Quantitative description of sleep stage electrophysiology using digital period analytic techniques

The purpose of the current study was to assess the overlap in variance of two procedures for the quantification of sleep electrophysiology: conventional stage scoring and computer quantification of tonic activity. Data were collected on 24 nights from eight subjects and were scored according to a modified set of Rechtschaffen and Kales criteria and submitted to a period-analytic computer analysis. Following this, discriminant function analyses were performed on the data for each night to predict the visual stage scores from the computer-generated data. The results indicate a very high degree of predictive accuracy (91.05%) supporting the contention that the computer-quantified data set includes the variance normally captured by stage scoring. The implications of computer quantification of sleep electrophysiology are discussed.