建構非侵入式連續血壓量測系統以評估血管系統之數學模型

Development of non-invasive continuous blood pressure monitor system for vascular system model evaluation

張峰碩(Feng-Shou chang)

指導教授 ： 徐良育

中原大學/工學院/醫學工程研究所/碩士(2005年)

https://doi.org/10.6840/cycu200500589

摘要

血壓就如同心電圖一般，可借助連續血壓訊號進行心臟方面疾病之診斷。在臨床應用上，侵入性血壓量測裝置可提供連續血壓訊號，雖精準卻不可能在一般例行性檢查中進行。非侵入性血壓量測計可應用於一般檢查或居家環境中，卻僅能提供收縮壓與舒張壓兩個特徵值，無法呈現完整血壓訊號的資訊。因此本研究提議研發一個非侵入式連續血壓量測系統，可在一般環境中取得連續血壓訊號。另一方面，人體不同部位的動脈所量測到的血壓波型，因壓力波在動脈系統中的傳送、反射、時間延遲與血管系統時變的特性，因而造成波型的不同。若能將兩不同部位之連續血壓訊號，利用自動控制領域中系統識別之理論建立動態血管模型，即可在模型中觀測血管系統之時變性。
本研究所研發之非侵入式低壓型連續血壓量測系統，主要包含兩組連續血壓訊號量測通道與一組心電訊號量測通道。在血壓量測通道之恆定壓力控迴路以MSP430F149微控制器為核心，利用動態平衡與回饋機制，將氣囊內壓維持於40 mmHg之恆定壓力。因血壓而引起之氣囊內壓力變化經由壓力感測器轉換為電訊號，經由增益為6250倍放大與截止頻寬為0到30Hz之濾波處理，以獲得連續血壓訊號。
本研究中針對10位正常年輕男性，設計三項實驗以改變人體生理狀態，實驗分別為穩態、冰敷與Valsalva maneuver，利用血壓量測系統分別量測此三個實驗中手腕與手臂連續血壓訊號進行分析。在穩態的狀況下，兩個量測點之收縮壓、舒張壓、舒張壓間期與振幅的平均相關度分別為0.83±0.15、0.81±0.14及0.91±0.07與0.48±0.21。在冰敷的狀況下，其收縮壓、舒張壓、舒張壓間期與振幅之相關度則略為下降。在Valsalva maneuver實驗後，其收縮壓、舒張壓、舒張壓間期與振幅之相關度則下降為0.16±0.54、0.25±0.59、0.77±0.24與0.30±0.32。在血管模型建立方面，本研究將每ㄧ個心搏視為一個系統，利用系統識別的方式建構血管模型，將不同心搏之脈衝響應函式與頻率響應函式組合成一個時變脈衝響應函式與時變頻率響應函式，並以3D圖形的方式呈現。同時，以描繪出累積能量90%的曲線可清楚的觀察血管系統在三種生理狀況下的動態變化。
本研究完成一非侵入式連續血壓量測系統，並以實驗證明此系統確實能長時間擷取連續血壓波型，實驗結果同時顯示本系統可運用於血管系統動態頻率響應之研究。然而，本系統對動作雜訊非常敏感，同時血壓訊號亦需要校正，以反映實際血壓，在未來可針對此兩缺點加以改進，以增加系統之實用性。

關鍵字

非侵入式 ； 連續血壓量測 ； 血管系統模型

並列摘要

Blood pressure, like electrocardiogram, can be used to diagnose cardiac disease espectively in the form of continual blood pressure signal. In clinical practice, invasive devices can be used to provide continual blood pressure signal. Although, these invasive methods are accurate, it cannot apply to routine inspection. In non-clinical circumstances, noninvasive blood pressure measurement devices are the only choice. However, most of these noninvasive devices can only provide systolic pressure and diastolic pressure readings. Therefore, this study proposes to develop a non-invasive continuous blood pressure monitor system that can be used to obtain continual blood pressure signal in every circumstances. In addition, it is known that the contour of blood pressure waveform is site dependent, due to wave propagation, reflection, time delay and time-varying characteristics of vascular system. Thus, vascular system dynamic model can be evaluated using system identification theory in the field of automatic control in order to observe the time-varying characteristics of vascular system.
In this study, the proposed non-invasive continuous blood pressure monitor system consists of two blood pressure measurement channels and one electrocardiogram measurement channel. In the blood pressure measurement channel, the MSP430F149 micro-controller is used to servo control the cuff pressure through dynamic balance and feedback such that the cuff pressure can be maintained at 40 mmHg. The cuff pressure variation provoked by blood pressure is converted to electronic signal via pressure transducer. The total gain of blood pressure measurement circuit is 6250 and the filter bandwidth is form 0 to 30Hz.
Ten normal young males were recurred in this study who want on to three physiological condition changes including steady-state, cold pressor, and Valsalva maneuver. The continual blood pressure signals from ulnar and brachial arteries were obtained by the proposed system. During steady-state, correlations of the systolic pressure, diastolic pressure, diastolic interval, and pressure change between ulnar and brachial arteries are 0.83±0.15, 0.81±0.14, 0.91±0.07, and 0.48±0.21, respectively. During cold pressor, correlations decrease slightly. On the other hand, after Valsalva maneuver experiment, correlations of the systolic pressure, diastolic pressure, diastolic interval, and pressure change are 0.16±0.54, 0.25±0.59, 0.77±0.24, and 0.30±0.32, respectively. Additionally, the vascular system model was evaluated using system identification theory. In this study, every heart beat was considered as an unique system. Combining impulse response and frequency response of different heart beat, we can obtain time-varying impulse response and time-varying frequency response and depict it in 3D graph. In order to assist the observation of the dynamic characteristic change of vascular system during different physiological conditions, the 90% accumulated energy curve was proposed and used. However, only some to the subject illustrated visible changes during tasks.
In conclusion, this study realized a non-invasive continuous blood pressure monitor system. It was used in three experiments to prove that the system can acquire continuous blood pressure waveform for long period of time. Additionally, this system can be applied to study the dynamic characteristics of vascular system. However, it is very sensitive to motion artifact. Moreover, the blood pressure signal must be calibrated to reflect the true pressure reading. Thus, in the future, it is necessary to correct these disadvantages in order to promote this system for any practical useage.

並列關鍵字

continuous blood pressure measurement ； vascular system model ； non-invasive

參考文獻

[1] Carr, J.J., Brown, J.M., “Introduction to biomedical equipment technology”, Prentice hall, fourth edition, 2001.

[2] Tanaka, S., Yamakoshi, K., “Ambulatory Instrument for Monitoring Indirect Beat-to-Beat Blood Pressure in Superficial Temporal Artery Using Volume-Compensation Method”, Medical & Biological Engineering & Computing, Vol. 34, pp.441-447, 1996.

[3] Yamakoshi, K., Kawarada, A., Kamiya, A., Shimazu, H., Ito, H., ”Long-term Ambulatory Monitoring of Indirect Arterial Blood Pressure Using a Volume-Oscillometric Method”, Medical & Biological Engineering & Computing, Vol. 23, pp.459-465, 1996.

[4] Kawarada, A., Shimazu, H., Ito, H., Yamakoshi, K., “Ambulatory Monitoring of Indirect Beat-to-Beat Arterial Pressure in Human Fingers by a Volume-Compensation Method”, Medical & Biological Engineering & Computing, Vol. 29, pp.55-62, 1991.

[5] Shimazu, H., Ito, H., Kawarada, A., Kobayashi, H., Hiraiwa, A., Yamakoshi, K., “Vibration Technique for Indirect Measurement of Diastolic Arterial Pressure in Human Fingers”, Medical & Biological Engineering & Computing, Vol. 27, pp.130-136, 1989