Training subjective insomniacs to accurately perceive sleep onset.

Subjective insomniacs overestimate sleep latency at the beginning of their nocturnal sleep period. It was hypothesized that subjective insomniacs could be trained to accurately estimate sleep latency by learning to differentiate wakefulness from sleep. Ten subjective insomniacs were randomly assigned to one of two groups. Group 1 subjects participated in both a control and a training week; group 2 subjects participated only during a training week. Each week consisted of a baseline lab night, a training lab night (treatment or control), a home (unmonitored night) and a recovery lab night. During training, subjects were taught to use sleep markers (A, B or C) to help them more accurately estimate sleep latency and were given feedback about the accuracy of their estimates. Marker A corresponded to an electroencephalographic level of wakefulness; marker B corresponded to the initial sleep spindle; marker C corresponded to 5 minutes of continuous sleep after the first sleep spindle. In the control condition, subjects had no feedback and were not taught to use markers to help them judge sleep from wakefulness. Total sleep time and percent stage 3 sleep increased, and objective sleep latency decreased on recovery nights. After training, subjective sleep latency, correctness of estimates of sleep versus wakefulness and perceived ability to fall asleep significantly improved. This study helps to establish that subjective insomniacs can learn to more accurately estimate sleep from wakefulness with the use of sleep-wake markers.