具生物反餽機制之智慧型長時睡眠監測系統---子計畫二：多重生理訊號分析與硬體實現(I)

Abstract

本計畫針對多重生理訊號提出一個硬體平台架構及其對應之訊號 處理晶片系統設計，以用於臨床上的異常事件偵測及其即時分析，並提 供長時多重生理訊號之記錄，以便於事後之分析與病症之判讀。在臨床 醫學上，生理訊號是診斷、照護、治療的重要依據，腦電波、肌電圖、 眼電圖可辦別睡眠時期，心電圖可以監控自主神經並反應出生活緊張狀 態，鼻孔、胸部與腹部的感測器可評估呼吸節率，血氧計評估血中含氧 濃度，麥克風評估打呼情形等。目前生理記錄大多依靠醫護人員判讀， 然在急診室、加護病房、嬰兒室與長期照護中心，具有大量患者與患者 生理狀態變化快速且變異度大的特性，可靠的即時生理異常示警裝置將 能在定時的照護程序外，讓患者在突發狀況可及時獲得有效的緊急照 護。同時，長時多重生理訊號的紀錄與分析，也對於許多特定的病症有 重要的意義。因此大量的使用於睡眠障礙分析、癲癇外科治療的輔助定 位等。本計畫所提出之硬體平台與晶片系統設計，即針對上述多重生理 訊號即時分析與異常事件監測，配合長時多重生理訊號及路的即時標記 技術，設計專屬之生理訊號處理電路設計，並配合整合型計畫及其他子 計畫各式感測器與儲存系統設計，達到多重生理訊號分析記錄之硬體微 小化的目標。

This project proposed an on-line multi-stream physiological signal processing and abnormal event detection system and its corresponding integrated circuits designs. In clinics, physiological signals play an important role in diagnosis and symptoms analysis. For example, EEG, EMG can be utilized to analyze sleeping step; ECG can be utilized to monitor autonomic nervous system. Traditionally, clients rely on health care staffs to monitor physiological signals. However, in emergency room, intensive care unit, and other long-term care units, a reliable physiological signals monitoring machine with automatic abnormal event detection is useful to provide the in time clinical care and reduce the loading of health care staffs. In the mean time, long term recording of multi-stream physiological and its labeling and analysis is utilized in diagnosis sleep scoring, sleep apnea, etc. Thus, this project proposed to design and implemented a bio-signal processing unit and its hardware/firmware integration. By integration with the analog front-end circuits and storage elements that design by other portion of this integration project. A miniaturized multi-stream physiological signal recording and reacting system is proposed. Which can be utilized as a polysomnography or other medical device, furthermore, the feed-back stimulation or alarm from the on-line monitoring can provides the function to prevent sudden infant death and other syndromes.