基于双路脉搏波建模的无袖带连续血压估计方法研究

Current researches show that besides the level of blood pressure, the dynamic change of blood pressure is also a very important risk factor of cardiovascular and cerebrovascular diseases. At present, continuous blood pressure can be estimated indirectly by ECG and single pulse wave without using a cuff, which could provide dynamic information of blood pressure conveniently. However, this method is affected by factors such as vascular elasticity, and the estimation accuracy is limit. This project proposed a cuff-less and continuous blood pressure estimation based on dual pulse waves modeling. The morphological changes of pulse wave while transmitting in the artery are affected by the vascular elasticity, so it is expected to reduce the influence of vascular elasticity on the blood pressure estimation accuracy by collaborative analysis of dual pulse wave recordings in two different positions of artery. Pulse waves in brachial artery and radial artery will be collected by flexible sensors, a mechanism-driven pulse wave propagation model will be developed (modeling 1), and system identification technology would be used to estimate the model parameters for exploring the mapping relationship of the pulse waves morphological changes and the time delay to blood pressure, while, a data-driven neural network model will be developed (modeling 2), and machine learning technology would be used to estimate the model parameters for exploring the internal relationship between the pulse wave waveform/characteristics and blood pressure. So that, cuff-less and continuous blood pressure could be estimated by using dual pulse waves. Comparing the two modeling methods to find out the best blood pressure estimation method for individual. This project is expected to develop a new cuff-less and continuous blood pressure estimation method to improve the accuracy of cuff-less and continuous blood pressure estimation, which is of great significance for the prevention and treatment of cardiovascular and cerebrovascular diseases.

最新研究表明，除血压的高低水平，血压的动态变化也是心脑血管疾病非常重要的风险因子。当前利用心电和单路脉搏波,间接估计连续血压，无需袖带，能方便地提供血压动态信息。然而，受血管弹性因素影响，估计精度不高。本项目提出基于双路脉搏波建模的无袖带连续血压估计方法。脉搏波在动脉传播过程中的形变受血管弹性影响，协同分析动脉不同位置的双路脉搏波有望减少血管弹性对血压估计精度的影响。拟利用柔性传感获取的肱、桡动脉双路脉搏波，建立基于机制传导的脉搏波传播模型（建模1），利用系统辨识解析模型参数，探明双路脉搏波的形变和时延与血压的映射关系；建立基于数据驱动的神经网络模型（建模2），利用机器学习估计模型参数，探明脉搏波波形/特征与血压的内在联系，实现无袖带连续血压估计。对比两种模型，找出适合个体的最优血压估计方法。本项目有望开发无袖带连续血压估计的新方法，提高连续血压估计精度，对心脑血管疾病的防治具有重要意义。

本项目针对当前无袖带连续血压估计受血管弹性等生理因素的影响，精度不高，难以达到临床需求的问题。从脉搏波的传感、多通道脉搏波的同步检测系统、用于血压估计的脉搏波特征的生理机制、到多阶多元脉搏波特征数据融合建模等方面开展了一系列的研究，提高了血压估计精度。为无袖带连续血压估计提供了新理论新方法，对心脑血管疾病的防治具有重要意义。在脉搏波的传感方面，开发了压阻-压电驻极体复合传感系统，展示了一种简化的脉搏波检测方法，实现了脉搏波检测在血压、心跳、呼吸等健康监测中的应用。在系统方面，开发了可穿戴多通道脉搏波同步检测系统，可根据不同的应用场景灵活配置传感器的选型和数量，用于血压等心血管健康信息的检测。项目还研究了用于血压估计的脉搏波特征的生理机制，设计了冷刺激实验和运动试验，分别改变外周总阻抗和心输出量。找出了在冷刺激阶段和运动后恢复期早期阶段发生了显著变化的脉搏波特征，解释了脉搏波特征可以用于血压估计的内在生理机制。研究了基于多阶多元脉搏波特征数据融合建模的血压估计方法，并提出了一种基于重要性和稳定性评分的特征选择方法。与最新的两项研究相比，采用本课题的方法可使收缩压估计的平均绝对误差的平均值从5.91 mmHg, 5.77 mmHg，降到4.59 mmHg；误差标准差的平均值从7.51 mmHg，7.39 mmHg，降到6.00 mmHg。舒张压估计的平均绝对误差的平均值从3.03mmHg, 2.99 mmHg，降到2.47 mmHg；误差标准差的平均值从3.96 mmHg，3.91 mmHg，降到3.30 mmHg。且仅使用本研究筛选出的重要性和稳定性较高的重要脉搏波特征，其血压估计精度基本与以往研究相当。但是本研究仅需脉搏波，而以往的研究需要联合使用心电和脉搏波。说明在个性化无袖带连续血压估计的应用中，仅使用具有高度重要性和稳定性的脉搏波特征具有巨大的潜力。

内容获取失败，请点击重试