Driver sleepiness--evaluation of reaction time measurement as a secondary task

The application of reaction time (RT) as a secondary task to determine sleepiness in drivers is of increasing interest, but is a problematic area. We assessed the extent to which RT reflected this sleepiness, and/or otherwise affected driving behaviour in sleep restricted, moderately sleepy people. They drove a real-car interactive simulator for two, two hour afternoon monotonous drives, with and without RT (counterbalanced). Simple auditory RT was used, with a semi-random inter-stimulus interval averaging 21/2 minutes. Lane wandering (driving "incidents"), subjective and EEG measures of sleepiness were obtained. For both conditions all three indices changed significantly during the course of the afternoon circadian "dip". However, this was not reflected in RT, which remained relatively stable. Nevertheless, RT provided more "stimulation" for the sleepy driver, and significantly reduced subjective sleepiness, with a trend for fewer incidents and a more alert EEG. Possible reasons for the disparity in sensitivity between RT and the other measures are discussed. Under this experimental protocol, RT did not provide a useful guide to driver sleepiness; it was merely a mechanism for increasing task load and reducing monotony. The drivers' own insight into their sleepiness had more validity as a tool for assessing sleepiness.