scholarly journals A Novel Clustering-Based Algorithm for Continuous and Noninvasive Cuff-Less Blood Pressure Estimation

2022 ◽  
Vol 2022 ◽  
pp. 1-13
Author(s):  
Ali Farki ◽  
Reza Baradaran Kazemzadeh ◽  
Elham Akhondzadeh Noughabi
Keyword(s):  
Blood Pressure ◽  
Clustering Algorithm ◽  
Machine Learning Algorithms ◽  
Weighted Averaging ◽  
Gradient Boosting ◽  
Data Set ◽  
Arterial Blood ◽  
Dependent Variables ◽  
Pressure Estimation ◽  
Blood Pressure Estimation

Extensive research has been performed on continuous and noninvasive cuff-less blood pressure (BP) measurement using artificial intelligence algorithms. This approach involves extracting certain features from physiological signals, such as ECG, PPG, ICG, and BCG, as independent variables and extracting features from arterial blood pressure (ABP) signals as dependent variables and then using machine-learning algorithms to develop a blood pressure estimation model based on these data. The greatest challenge of this field is the insufficient accuracy of estimation models. This paper proposes a novel blood pressure estimation method with a clustering step for accuracy improvement. The proposed method involves extracting pulse transit time (PTT), PPG intensity ratio (PIR), and heart rate (HR) features from electrocardiogram (ECG) and photoplethysmogram (PPG) signals as the inputs of clustering and regression, extracting systolic blood pressure (SBP) and diastolic blood pressure (DBP) features from ABP signals as dependent variables, and finally developing regression models by applying gradient boosting regression (GBR), random forest regression (RFR), and multilayer perceptron regression (MLP) on each cluster. The method was implemented using the MIMIC-II data set with the silhouette criterion used to determine the optimal number of clusters. The results showed that because of the inconsistency, high dispersion, and multitrend behavior of the extracted features vectors, the accuracy can be significantly improved by running a clustering algorithm and then developing a regression model on each cluster and finally weighted averaging of the results based on the error of each cluster. When implemented with 5 clusters and GBR, this approach yielded an MAE of 2.56 for SBP estimates and 2.23 for DBP estimates, which were significantly better than the best results without clustering (DBP: 6.27, SBP: 6.36).

scholarly journals Continuous Blood Pressure Estimation Using Exclusively Photopletysmography by LSTM-Based Signal-to-Signal Translation

Sensors ◽  
2021 ◽  
Vol 21 (9) ◽  
pp. 2952
Author(s):  
Latifa Nabila Harfiya ◽  
Ching-Chun Chang ◽  
Yung-Hui Li
Keyword(s):  
Blood Pressure ◽  
Absolute Error ◽  
Estimation Methods ◽  
British Hypertension Society ◽  
Arterial Blood ◽  
Proposed Model ◽  
Pressure Estimation ◽  
Effective Signal ◽  
Blood Pressure Estimation ◽  
Continuous Blood Pressure

Monitoring continuous BP signal is an important issue, because blood pressure (BP) varies over days, minutes, or even seconds for short-term cases. Most of photoplethysmography (PPG)-based BP estimation methods are susceptible to noise and only provides systolic blood pressure (SBP) and diastolic blood pressure (DBP) prediction. Here, instead of estimating a discrete value, we focus on different perspectives to estimate the whole waveform of BP. We propose a novel deep learning model to learn how to perform signal-to-signal translation from PPG to arterial blood pressure (ABP). Furthermore, using a raw PPG signal only as the input, the output of the proposed model is a continuous ABP signal. Based on the translated ABP signal, we extract the SBP and DBP values accordingly to ease the comparative evaluation. Our prediction results achieve average absolute error under 5 mmHg, with 70% confidence for SBP and 95% confidence for DBP without complex feature engineering. These results fulfill the standard from Association for the Advancement of Medical Instrumentation (AAMI) and the British Hypertension Society (BHS) with grade A. From the results, we believe that our model is applicable and potentially boosts the accuracy of an effective signal-to-signal continuous blood pressure estimation.

scholarly journals Non-Invasive Risk Stratification of Hypertension: A Systematic Comparison of Machine Learning Algorithms

2020 ◽  
Vol 9 (3) ◽  
pp. 34
Author(s):  
Giovanna Sannino ◽  
Ivanoe De Falco ◽  
Giuseppe De Pietro
Keyword(s):  
Machine Learning ◽  
Blood Pressure ◽  
Risk Stratification ◽  
Learning Algorithms ◽  
Circulatory System ◽  
Machine Learning Algorithms ◽  
Learning Mechanisms ◽  
Data Set ◽  
Non Invasive ◽  
Blood Pressure Estimation

One of the most important physiological parameters of the cardiovascular circulatory system is Blood Pressure. Several diseases are related to long-term abnormal blood pressure, i.e., hypertension; therefore, the early detection and assessment of this condition are crucial. The identification of hypertension, and, even more the evaluation of its risk stratification, by using wearable monitoring devices are now more realistic thanks to the advancements in Internet of Things, the improvements of digital sensors that are becoming more and more miniaturized, and the development of new signal processing and machine learning algorithms. In this scenario, a suitable biomedical signal is represented by the PhotoPlethysmoGraphy (PPG) signal. It can be acquired by using a simple, cheap, and wearable device, and can be used to evaluate several aspects of the cardiovascular system, e.g., the detection of abnormal heart rate, respiration rate, blood pressure, oxygen saturation, and so on. In this paper, we take into account the Cuff-Less Blood Pressure Estimation Data Set that contains, among others, PPG signals coming from a set of subjects, as well as the Blood Pressure values of the latter that is the hypertension level. Our aim is to investigate whether or not machine learning methods applied to these PPG signals can provide better results for the non-invasive classification and evaluation of subjects’ hypertension levels. To this aim, we have availed ourselves of a wide set of machine learning algorithms, based on different learning mechanisms, and have compared their results in terms of the effectiveness of the classification obtained.

scholarly journals Pulse Transit Time Derivation Using Xiphoidal and Carotid Seismocardiograms

2017 ◽  
Author(s):  
Ajay K. Verma ◽  
John Zanetti ◽  
Reza Fazel-Rezai ◽  
Kouhyar Tavakolian
Keyword(s):  
Blood Pressure ◽  
Aortic Valve ◽  
Transit Time ◽  
Pulse Wave ◽  
Blood Pressure Measurement ◽  
Pulse Transit Time ◽  
Non Invasive ◽  
Arterial Blood ◽  
Pressure Estimation ◽  
Blood Pressure Estimation

Blood pressure is an indicator of a cardiovascular functioning and could provide early symptoms of cardiovascular system impairment. Blood pressure measurement using catheterization technique is considered the gold standard for blood pressure measurement [1]. However, due its invasive nature and complexity, non-invasive techniques of blood pressure estimation such as auscultation, oscillometry, and volume clamping have gained wide popularity [1]. While these non-invasive cuff based methodologies provide a good estimate of blood pressure, they are limited by their inability to provide a continuous estimate of blood pressure [1–2]. Continuous blood pressure estimate is critical for monitoring cardiovascular diseases such as hypertension and heart failure. Pulse transit time (PTT) is a time taken by a pulse wave to travel between a proximal and distal arterial site [3]. The speed at which pulse wave travels in the artery has been found to be proportional to blood pressure [1, 3]. A rise in blood pressure would cause blood vessels to increase in diameter resulting in a stiffer arterial wall and shorter PTT [1–3]. To avail such relationship with blood pressure, PTT has been extensively used as a marker of arterial elasticity and a non-invasive surrogate for arterial blood pressure estimation. Typically, a combination of electrocardiogram (ECG) and photoplethysmogram (PPG) or arterial blood pressure (ABP) signal is used for the purpose of blood pressure estimation [3], where the proximal and distal timing of PTT (also referred as pulse arrival time, PAT) is marked by R peak of ECG and a foot/peak of a PPG, respectively. In the literature, it has been shown that PAT derived using ECG-PPG combination infers an inaccurate estimate of blood pressure due to the inclusion of isovolumetric contraction period [1–3, 4]. Seismocardiogram (SCG) is a recording of chest acceleration due to heart movement, from which the opening and closing of the aortic valve can be obtained [5]. There is a distinct point on the dorso-ventral SCG signal that marks the opening of the aortic valve (annotated as AO). In the literature, AO has been proposed for timing the onset of the proximal pulse of the wave [6–8]. A combination of AO as a proximal pulse and PPG as a distal pulse has been used to derive pulse transit time and is shown to be correlated with blood pressure [7]. Ballistocardiogram (BCG) which is a measure of recoil forces of a human body in response to pumping of blood in blood vessels has also been explored as an alternative to ECG for timing proximal pulse [5, 9]. Use of SCG or BCG for timing the proximal point of a pulse can overcome the limitation of ECG-based PTT computation [6–7, 9]. However, a limitation of current blood pressure estimation systems is the requirement of two morphologically different signals, one for annotating the proximal (ECG, SCG, BCG) and other for annotating the distal (PPG, ABP) timing of a pulse wave. In the current research, we introduce a methodology to derive PTT from seismocardiograms alone. Two accelerometers were used for such purpose, one was placed on the xiphoid process of the sternum (marks proximal timing) and the other one was placed on the external carotid artery (marks distal timing). PTT was derived as a time taken by a pulse wave to travel between AO of both the xiphoidal and carotid SCG.

scholarly journals Progressive Dynamic Time Warping for Noninvasive Blood Pressure Estimation

2020 ◽  
Vol 6 (3) ◽  
pp. 579-582
Author(s):  
Alexandru-Gabriel Pielmus ◽  
Michael Klum ◽  
Timo Tigges ◽  
Reinhold Orglmeister ◽  
Mike Urban
Keyword(s):  
Blood Pressure ◽  
Dynamic Time Warping ◽  
Time Warping ◽  
Physical Strain ◽  
Test Protocol ◽  
Arterial Blood ◽  
Pulse Waves ◽  
Pressure Estimation ◽  
Dynamic Time ◽  
Blood Pressure Estimation

AbstractArterial blood pressure is one of the most important cardiovascular parameters. Yet, current-generation devices for continuous, noninvasive acquisition are few, expensive and bulky. Novel signal processing applied to easily acquired unimodal signals can alleviate this issue, reducing size, cost and expanding the use of such devices to ambulatory, everyday settings. The features of pulse waves acquired by photo- or impedance-plethysmography can be used to estimate the underlying blood pressure. We present a progressive dynamic time warping algorithm, which implicitly parametrizes the morphological changes in these waves. This warping method is universally applicable to most pulse wave shapes, as it is largely independent of fiducial point detection or explicit parametrization. The algorithm performance is validated in a feature selection and regression framework against a continuous, noninvasive Finapres NOVA monitor, regarding systolic, mean and diastolic pressures during a light physical strain test protocol on four clinically healthy subjects (age18- 33, one female). The obtained mean error is 2.13 mmHg, the mean absolute error is 5.4 mmHg and the standard deviation is 5.6 mmHg. These results improve on our previous work on dynamic time warping. Using single-sensor, peripherally acquired pulse waves, progressive dynamic time warping can thus improve the flexibility of noninvasive, continuous blood pressure estimation.

Sign in / Sign up

Export Citation Format

Share Document