Correlation between EOG and EEG in the detection of sleep onset

中山醫學大學機構典藏 CSMUIR > 行政單位 > 研發處 > 期刊論文 > Item 310902500/19173

Please use this identifier to cite or link to this item: https://ir.csmu.edu.tw:8080/ir/handle/310902500/19173

Title:	Correlation between EOG and EEG in the detection of sleep onset
Authors:	Huang, Ren-Jing;Hsiao, Ya-Yun;Ting, Hua
Contributors:	中山醫大
Keywords:	Driver;slow eye movement;blink
Date:	2017-12-01
Issue Date:	2018-04-03T08:18:57Z (UTC)
Publisher:	研發處育成中心暨產學合作組
Abstract:	Purpose: Bus driver sleepiness is a serious and urgent public safety issue. Manually analyzing electroencephalogram (EEG) changes to determine sleep onset latency, Maintenance of Wakefulness Test (MWT), is the gold standard for evaluating subject's ability to keep alert at twilight and in quiet situations. However, this test cannot be applied to drivers at work. Therefore, we attempted to develop an auto-program that analyzes sole electrooculogram (EOG) signals in a quasi-real-time manner, and sends out an immediate warning when sleepiness is detected. Methods: Eighty-one bus drivers were recruited. Blink-type waves were characterized by pulse-type, high amplitude and high frequency (0.6-1.3Hz), and corresponded to fast eye closing and opening. Slow eye movement (SEM) was characterized by low frequency (0.2-0.6Hz), and corresponded to rolling, horizontal, bidirectional and conjugate eye movements. Sleep onset was determined by Blink index level lower than a certain threshold, concomitant with SEM index level higher than another threshold. These two thresholds were set progressively by corresponding MWT-determined sleep onset. Results: Thresholds of these two indices were set at 79 and 985 from the data of 43 subjects. Sleep latency determined by EOG showed good agreement and sensitivity in comparison with MWT (91.2% and 93.3%, respectively). Following further verification with 38 subjects, agreement and sensitivity of sleep latency detected by EOG reached 88.3% and 93.6%, respectively. Conclusion: This auto program computed sleep onset latency using indices of Blink and SEM. The results are in strong agreement with those of MWT. Our innovative EOG-based program has the potential to be applied to drivers and working individuals to warn of sleep onset in almost real time.
URI:	https://ir.csmu.edu.tw:8080/ir/handle/310902500/19173
Relation:	中山醫學雜誌 27卷2期 , P69~74
Appears in Collections:	[研發處] 期刊論文

Files in This Item:

File	Description	Size	Format
(2校)中山雜誌28-2內頁-27-32 (1).pdf		871Kb	Adobe PDF	232	View/Open

View Licence

Loading...