The effects of day-to-day variability of physiological data on operator functional state classification

► Psychophysiological data exhibits variability or nonstationarity across days. ► We applied pattern classification to EEG data in order to monitor mental workload. ► Accuracy was low across days, but improved with multiple days in the training set. ► Including small amounts of data from a new day boosted accuracy as well. ► Either method produces 80% + classification accuracy, adequate for many applications.