Page:Risk of performance errors due to sleep loss, circadian desynchronization, fatigue, and work overload.pdf/20

recent studies have documented large stable (trait-like) differences among individuals in the degree of cognitive deficits that are experienced during sleep loss (Van Dongen et al., 2004; Klerman and Dijk, 2005). Preliminary validation of these techniques indicates that as the number of past data points increases, the model increases the accuracy with which the trait parameters are estimated, resulting in significant improvements in performance prediction accuracy over population average models (figure 3-7) (Van Dongen et al., 2007).

Figure 3-7. Future performance of three individuals, measured with the 10-minute PVT, during total sleep deprivation condition is predicted starting from t = 44h of wakefulness, with mean (thick black line) and 95% confidence interval (vertical bars). Individual predictions are based on traits that are identified from prior performance measurements up to t = 44h (block dots). The individualized predictions more accurately forecast the actual future performance of each individual (gray dots) than does the population average prediction (red line).

The second model that was mentioned previously—the Circadian, Neurobehavioral Performance, and Subjective Alertness Model— predicts the effects of different light/dark and sleep/wake patterns on the circadian biological clock, performance, and alertness. Astronaut performance or alertness for an entire schedule or for a mission-critical time can thus be predicted. The model has been validated with data from shifted sleep-wake (e.g., jetlag or night work), low-light conditions, intermittent bright-light exposure data, and non-24-hour schedules (e.g., Mars), all of which apply to NASA operations. This model has also been used successfully to design a pre-flight light exposure regimen that is associated with the early-morning launch times that are often necessary for shuttle flights.

These methods can be used to design a variety of schedules that are relevant to NASA operations, including shifting sleep/wake (slam shifting) and non-24-hour schedules. Critically, these methods will be able to satisfy the variety of schedules that will be encountered during a Mars mission, where a day is 24 hours and 39 minutes.

Current work includes quantifying individual differences in response to circadian and sleep/wake factors, and incorporating non-light stimuli (e.g., posture and social cues) and information concerning the various wavelengths of light into the model, since the circadian system is responsive to specific wavelengths of light and the wavelength distribution that is found in space differs from that found on Earth (both indoors or outdoors).

This work allows for mathematical simulations that assess the impact of circadian alignment and sleep disruption on performance and alertness.