Behavioral,subjective, and electroencephalographic indices of sleep onset latency and sleep duration

Electrographic (EEG, EOG, EMG) indices have been used for some decades in the definition of the stages of sleep and more recently in the diagnosis of sleep-related disorders, e.g., insomnia, despite the lack of detailed information concerning the precise relationship between such electrographic indices and behavioral and subjective criteria of sleep. Evaluation of the relationship between EEG, behavioral, and self-report measures of sleep onset latency and sleep duration was conducted using 17 young normal sleepers. The behavioral measure was provided by an apparatus which records subjects' button-press responses to an auditory stimulus presented at various interstimulus intervals (2, 5, and random 1, 2, 5, 7, and 10 min). The behavioral and stage 2 EEG estimates of sleep onset latency (SOL) and sleep duration (SD) were almost identical. The stage 1 EEG provided the shortest estimate of SOL; the self-report measure, the longest. The SD measures were in reverse order. There were no significant differences among the three interstimulus interval conditions (2, 5, and random min). Average response rates to the 50-dB chime were 100% during wakefulness, 81% during EEG stage 1, and 8% during EEG stage 2. Almost all stage 2 responding occurred during the first 5 min of each stage 2 period. The validity of electrographic indices as sleep criteria and the implications of the findings for the formulation of an adequate definition of sleep and its clinical measurement are discussed.