基于LightGBM的血压检测方法研究

为了提高血压的检测精度,本研究基于光电容积脉搏波(photoplethysmography,PPG)波形,提取该波形的时域特征参数15个,并在此基础上,加入前一时刻的血压值作为特征,通过LightGBM(light gradient boosting machine,LightGBM)模型进行10折交叉验证。结果表明,舒张压和收缩压的平均绝对误差均小于5 mmHg,均方根误差均小于8 mmHg,满足了AAMI国际电子血压计标准。本研究所用的检测方法简单高效、易于实现,有助于个性化血压检测的应用以及临床应用。     

皮质下动脉硬化性脑病血压及心率变异性分析

目的和方法：采用动态检测技术观察３１例皮质下动脉硬化性脑病和对照组３１例健康人血压和心率的变异性。结果：患者均存在动态血压异常，其中平均收缩压、平均舒张压超过正常值者２７例（８７％），夜间基底血压异常升高２２例（７１％），血压异常波动２０例（６４％），昼夜节律逆转４例（１３％）。患者心率变异时域指标２４小时内全部正常心动周期的标准差、２４小时内５分钟节段平均正常心动周期的标准差、在一定时间内相邻两正常心动周期差值大于５０毫秒的个数所占的百分比均明显低于对照组，收缩压与心率无显著相关性。结论：提示血压波动及持续的夜间升压现象在该病中起着一定作用，而这种异常可能与患者植物神经系统功能失调有关     

基于心电信号-光电容积脉博波信号多特征参数无创血压监测系统设计

针对传统血压计袖带使用给患者带来的不便以及现有的利用单一脉搏波传导时间(pulse wave translation time,PWTT)进行无创血压监测稳定性较差等问题,本研究提出基于心电信号(electrocardiograph,ECG)-电容积脉搏波信号(photoplethysmo graphy,PPG)多特征参数融合的无创血压监测、检测方法,并设计相应的血压监测系统。系统通过采集ECG和PPG提取特征并建立血压计算模型。模型以PPG传导时间PWTT、PPG上升支时间与PPG周期之比λ和PPG峰峰值φ为自变量,实测血压值为因变量进行回归分析,得到收缩压(systolic blood pressure,SBP)和舒张压(diastolic blood pressure,DBP)回归方程,实现无袖带血压连续监测、检测。利用该方法对30名志愿者进行验证实验,结果显示,本系统所测SBP与实测SBP绝对误差均值E_d=2. 3067 mmHg,标准差S_D=1. 4633 mmHg; DBP绝对误差均值E_d=2. 4133 mmHg,标准差S_d=1. 9894 mmHg,符合AAMI推荐标准。通过Bland-Altman一致性分析,本系统与水银血压计所测SBP、DBP相关系数分别为0. 9878、0. 9730,一致性界限分别为(-4. 9300,5. 6770)、(-6. 1950,6. 0620)。实验结果证明,通过本方法可以实现无袖带血压的连续监测、检测,为便携式可穿戴设备血压实时监测及临床无袖带血压检测提供了可参考的实施方法。     

基于脉搏-心电的便携式血压仪的研制

设计一种体积小巧,测量过程无任何束缚的基于脉搏波传导时间(pulse wave transit time,PWTT)的血压仪。从反射式容积脉搏波(photoplethysmography,PPG)和心电波形(electrocardiogram,ECG)中提取脉搏波传导时间,再综合心率、每搏输出量和外周阻力对血压的影响,通过回归分析建立血压模型,最终实现血压测量。应用样机对20名志愿者进行血压测量,同时以传统水银血压计的测量结果作为标准,结果显示收缩压和舒张压的95%一致性界限分别为(-8.3,11.6),(-9.9,12.7),说明两种方法所测的血压值有较好的一致性。血压仪样机实体小巧,使用方便,使用过程没有任何束缚,同时在理论上对基于脉搏波传导时间的血压测量方法进行优化,实现了收缩压和舒张压的测量。     

基于深度学习的ECG/PPG血压测量方法

近年来,基于ECG/PPG信号的血压测量方法已经在某些可穿戴设备上实现.但此类方法的检测精度尚未达到相关国际标准.本研究利用深度神经网络模型,对基于ECG/PPG信号的血压测量方法进行了深入研究,提高了该类方法的检测精度.首先,采用基于小波包的模态分解技术,从PPG信号中提取出心脏信号和呼吸信号,并将其与ECG信号同步...     

基于脉搏波传导时间的连续血压监测系统

为了实现无创连续血压测量,提出了一种基于脉搏波传导时间(pulse transit time,PTT)的连续血压测量方案。通过同步采集心电(electrocardiogram,ECG)信号与光电脉搏波(photoplethysmograph,PPG)信号,以ECG的R波峰值点作为PTT的开始点,PPG信号的最大值点作为PTT的结束点,得到PTT,与水银血压计测得舒张压(diastolic blood pressure,DBP)与收缩压(systolic blood pressure,SBP)进行回归分析,得到了DBP和SBP的数学模型。利用该方法对41名身体健康的青年人进行实验,利用本方案得到的SBP与水银血压计的相关系数为0.82,其差值的平均数与标准偏差为0.15±2.05 mm Hg;得到的DBP与水银血压计的相关系数为0.73,其差值的平均数与标准偏差为0.12±2.16 mm Hg。利用Bland-Altman差值法对本系统血压测量方法与水银血压计测血压方法进行一致性检验,结果显示两种血压测量方法具有很好的一致性。     

ACEI及ARB对寒冷所致的高血压病人循环系统反应的干预

目的 为了探讨血管紧张素转换酶抑制剂（ACEI）和血管紧张素受体阻滞剂（ARB）对寒冷所致的高血压病人循环系统反应的影响。方法 111例血压控制稳定的高血压病人分成ACEI治疗、ARB治疗组及对照组3组，观察气温骤降前后血压和心率变化。结果 ①3组病人气温骤降前血压控制均达到目标血压；②气温骤降后，ARB治疗组收缩压、舒张压无明显上升，其余各组收缩压、舒张压均明显上升；心率3组均无明显变化；③与对照组比较，ARB组寒冷所致的收缩压上升幅度明显减少；④与对照组比较，ACEI及ARB组寒冷所致的舒张压上升均轻度减少（P＞0．05），但无统计差异，心率变化亦无差异。结论 ACEI未能有效地阻断高血压病人寒冷所致的循环系统反应，亦不比其它降压药优胜；但ARB能有效地阻断高血压病人寒冷所致的循环系统反应，且明显优胜于其它降压药。     

高血压无左室肥厚患者舒张功能改变的观察

目的：探讨高血压无左室肥厚患者舒张功能的改变。方法：高血压无左室肥厚患者、高血压伴左室肥厚患者与正常血压者各４０例,均进行超声心动图检查和动态血压检测。结果：高血压伴左室肥厚组、高血压无左室肥厚组与正常对照组比较,二尖瓣血流Ｅ峰、Ａ峰、Ｅ／Ａ比率、等容舒张时间及舒张早期减速度均有明显差异（Ｐ＜００５,Ｐ＜００１）,而高血压左室肥厚组与高血压无左室肥厚组两组间,则无明显差异（Ｐ＞００５）。３个组的左室射血分数（ＥＦ）无明显差异（Ｐ＞００５）。动态血压监测显示,高血压左室肥厚组与高血压无左室肥厚组２４ｈ平均收缩压和舒张压、白昼平均收缩压和舒张压、血压负荷有显著差异（Ｐ＜００５,Ｐ＜０．０１）,而夜间平均收缩压、舒张压无显著差异（Ｐ＞００５）。结论：高血压患者在出现左室肥厚前可出现舒张功能异常,可能与夜间血压持续升高有关。     

基于视频流的非接触式连续血压检测算法研究

高血压是危害人类健康的首要疾病，方便准确的血压测量方法将有助于高血压的防控。本文提出了一种基于面部视频信号的连续血压测量方法。采用颜色失真滤波与独立成分分析法提取面部视频信号中感兴趣区域的视频脉搏波，基于时频域以及生理学原理对脉搏波进行多维特征提取；设计了一种集成特征选择方法提取具有通用性的最优特征子集；比较基于粒子群优化的Elman神经网络、支持向量机与深度信念网络所建立的单人血压测量模型；采用支持向量回归算法构建通用血压预测模型，并与真实血压值进行比较与评价。实验结果表明：基于面部视频的血压测量结果与标准血压值具有较好的一致性，由视频估计出的收缩压与标准收缩压的平均绝对误差（MAE）为4.9 mm Hg，标准差（STD）为5.9 mm Hg；舒张压的MAE为4.6 mm Hg,STD为5.0 mm Hg，符合AAMI标准。本文所提出的基于视频流的非接触式血压检测方法可以用于血压的测量。     

人体血压发育的遗传学研究

以１０种血型物质反应鉴定了１１１对双生子的卵性，计同卵双生（ＭＺ）６８对，异卵双生（ＤＺ）４３对以及依ＭＺ组的性别与年龄随机配成的无关对照（ＣＰ）６８对。作者对比三组学生进行了收缩压、舒张压的测定，并计算了三组人群血压的偶内差均值、百分比差均值、相似度系数和遗传度值。对比三组人群收缩压和舒张压的相似程度，依次为ＭＺ＞ＤＺ＞ＣＰ，提示发育期血压与遗传有关。本文依ＣＰ组与ＭＺ组资料所得血压遗传度值（ｈ２）收缩压为０．８３，舒张压为０．８１。进一步证实我市发育期血压值主要受遗传因素决定，但环境因素不容忽视。指出对有高血压家族倾向的小儿，从儿童期就应开始进行防止或推迟高血压发病的健康指导。     

基于脉搏波信号和血管弹性腔模型的动脉血压连续测量方法

目的为满足健康监护中的连续测量血压的要求,研究并实现一种基于脉搏波信号和血管弹性腔模型的动脉血压拟合计算方法。方法利用自制的穿戴式人体生理参数监测系统收集测试对象的脉搏波信号、心电信号以及血压数据。根据心电信号与脉搏波信号的时间关系,推导出收缩压和脉搏波传导时间的回归分析方程,而舒张压的测量,则是通过脉搏波的波形系数分析以及血管单弹性腔模型的参数计算完成。结果试验结果表明,该方法血压测量结果的平均偏差和标准偏差为(0.51±0.74)kPa([384±5.54)mmHg],达到了美国医疗仪器促进协会建议的(0.665±1.064)kPa([5±8)mmHg]标准。结论结合脉搏波信号和弹性腔模型可以估算人体血压值,为连续血压测量提供了新的实现方法。     

基于光电容积脉搏波描记法的无创连续血压测量

为了摆脱传统血压计充气袖带的束缚、实现长期连续的血压监测,很多学者开展了基于PPG实现无创、连续血压监测的研究。阐述基于光电容积脉搏波描记法(PPG)实现无创、连续血压测量的基本原理。将当前基于PPG无创血压监测的研究分为3类,分别为心电(ECG)与PPG结合的血压测量技术、两路PPG结合的血压测量技术、脉搏波特征参数血压测量技术,分析这3类技术的基本测量原理、测量精度及其优缺点。在此基础上,论述基于PPG实现无创、连续监测血压的发展方向。     

AVR单片机下脉搏传导时间与收缩压的关系

目的利用MIMIC Datebase(mimicdb)数据分析有创收缩压和脉搏传导时间(pulse transit time,PTT)的关系。制作AVR信号采集系统,并在PC下采集心电和脉搏的实时信号,分析无创收缩压和PTT的关系。方法利用Matlab软件分析MIMIC数据库内5个个体的数据,得出PTT和有创收缩压的相关系数。利用AVR单片机采集用6名正常血压的男性(年龄27~31周岁)的心电和脉搏信号,用蓝牙发送至计算机。通过两个月对比试验,比较PTT和收缩压之间的关系是否发生很大的变化。结果PTT同有创和无创收缩压都有较好的线性相关性,波谷作为PTT特征点和收缩压有更好的线性关系。结论 AVR数据采集系统小巧、便携,可应用于家用监护。     

Continuous and noninvasive measurement of systolic and diastolic blood pressure by one mathematical model with the same model parameters and two separate pulse wave velocities

There is keen interest in continuous and noninvasive blood pressure (BP) measurement. However, many technologies have a shortcoming of complex mechanical structure. In our study, two arterial pulses are acquired by photoplethysmography (PPG) at ear and toe in order to explore a new method of measuring BP by pulse wave velocity (PWV). We previously validated and reported a BP-PWV mathematical model with measurements from humans with no evidence of cardiovascular disease, but were only able to determine PWV related to diastolic blood pressure (DBP). In this paper, we propose methods of identifying pulse transmit time (PTT) in low, normal and high systolic blood pressure (SBP) conditions. By averaging the PTT’s of incident wave and reflected wave for non-systematic error reduction, we obtain a PWV that is suitable for estimating SBP. SBP and DBP are estimated by two separate PWV’s based on the previously calibrated models. Experimental measurements are conducted on 26 subjects (age 19 ± 1 and 60 ± 1) with no evidence of cardiovascular disease. The measurement errors (Mean Deviation = 2.16 mmHg (SBP) and 1.49 mmHg (DBP); Standard Deviation = 6.23 mmHg (SBP) and 6.51 mmHg (DBP)) satisfy the accuracy criteria of Association for the Advancement of Medical Instrumentation. The results verify that SBP and DBP can be estimated by one mathematical model with the same model parameters and two separate PWV’s.     

Verylow frequency variability in arterial blood pressure and blood volume pulse

Several parameters of the cardiovascular system fluctuate spontaenously owing to the activity of the autonomic nervous system. In the study, the simultaneous very low frequency (VLF) fluctuations of the arterial blood pressure, the tissue blood content and the tissue blood volume pulse are investigated. The latter two parameters are derived from the baseline BL and the amplitude AM of the photoplethysmographic (PPG) signal, measured on the fingertips of 20 healthy male subjects: the changes in the PPG parameters AM and BV, defined by BV=const.-BL, are related to the change in the tissue blood volume pulse and the total tissue blood volume, respectively. The VLF fluctuations in BV and AM are directly correlated, those of AM preceding those of BV by 4–13 heart-beats. The VLF fluctuations in the systolic (SBP) and the diastolic (DBP) blood pressure are inversely correlated to those of AM and BV, those of AM preceding those of SBP and lagging behing those of DBP by about one heart-beat. For most subjects, the period P of the PPG pulse, which is equal to the cardiac cycle period, directly correlates with AM and BV and inversely correlates with DBP and SBP. On average, the fluctuations fluctuations in tissue blood volume, systolic blood volume pulse, diastolic and systolic blood pressure, and heart period, together with their interrelationship, can provide a better understanding of the autonomic nervous control of the peripheral circulation and a potential tool for the evaluation of its function.     

基于舒张期时间的个体化血压建模方法研究

本研究提出了一种基于脉搏波舒张期时间(diastolic time,DT)的无袖带式血压估计方法,能够有效提升可穿戴式设备的血压测量精度。该方案首先从脉搏波(photoplethysmography,PPG)中提取DT,再利用线性回归方法建立个体化血压估计模型。本研究对采集的30个样本的实验数据,进行血压建模并加以验证,实验结果显示收缩压的平均误差为1.859 mmHg,标准差为5.640 mmHg;舒张压的平均误差为1.049mmHg,标准差为6.107 mmHg,证明了该方案的可行性。通过将该方案的估计结果与基于脉搏波传导时间(pulse transit time,PTT)的血压模型的估算结果进行对比研究,结果表明交感神经活动较强时,基于DT的血压模型能够有效提升血压估计精度。     

Pulse transit time as a surrogate measure of changes in systolic arterial pressure in children during sleep

Pulse transit time has been proposed as a surrogate measure of systolic arterial pressure, as it is dependent upon arterial stiffness. Past research has shown that pulse transit time has a significant inverse relationship to systolic arterial pressure in adults; however, studies in children are limited. This study aimed to explore the relationship between systolic arterial pressure and pulse transit time in children during sleep. Twenty‐five children (13.1 ± 1.6 years, 48% male) underwent overnight polysomnography (PSG) with a simultaneous recording of continuous systolic arterial pressure and photoplethysmography. Pulse transit time was calculated as the time delay between the R‐wave peak of the electrocardiogram (ECG) to the 50% point of the upstroke of the corresponding photoplethysmography waveform; 500 beats of simultaneous systolic arterial pressure and pulse transit time were analysed in each sleep stage for each child. Pulse transit time was normalized to each subject's mean wake pulse transit time. The ability of pulse transit time to predict systolic arterial pressure change was determined by linear mixed‐effects modelling. Significant negative correlations between pulse transit time and systolic arterial pressure were found for individual children for each sleep stage [mean correlations for cohort: non‐rapid eye movement (NREM) sleep 1 and 2 r = ?0.57, slow wave sleep (SWS) r = ?0.76, REM r = ?0.65, P < 0.01 for all]. Linear mixed‐model analysis demonstrated that changes in pulse transit time were a significant predictor of changes in systolic arterial pressure for each sleep stage (P < 0.001). The model of pulse transit time‐predicted systolic arterial pressure closely tracked actual systolic arterial pressure changes over time. This study demonstrated that pulse transit time was accurate in tracking systolic arterial pressure changes over time. Thus, the use of pulse transit time as a surrogate measure of changes in systolic arterial pressure in children is a valid, non‐invasive and inexpensive method with many potential applications.     

Design and development of a new electronic sphygmomanometer

The paper introduces a new technique for the indirect measurement of the systolic and diastolic blood pressure in humans. The technique is based upon a statistically consistent relationship between the amplitude of the pulsatile pressure waveform at the systolic and diastolic points and the amplitude of pulse signals detected when the artery is fully occluded. An adaptive measurement philosophy has been implemented in the design of an electronic sphygmomanometer which, in addition to a pressure transducer, contains suitable electronic instrumentation for processing and displaying the electronic signals. Verification of overall system accuracy is accomplished with direct comparison with manual auscultatory measurements. Clinical testing of a prototype indicates a satisfactory performance; measurement errors are maintained well within proposed standards for automated sphygmomanometers.