg method on 16 bipolar derivations to form an EEG feature matrix. PCA was used to reduce the dimensionality of the feature matrix and linear discriminant analysis used to form a classifier for each subject. The 8 classifiers were combined using stacked generalization to create an overall detection model and N-fold cross-validation used to determine its performance (Phi=0.30 +/- 0.05, mean +/- SE). While modest, the detection of BMs at such a high temporal resolution (1 s) has not been achieved previously other than by our group.
