48時間断眠状態におけるヴイジランスパフォーマンスの変動 不安定状態仮説の検証を中心に

摘要

The purpose of the present experimental study was to clarify whether the "state instability" hypothesis was valid for 48-hour sleep deprivation. The hypothesis,which has been recommended by Dinges DF and his colleagues,stresses that variability in sustained attention performance,such as vigilance performance, will reflect a combination of normal timely responses,"omission errors" like misses or lapses, and "commission errors" such as committed errors or false alarms, in spite of long-term sleep deprivation (SD).In the hypothesis,such performance variability is assumed to arise as human performance during SD changes with the interaction of homeostatic drive to sleep elevating during it, endogenous circadian promotion of wakefulness,and increasing compensatory effort to respond as sleep loss progresses.The last factor means that as sleep-deprived hours increase,subjects tend to respond less and worse in a task due to sleepiness,but there often occur very short periods when they will make an effort to respond normally and correctly in order to achieve good results in the task.Six male healthy university students served as subjects of the experiment.The experiment was carried out for four days, and the subjects were totally deprived of sleep for 48 hours from 08:00 of the 2nd day till 08:00 of the 4th day.The subjects were asked to respond to the stimuli of a vigilance task during 20 minutes every two hours during the 48-hour period. The vigilance task used in the experiment was a relatively high-signal loaded,sustained attention performance task requiring the subjects to press a button corresponding to the presentation of one of four different numbers, presented in the center of a computer monitor.The presentation time of the stimulus was 2 seconds and time intervals between the presentations were set at random, 5 to 25 seconds.The results of the experiment showed that mean reaction times for correct responses in the 2 SD session (08:00 of the 3rd day to 08:00 of the 4 day) were significantly longer than those in the 1st SD session (08:00 of the 2nd day to 08:00 of the 3 day).The results also showed that mean numbers of lapses,misses, and false alarms in the 2 SD session also increased significantly compared to those in the 1 SD session. As for circadian variations,the mean reaction times were significantly longer in the midnight and early morning hours (01:00-07:00) than in other hours during the 48-hour period.The mean numbers of "omission errors", i.e. lapses and misses,also revealed such significant circadian variations.In this study,the results of performance during the vigilance task done in the SD sessions were compared between five 4-minute phases. And the state instability phase was defined as a 4-minute period having a combination of normal timely correct responses, with two or more misses, two or more lapses, and two or more errors and/or false alarms.It was revealed that the mean number of the state instability phases increased significantly in the 2 SD session compared with the 1 SD session, with a tendency to increase during the periods of 09:00-13:00 and 01:00-07:00 in the 2 SD session.It was concluded that the state instability hypothesis was considered to be valid for 48-hour sleep deprivation.