利用單通道腦電波進行自動睡眠分期之快速動眼期睡眠剝奪

Automated Sleep Staging using Single EEG Channel for REM Sleep Deprivation

李郁萱(Yu-Shiuan Lee)

指導教授 ： 陳永昇

國立交通大學/資訊學院/生醫工程研究所/碩士(2009年)

https://doi.org/10.6842/NCTU.2009.00762

摘要

本論文目的為發展以單通道腦電波為分析訊號來發展即時自動睡眠分期系統，並利用此系統進行快速動眼期之睡眠剝奪實驗。醫學文獻顯示，針對憂鬱症的睡眠腦波，其中一項特徵為憂鬱症患者的快速動眼期發生頻率會比正常人頻繁，因此在他們所服用的抗憂鬱劑具有抑制快速動眼期的效果。相關醫學文獻指出，服用抗憂鬱藥物和快速動眼期睡眠剝奪的機制是相同的，在過去有少量針對憂鬱症患者進行睡眠剝奪的實驗，其實驗結果也證實了睡眠剝奪對於內生型憂鬱症(endogenous depression) 的改善是具有一定效果的。由於在進行睡眠剝奪的實驗中，睡眠分析師必須整夜監測受試者的睡眠狀態，利用人工方式判讀睡眠分期，並在受試者的睡眠狀態處於快速動眼期時將之吵醒。但由於費時費力的缺點，過去僅有少數研究提出睡眠剝奪實驗的治療方式對於憂鬱症治療的效果。因此，本論文欲發展以單通道腦電波為基礎的即時自動睡眠分期系統，以即時偵測到快速動眼期並對受測者進行睡眠剝奪。
　　在此系統中，我們利用支援向量機作為分類器，對擷取之腦電波特徵做分類。我們利用二十五位受測者的睡眠腦波作為挑選腦電波特徵的測試資料，在嘗試過各種不同的特徵擷取方式後，採用可以得到最佳睡眠分期準確度的特徵擷取方法。利用本論文所建構之自動睡眠分期系統，針對二十五位不同受測者，平均可達到百分之八十五的分期準確度。我們對此自動睡眠分期系統進一步發展成即時自動睡眠分期系統，一旦偵測到快速動眼期，系統即會自動發出聲音以剝奪受測者之快速動眼期的睡眠。在睡眠剝奪的實驗中，我們以六位健康狀況良好，無失眠狀況及憂鬱傾向的受試者來進行本實驗。本實驗目的在於驗證我們所設計的自動睡眠剝奪機制是否能夠達到一定的睡眠剝奪效果，未來預期能運用此系統以進行憂鬱症患者的睡眠剝奪，並探討睡眠剝奪對於憂鬱症的療效。

關鍵字

腦波 ； 自動分期 ； 快速動眼期 ； 非快速動眼期 ； 睡眠剝奪 ； 多重睡眠訊號分析圖 ； 睡眠分期圖 ； 憂鬱症

並列摘要

The objective of this study is to develop an on-line　automated sleep staging system based on single EEG analysis to assist REM sleep deprivation. Polysomnographic sleep research has demonstrated that the increased rapid eye movement (REM) density is one of the characteristics of depressed sleep. Some experiments were conducted to confirm that REM sleep deprivation (REM-SD) for a period of time is therapeutic for endogenous depressed patients. However, because of its intensive labor requirement, validity of this therapy has not yet been assessed by a sufficient amount of depression patients. Therefore, we propose to
develop an automated sleep staging system using only single EEG channel to achieve on-line detection for REM state during sleep. For our sleep staging system, it is based on the supervised classification method with support vector machine. In order to select the feature extraction method which can achieve the best classification result within single EEG channel, we have implemented some feature extraction methods and tested with 25 sleep records. Feature sets derived from different feature extraction method which can achieve the best classification accuracy is the one we will adopted in the proposed system. The average accuracy of classification of all 25 records can achieve 85% that is feasible to REM sleep deprivation. After the algorithm of sleep staging is established, we extend the system to on-line staging which can output the scoring result right away as every 30-second epoch is acquired. Once the REM state is detected by the system, the system will make a shrill sound
to disturb the subject until the REM sleep state is changed to other states. Six healthy subjects enrolled in this experiment and have verified the feasibility of the procedure of REM sleep deprivation assisting with the on-line automated sleep staging system. The experimental results demonstrate that our on-line automated sleep staging system is reliable and applicable for on-line REM sleep deprivation.

並列關鍵字

electroencephalogram (EEG) ； automated sleep staging ； rapid eye movement (REM) ； non-rapid eye movement (NREM) ； sleep deprivation (SD) ； polysomnogram (PSG) ； sleep hypnogram ； depression

參考文獻

[1] Flexer Arthur, Gruber Georg, and Dorffner Georg. A reliable probabilistic sleep stager based on a single eeg signal. Artificial Intelligence in Medicine.

[2] Chang Chih-Chung and Lin Chih-Jen. Libsvm: a library for support vector machines. 2007.

[3] Chih-Chung Chang Chih-Wei, Hsu and Lin Chih-Jen. A practical guide to support vector classification. 2008.

Marie-Pia. Automatic analysis of single-channel sleep eeg: Validation in healthy individuals. SLEEP, 30(11), 2007.

[5] F. Lopes da Silva. Computer-assisted eeg diagnosis: Pattern recognition techniques. Electroencephalography, Basic Principles, Clinical Applications and Related Fields.,