M.W. Johns. Assessing the Drowsiness of Drivers. Commissioned by VicRoads, Australia. 2001
Summary
This is a review of psychophysiological evidence relevant to the problem of drowsy drivers who "fall asleep at the wheel". Although wide-reaching, it is not an exhaustive review. It does not include biochemical and hormonal changes at sleeponset. If focuses rather on several points of view that I believe are potentially helpful, as well as others that have been tried in the past and which I believe are not very helpful. Several conceptual and methodological problems are identified that are impeding our progress in this field. They relate particularly to the currently accepted model of sleep and wakefulness, and to the definitions of sleep stages by Rechtschafen and Kales (1968) that are quite inadequate to describe drowsiness as a fluctuating state between alert wakefulness and R&K stage-1 sleep.
An understanding of the anatomy, physiology, and movements (both normal and abnormal) of the eyes and eyelids can explain their role in intentional and unintentional sleep-onset in a way that has not been described before. The concept of a microsleep, which is a brief period of stage-1 sleep with theta-waves dominating the EEG, is useful but does not explain many episodes of performance failure by drowsy subjects. Many such "lapses" are associated with alpha-waves in the EEG when the eyes are open. These waves would normally be blocked when the eyes are open in an alert subject. The "lapses" evidently involve the impairment of visual perception, especially at the periphery of the visual field, and an inability to maintain attention. Most cognitive processes are slowed down by drowsiness.
In the drowsy state, a driver who is striving to stay awake can fall asleep with his eyes open. This poses a potential problem for methods of monitoring drowsiness that rely solely on the duration of eyelid closures and blinks, such as PERCLOS. The recognition of slow eye movements that are unique to the drowsy state is a promising new method. However, much more research is needed.
