EXECUTIVE SUMMARY
BackgroundFatigue is a deadly problem for U.S. Naval Aviation (Naval Safety Center, 2006), and receives a correspondingly large amount of research attention in military RDT&E. Though the basic consequences of fatigue are well known, significant measurement challenges remain in the applied laboratory, where an optimal combination of scientific rigor and operational relevance can be elusive. Completing a flight simulation (FS) while fatigued is used to maximize ecological validity for the flight environment (Caldwell et al., 2003;Russo et al., 2005;Van Dongen, Caldwell, & Caldwell, 2006), but this approach has psychometric shortcomings. The Psychomotor Vigilance Task (PVT) has significant psychometric strength as the gold standard instrument for assessing the cognitive effects of fatigue, but lacks ecological validity for flight. There is a clear need for a fatigue assessment tool combining the operational utility of a flight simulator with the control of the PVT.
PurposeThe purpose of this report is twofold: 1) to describe the development and execution of a flight simulation tool for quantifying vigilance during a fatigue study, and 2) to describe the effort to balance the control and diagnosticity of the PVT with the ecological validity of flight simulation in an inexpensive, offthe-shelf, open-source format.
MethodFifteen active duty military personnel from the Naval Aviation Preflight Indoctrination (API) program at NAS Pensacola volunteered for the study. Subjects completed a battery of neurocognitive and physiological assessments over the course of 25 hours of continual wakefulness. As part of that battery, subjects completed eight trials of a simple flight profile using X-Plane 9 (Laminar Research, Columbia, SC), an inexpensive, off-the-shelf flight simulator. Subjects were instructed to fly "straight and level" at a specified altitude, airspeed and heading (i.e., 2000 ft, 140 knots, due North) for 15 minutes each session. To monitor second-by-second performance on the task, a central data capture server was connected to the four PC-based flight simulation stations. Deviations from the specified flight parameters were monitored by the data capture server. Lapse times were calculated for each parameter as the number of seconds during a simulator trial that subjects deviated from the flight goal by greater than one intraindividual standard deviation (determined at baseline). Flight Simulator Total Lapse Time was the sum of lapse times for each parameter. A series of Visual Basic (VB) programs were then written in Microsoft Excel to calculate descriptive statistics based on those lapses. Results were compared to concurrent subject performance on the PVT.
ResultsSignificant inter-trial correlations for the Flight Simulator Performance Task (FSPT) scores were moderately strong (r = .53 -.81, p < .05), providing initial evidence for test-retest reliability. However, there were some notable non-significant correlations which are discussed in terms of significant, trait-like individual differences...