Sleep and Biological Rhythms 2004; 2: S58

POSTER PRESENTATION

Accuracy evaluation of sleep-wake stage analysis with SleepSign Ver2.0 Yujiro TAGUCHI,1 Shima HANDO,1 Mie SAKATA,2 Naomi EGUCHI2 and Yoshihiro URADE2 1

Kissei Comtec Co., Ltd, Nagano, 2Osaka Bioscience Institute, Osaka, Japan

INTRODUCTION The survey of sleep in mice differs from that in humans in several aspects. Mice fall asleep regardless of day or night and transitions between sleep and wakefulness occur frequently. Also, in the case of mice, the survey must be carried out for at least 24 h, even for a week at times; and available bio-signals are limited to the EEG and EMG. These facts make the evaluation of mouse sleep complicated and troublesome. In addition, the minimum time unit for sleep stage scoring (epoch) for mice varies from 4 s to 30 s, depending on the researcher, whereas that for humans is standardized as 20 s or 30 s. Our recent study on histamine H1 receptor gene-knockout mice revealed that their frequency of short-term arousal (<16 s length) was significantly lower than that of wild-type mice. This finding implies the importance of scrutinizing more minute transitions of sleep-wake stages. Therefore, we evaluated the impact of the difference in epoch length (4 Section vs. 20 s) on the accuracy of sleep analysis results, such as duration and frequency of each stage. We used SleepSign Ver.2.0 software, which we developed for sleep analysis. It is characterized by its automatic sleep-wake stage scoring function and video recording function.

METHODS EEG and EMG were recorded for 24 h (a 12-h light/12h dark cycle: light on at 08.00) from five wild-type mice (C57BL/6). We carried out sleep-wake stage analysis on 4 s/epoch and 20 s/epoch bases, and compared the results in the following respects: (i) Total duration and

Correspondence: Yujiro Taguchi, Kissei Comtec Co., Ltd, 4010–10 Wada, Matsumoto, Nagano 390–1293, Japan. Email: [email protected]

S58

average duration of each sleep stage (REM sleep, nonREM sleep, Wake) in each light and dark period; (ii) the number of arousal periods shorter than 16 s and that occurred during non-REM sleep in each light and dark period; and (iii) EEG frequency spectrum for non-REM sleep and REM sleep. Sleep-wake stage scoring was carried out automatically with SleepSign software and reviewed manually by referring to EEG frequency analysis results.

RESULTS AND DISCUSSION 1 No apparent difference in the EEG spectrum (0.75Hz~25Hz) in each non-REM sleep and REM sleep was observed. 2 No apparent difference in the total duration of each sleep stage (REM sleep, non-REM sleep, Wake) was observed in each light and dark period. 3 The number of arousals with the 4 s/epoch was about 6 times greater than that obtained with 20 s/epoch. 4 The average duration of wakefulness with the 4 s/ epoch was about 50% of that with the 20 s/epoch in both light and dark periods. 5 The average duration of non-REM sleep with the 4 s/ epoch was also about 50% of that with the 20 s/epoch in both light and dark periods. However, no apparent difference in the average duration of REM sleep was observed. These results suggest that the number of arousals can be measured and the stage transition can be grasped more accurately with 4 s/epoch than with 20 s/epoch. We believe that the use of 4 s/epoch will be essential for sleep-wake stage analysis of various knockout mice that will be conducted in the near future. Besides, we have added a simultaneous video recording-playback function to SleepSign Ver.2.0. We believe with confidence that these elements will become powerful tools to discover unknown sleep-wake disorders.

© 2004 Japanese Society of Sleep Research