穿戴式PPG裝置於癲癇病患之智能臨床應用;Intelligent Clinical Applications of Wearable PPG Devices for Patients with Epilepsy

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Electrical Engineering > Electronic Thesis & Dissertation > Item 987654321/98568

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/98568

Title:	穿戴式PPG裝置於癲癇病患之智能臨床應用;Intelligent Clinical Applications of Wearable PPG Devices for Patients with Epilepsy
Authors:	洪嘉均;Hong, Chia-Chun
Contributors:	電機工程學系
Keywords:	光體積變化描記圖法;癲癇;機器學習;深度學習;心率變異性;PPG;Seizure;Machine Learning;Deep Learning;heart rate variability
Date:	2025-08-25
Issue Date:	2025-10-17 12:56:10 (UTC+8)
Publisher:	國立中央大學
Abstract:	癲癇的成因包括大腦的先天性或後天性異常，這些異常可能導致神經迴路產生異常放電，進而引發一系列陣發性或慢性的神經功能失調。癲癇的臨床表現多樣，常見症狀包括肢體抽搐、意識喪失、感覺異常等，部分患者甚至無法察覺自己正在經歷發作。因此，找出癲癇發作前後的關鍵生理指標，具有重要的臨床與研究價值。本研究招募因診斷需求、抗癲癇藥物調整或手術前評估而住院的癲癇患者，進行錄影腦電圖（electroencephalography，EEG）檢查，作為研究受試者。同時蒐集光體積變化描記圖法（photoplethysmography，PPG）與EEG訊號，並根據腦電圖紀錄進行事件標記，分析發作前、發作中與發作後，以及清醒與睡眠等不同生理狀態下的心率變異（heart rate variability，HRV）與光體積變化描記圖法波形參數變化，藉以探討自律神經系統在癲癇發作相關過程中的反應與影響。研究進一步運用機器學習技術，建立模型以偵測與預測癲癇發作事件，期望藉由模型辨識不同癲癇類型患者在發作前後與發作間期的自律神經活動特徵。並部署於穿戴式裝置中，供癲癇患者日常配戴使用，以即時監測身體狀況、預警發作風險，進而降低癲癇發作所造成的傷害與影響。此外，本研究亦致力於找出具鑑別性的HRV參數作為潛在的生物標記（biomarker），以輔助臨床醫師更早為患者選擇最適合的抗癲癇藥物，提升治療效率並減少藥物相關副作用。 ;The causes of epilepsy include congenital or acquired abnormalities of the brain. These abnormalities can lead to abnormal electrical discharges in neural circuits, which in turn trigger a series of paroxysmal or chronic neurological dysfunctions. The clinical manifestations of epilepsy are diverse, with common symptoms including limb convulsions, loss of consciousness, and sensory abnormalities. Some patients may not even be aware that they are experiencing a seizure. Therefore, identifying key physiological indicators before and after epileptic seizures holds significant clinical and research value. This study recruits epilepsy patients hospitalized for diagnostic purposes, anti-epileptic drug adjustments, or pre-surgical evaluation to serve as research subjects, undergoing video electroencephalography (EEG) monitoring. Photoplethysmography (PPG) and EEG signals are collected simultaneously. Based on the EEG recordings, events are annotated to analyze changes in heart rate variability (HRV) and PPG waveform parameters before, during, and after seizures, as well as in different physiological states such as wakefulness and sleep. This is done to investigate the response and influence of the autonomic nervous system during processes related to epileptic seizures. The study further utilizes machine learning techniques to build models for detecting and predicting seizure events. The goal is for these models to identify the characteristics of autonomic nervous activity in patients with different types of epilepsy before, after, and between seizures. These models are intended for deployment in wearable devices for daily use by epilepsy patients to monitor their physical condition in real-time and provide early warnings of seizure risk, thereby reducing the harm and impact caused by seizures. Furthermore, this research is also dedicated to identifying discriminative HRV parameters as potential biomarkers. These biomarkers could assist clinicians in selecting the most suitable anti-epileptic drugs for patients at an earlier stage, improving treatment efficacy and reducing drug-related side effects.
Appears in Collections:	[Graduate Institute of Electrical Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	9	View/Open

社群 sharing

Loading...