Windows into the mind: From neurophysiological to multimodal measurements in the cockpit

Increasing automation in modern aircraft is intended to alleviate the physical and cognitive demands put on the pilots and thereby to reduce human error. With new technologies available, it might be possible to design adaptive assistance systems that offer support tailored to the situation and the pilots’ needs. In order to do so, however, the systems need information about the pilots’ state and situation. Neurophysiological measurements like electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) seem a powerful tool for enabling adaptive automation by monitoring pilots’ cognitive states. Such measurements need to be reliable and valid in order to gain interpretable and robust results. Our research, however, suggests that this may not be as easy as it seems. In this paper, we give a short overview of the methodological approach of our research and results, identify problems and research gaps concerning the validity and reliability of neurophysiological assessments and discuss possible mitigations to make these measurements the robust and dependable sources of information we wish for.