The impact of rapid transItions in mental workload in a supervisory control setting

Abstract

With improving and expanding automation in domains such as unmanned aerial vehicles, nuclear power plants, and commercial aviation, operators will be expected to handle long periods of low task load while monitoring these highly automated systems with only an occasional need to respond to critical and emergent situations. In most of these situations operators must go from periods of low cognitive engagement to ones of high stress, high mental workload, and limited time to respond. This research investigates the mental workload transition process from low to high workload by measuring the hemodynamic response of subjects to a simulated missile defense task using functional near-infrared spectroscopy, or fNIRS. NIRS is a non-invasive neuroimaging technique that uses specific wavelengths of near-infrared light to measure concentrations of oxygenated and deoxygenated hemoglobin in the prefrontal cortex, the region of the brain commonly associated with "higher level" mental activities. This missile defense simulation consisted of low task loading through system monitoring and two critical events. For these two critical events, participants had to dynamically allocate assets to rapidly respond to incoming missiles in just 100 seconds. The two factors controlled were critical event difficulty (2 levels- easy and hard) and critical event onset time (3 levels - 40, 100, 160 minutes). Thirty subjects participated in the 3-hour experiment. Subjects who received their first event at the middle onset time (100 minutes) had a lower hemodynamic response than those in the early or late groups. Subjects in the "100-minute, hard" condition performed worse than other groups, indicating that the diminished hemodynamic response may be correlated with diminished performance. There was no significant difference for hemodynamic response between the difficulty levels, despite one scenario difficulty level being significantly harder, which was reflective in performance scores. The most important factors to predicting performance in this supervisory control application were NEO Five Factor Index Agreeableness, video game usage, pre-event attention state, and deoxygenated hemoglobin levels. These results indicate that performance can be correlated to the physiological response, and that physiological responses can change over time during periods of very low workload. Additionally, hemodynamic trends correlating with the 30-minute vigilance decrement indicate that physiological changes may be occurring during a vigilance task and that fNIRS may be a suitable method for tracking mental state and engagement over time.