Sleep EEG has been categorised into sleep stages by standardised scoring rules since 1968. The rules were assessed by the Committee led by Allan Rechtschaffen and Anthony Kales. The standardised scoring system, often called "the manual of Rechtschaffen and Kales" (RKS) has become practically the only method of visual sleep analysis. It is characterised by fixed, long epochs, a discrete number of sleep stages and ignorance of EEG topography.

In order to overcome these limitations a visual adaptive scoring system (VASS) was developed. Epochs of variable length and more stage categories than in the standard system are used. The purpose is to have as electrophysiologically stationary epochs as possible.

In the present study VASS was applied to the analysis of the multiple sleep latency tests (MSLT) of 17 subjects. Ten subjects had a clinical history of obstructive sleep apnea syndrome and an apnea-hypopnoea index (AHI) of > 10/h in a preceding polygraphic whole-night recording. In addition 7 healthy subjects with no history of EDS or sleep complaints and AHI Each MSLT consisted of four naps. The naps were terminated 20 min. after lights out if there was no sleep at all. If the subject fell asleep, the recording was continued for 15 minutes from the first S1 epoch. Frontal, central and occipital EEG derivations were recorded. In addition two channels of eye movements (EOG), submental muscle tonus and respiratory variables were measured.

Each nap was scored by RKS and VASS and the mean sleep latencies of the MSLTs were determined. The VASS stages used were 3 categories of clear wakefulness, 2 categories of drowsy wakefulness with slow eye movements, 4 types of movements and EMG augmentations, 2 categories of light sleep (Drowsy low and S1VASS), S2VASS, REMVASS and several types of arousals. Electrophysiological stage changes shorter than 1 s were not separately scored.

Quality of life, subjective sleepiness and performance were measured by questionnaires and psychometric tests. Spectral analysis by FFT was made from stages scored both by RKS and VASS.

VASS revealed that sleep onset consists of fluctuations between wakefulness and sleep. This fluctuation is already visible before sleep can be scored by the conventional method. VASS provides more precise information on sleep staging, as eye movements and topography are also taken into account. The nine VASS stages in general can be differentiated from each other by spectral means. RKS stages were found to consist of several different VASS stages. Therefore VASS can provide a better reference for the validation of automatic sleep analysis.

The MSLT latencies scored by VASS were in general shorter than by RKS, especially in the patient group. It seems that the sensitivity of MSLT can be improved by VASS. However, in clinical practice the use of VASS is not unproblematic. Although VASS scored latencies presented abundant correlations to psychometric tests, no model to predict sleepiness could be derived. Therefore no single optimal parameter to be used in clinical practice was found.

The greatest value of VASS is that it proved to be a sensitive method to examine sleep dynamics both in healthy subjects and in sleep apnea patients. If it is assumed that sleep disorders are disturbances of the sleep process, then VASS provides an efficient tool for scientific purposes.