Computational Modeling of the Effects of Sleep Deprivation on the Vigilance Decrement.

Abstract

We developed a computational model of the effects of sleep deprivation on the vigilance decrement by employing the methods of system dynamics modeling. Situations that require sustained attention for a prolonged duration can cause a decline in cognitive performance, the so-called vigilance decrement. One factor that should influence the vigilance decrement is fatigue in the form of sleep deprivation. We employed the methods of system dynamics modeling (numerical-integration techniques for modeling complex feedback systems) to create a computational model of the vigilance decrement. We then simulated the computational effects of sleep deprivation on the behavior of that model, using empirical data obtained from the literature for calibrating such effects. Sleep deprivation of 2 hr over a 14-day period should produce an additional decline of 9% in detection performance over that found with the typical vigilance decrement, whereas 4 hr of sleep deprivation over 14 days should produce an additional decline of 14% in detection performance. With respect to dual-process theory, it is through its deleterious effects on analytical cognition that sleep deprivation should impact the vigilance decrement. Such computational modeling may be advantageous for human-machine teaming by theoretically allowing a future autonomous software agent to anticipate the decline of human performance and compensate accordingly.