Translation of EEG-based performance prediction models to rapid serial visual presentation tasks

Abstract

Brain wave activity is known to correlate with decrements in behavior brought on by fatigue, boredom or low levels of alertness. Being able to predict these behavioral changes from the neural activity via electroencephalography (EEG) is an area of ongoing interest. In this study we used an established approach to predict time-on-task decrements in behavior for both a realistic driving simulator and a difficult perceptual discrimination task, utilized in many brain-computer interface applications. The goal was to quantify how well EEG-based models of behavior, developed for a driving paradigm, extend to this non-driving task. Similar to previous studies, we were able to predict time-on-task behavioral effects from the EEG power spectrum for a number of participants in both the driving and perception tasks. © 2013 Springer-Verlag Berlin Heidelberg.