Reduced alertness while driving: Towards a composite behavioral/ psychophysiological measure

N. González¹, I. Kalyakin², A. Kauppi³, D. Ivanov² and H. Lyytinen³

¹AGORA Human Technology Center, University of Jyväskylä, Jyväskylä, Finland
²Department of Mathematics Information Technology, University of Jyväskylä, Jyväskylä, Finland
³Department of Psychology, University of Jyväskylä, Jyväskylä, Finland

Drivers’ performance is greatly impaired under reduced alertness states, which can lastly result in fatal accidents. In the endeavor of developing reliable alertness monitoring systems, some authors advocate the use of multiple measures to monitor alertness while others concentrate on single measures. This research seeks a composite measure of the psychophysiological and behavioral precursors of suboptimal alertness states the drivers might experience during prolonged drives. In this study, 7 sleep-deprived paid voluntaries (3 male, 4 female) performed a monotonous simulated driving task (approximately 3 hours) together with an embedded vigilance task. In the latter, participants responded to a red or green LED light by pressing a corresponding red or green button on the steering wheel. The light was presented either on the left or right side of the driving scene. Light color, side of presentation, and inter stimulus interval were randomized to reduce stimulus predictability. Simultaneously, an unattended oddball stream of auditory stimuli differing in frequency and intensity was presented through headphones. Different measures, i.e. electroencephalography (EEG), electro-oculography (EOG), heart rate (HR), respiration, skin potential (SP), reaction time (RT), and driving incident index (DIX) were recorded during the experiment. Consistent with the literature, performance fluctuated widely within and between participants. Overall, alertness decreased as a function of time on task (TOT). This was indicated by increased RT in the vigilance task throughout the session and omitted responses. These were infrequent and occurred more often during the second half of the drive. DIX was irregularly distributed across participants. Only four of them showed increases towards the end of the session. Significant correlations resulted between RT and TOT, DIX, blink rate, and the EEG ratio beta/alpha + theta. Despite the caution taken in the interpretation of the results, these suggest that a composite behavioral/psychophysiological measure of suboptimal alertness can be more reliable than single ones.

Paper presented at Measuring Behavior 2005 , 5^th International Conference on Methods and Techniques in Behavioral Research, 30 August - 2 September 2005, Wageningen, The Netherlands.

© 2005 Noldus Information Technology bv