- 學位論文
利用不同生理條件對腦血流自動調節進行分類
Classification of dynamic cerebral autoregulation using different physiological conditions.
摘要
本研究旨在開發和評估一個二元分類器，以根據受試者的血壓（BP）和腦血流速度（CBFV）數據來判斷其健康狀態。我們使用經顱多普勒超聲波（TCD）和轉移函數分析（TFA）測量了受試者在基線量測、高碳酸血症量測和大腿袖帶測試條件下的BP和CBFV數據。在分類器的開發過程中，我們使用支持向量機（SVM）對數據集進行訓練和測試，並分析了分類器的性能指標和特徵貢獻。我們的研究結果表明，使用基線量測和高碳酸血症量測訓練的分類器相比於大腿袖帶測試和高碳酸血症量測訓練的分類器，展示出了更高的準確性，這突顯了在這些狀態下BP和CBFV關係的穩定性和可預測性。
並列摘要
The brain's ability to maintain stable cerebral blood flow (CBF) despite fluctuations in blood pressure (BP) is crucial for preventing damage and ensuring normal brain function. This study aims to develop and evaluate a binary classifier to determine the health status of subjects based on their blood pressure (BP) and cerebral blood flow velocity (CBFV) data, with the assumption that subjects can be classified as either baseline or impaired. Using Transcranial Doppler (TCD) ultrasound and Transfer Function Analysis (TFA), we measured BP and CBFV under normocapnia, hypercapnia, and thigh cuff testing conditions. For classifier development, we trained and tested the dataset using Support Vector Machine (SVM) and analyzed the performance metrics and feature contributions of the classifier. Our findings indicate that classifiers trained under normocapnia and hypercapnia conditions demonstrate superior accuracy compared to those trained under thigh cuff testing conditions, highlighting the stability and predictability of BP and CBFV relationships in these states.