Using Eye and Head Based Psychophysiological Cues to Enhance Screener Vigilance

2011
Brent Langhals
Position: 
Assistant Professor, LT COL, Air Force Institute of Technology
Abstract: 
The objective of this study was to determine if eye and head based psychophysiological cues can be used to maintain screener vigilance levels during long duration tasks. In two parts, this study first explored which cues are predictive of vigilance levels. The second part of the study developed a vigilance feedback system using the cues established in part one and compared the results of subjects using the feedback system with those subjects having received no feedback or those who received random feedback. In part one, 48 participants reviewed 600 simulated x-rays to determine if any hazardous items (guns or knives) were present. Individual vigilance levels were determined by scoring the number of correct detections during eight 5-minute periods (total study length = 40 minutes). Using an eye-tracking machine, four concurrent eye and head activity measures (blinks, saccades, pupil diameter, and head position) were used to model changes in vigilance level throughout a simulated baggage screening task. At the end of the study, changes in blink and saccade rates proved to be significant predictors of an individual's ability to detect the presence of these hazardous items among other non-significant baggage items. Part two required 126 participants equally distributed across three conditions to repeat the same screening task. For one condition the monitoring system, instead of passively recording the individual's cues, provided near real-time feedback of vigilance levels to a condition. Participants in other conditions received random feedback on their vigilance levels while a third group received no feedback. At the conclusion of the study the subjects who received the real-time feedback performed significantly better than those who received no feedback. However, they did not perform better than the subjects who received random feedback. Perhaps more significantly, the subjects who received random feedback, while performing better than the no feedback group, also experienced a significantly higher number of false detects. The results of this study indicate a vigilance feedback system based upon subject psychophysiological cues may be an effective method to maintain attention levels during long duration vigilance tasks while preventing a corresponding increase in error detection rates.