Estimating Blood Pressure from the Photoplethysmogram Signal and Demographic Features Using Machine Learning Techniques

Sensors, Journal Year: 2020, Volume and Issue: 20(11), P. 3127 - 3127

Published: June 1, 2020

Hypertension is a potentially unsafe health ailment, which can be indicated directly from the Blood pressure (BP). always leads to other complications. Continuous monitoring of BP very important; however, cuff-based measurements are discrete and uncomfortable user. To address this need, cuff-less, continuous non-invasive measurement system proposed using Photoplethysmogram (PPG) signal demographic features machine learning (ML) algorithms. PPG signals were acquired 219 subjects, undergo pre-processing feature extraction steps. Time, frequency time-frequency domain extracted their derivative signals. Feature selection techniques used reduce computational complexity decrease chance over-fitting ML The then train evaluate best regression models selected for Systolic (SBP) Diastolic (DBP) estimation individually. Gaussian Process Regression (GPR) along with ReliefF algorithm outperforms algorithms in estimating SBP DBP root-mean-square error (RMSE) 6.74 3.59 respectively. This model implemented hardware systems continuously monitor avoid any critical conditions due sudden changes.

Language: Английский

---

Evaluation of the Accuracy of Cuffless Blood Pressure Measurement Devices: Challenges and Proposals

Hypertension, Journal Year: 2021, Volume and Issue: 78(5), P. 1161 - 1167

Published: Sept. 13, 2021

Several novel cuffless wearable devices and smartphone applications claiming that they can measure blood pressure (BP) are appearing on the market. These technologies very attractive promising, with increasing interest among health care professionals for their potential use. Moreover, becoming popular patients hypertension healthy people. However, at present time, there serious issues about BP measurement accuracy of 2021 European Society Hypertension Guidelines do not recommend them clinical Cuffless have special validation issues, which been recently recognized. It is important to note 2018 Universal Standard automated developed by American Association Advancement Medical Instrumentation, Hypertension, International Organization Standardization inappropriate devices. Unfortunately, an number publications presenting data devices, inadequate methodology potentially misleading conclusions. The objective this review facilitate understanding capabilities limitations emerging First, types these described. Then, unique challenges in evaluating explained. Studies from literature computer simulations employed illustrate challenges. Finally, proposals given how evaluate including interpreting relevant study results.

Language: Английский

---

Cuffless blood pressure measuring devices: review and statement by the European Society of Hypertension Working Group on Blood Pressure Monitoring and Cardiovascular Variability

Journal of Hypertension, Journal Year: 2022, Volume and Issue: 40(8), P. 1449 - 1460

Published: June 16, 2022

Background: Many cuffless blood pressure (BP) measuring devices are currently on the market claiming that they provide accurate BP measurements. These technologies have considerable potential to improve awareness, treatment, and management of hypertension. However, recent guidelines by European Society Hypertension do not recommend for diagnosis Objective: This statement Working Group Monitoring Cardiovascular Variability presents types technologies, issues in their validation, recommendations clinical practice. Statements: Cuffless monitors constitute a wide heterogeneous group novel with different intended uses. specific accuracy issues, which render established validation protocols cuff inadequate validation. In 2014, Institute Electrical Electronics Engineers published standard devices, International Organization Standardization is developing another standard. The should address related need individual calibration, stability measurements post ability track changes, implementation machine learning technology. Clinical field investigations may also be considered regarding readings investigated. Conclusion: changing fundamental questions accuracy, performance, carefully addressed before can recommended use.

Language: Английский

---

Cuffless Blood Pressure Measurement

Annual Review of Biomedical Engineering, Journal Year: 2022, Volume and Issue: 24(1), P. 203 - 230

Published: April 1, 2022

Cuffless blood pressure (BP) measurement has become a popular field due to clinical need and technological opportunity. However, no method been broadly accepted hitherto. The objective of this review is accelerate progress in the development application cuffless BP methods. We begin by describing principles conventional measurement, outstanding hypertension/hypotension problems that could be addressed with methods, recent advances, including smartphone proliferation wearable sensing, are driving field. then present all major methods under investigation, their current evidence. Our presentation includes calibrated (i.e., pulse transit time, wave analysis, facial video processing) uncalibrated oscillometry, ultrasound, volume control). can offer convenience advantages, whereas do not require periodic cuff device usage or demographic inputs. conclude summarizing highlighting potentially useful future research directions.

Language: Английский

---

Diagnostic Features and Potential Applications of PPG Signal in Healthcare: A Systematic Review

Healthcare, Journal Year: 2022, Volume and Issue: 10(3), P. 547 - 547

Published: March 16, 2022

Recent research indicates that Photoplethysmography (PPG) signals carry more information than oxygen saturation level (SpO2) and can be utilized for affordable, fast, noninvasive healthcare applications. All these encourage the researchers to estimate its feasibility as an alternative many expansive, time-wasting, invasive methods. This systematic review discusses current literature on diagnostic features of PPG signal their applications might present a potential venue adapted into health fitness aspects human life. The methodology is based Preferred Reporting Items Systematic Reviews Meta-Analysis (PRISMA) guidelines 2020. To this aim, papers from 1981 date are reviewed categorized in terms application domain. Along with consolidated areas, recent topics growing popularity also discovered. We highlight impact using individual's quality life public health. state-of-the-art studies suggest years come wearables will become pervasive fields medical practices, main domains include cardiology, respiratory, neurology, fitness. Main operation challenges, including performance robustness obstacles, identified.

Language: Английский

---

A review of machine learning in hypertension detection and blood pressure estimation based on clinical and physiological data

Biomedical Signal Processing and Control, Journal Year: 2021, Volume and Issue: 68, P. 102813 - 102813

Published: June 1, 2021

The use of machine learning techniques in medicine has increased recent years due to a rise publicly available datasets. These have been applied high blood pressure studies following two approaches: hypertension stage classification based on clinical data and estimation related physiological signals. This paper presents literature review such studies. We aimed identify the best practices, challenges, opportunities developing models detect or estimate using Hence, we identified examined techniques, datasets, predictors used previous feature selection reduce model complexity are also reviewed. found lack combining socio-demographic with signals, despite correlation photoplethysmography waveforms variables as age, gender, body mass index, heart rate. Therefore, there is an opportunity increase performance by both types for detection monitoring.

Language: Английский

---

Assessing hemodynamics from the photoplethysmogram to gain insights into vascular age: a review from VascAgeNet

AJP Heart and Circulatory Physiology, Journal Year: 2021, Volume and Issue: 322(4), P. H493 - H522

Published: Dec. 24, 2021

The photoplethysmogram (PPG) signal is widely measured by clinical and consumer devices, it emerging as a potential tool for assessing vascular age. shape timing of the PPG pulse wave are both influenced normal aging, changes in arterial stiffness blood pressure, atherosclerosis. This review summarizes research into age from PPG. Three categories approaches described:

Language: Английский

---

Cuff-Less Blood Pressure Estimation From Photoplethysmography via Visibility Graph and Transfer Learning

IEEE Journal of Biomedical and Health Informatics, Journal Year: 2021, Volume and Issue: 26(5), P. 2075 - 2085

Published: Nov. 16, 2021

This paper presents a new solution that enables the use of transfer learning for cuff-less blood pressure (BP) monitoring via short duration photoplethysmogram (PPG). The proposed method estimates BP with low computational budget by 1) creating images from segments PPG visibility graph (VG), hence, preserving temporal information waveform, 2) using pre-trained deep convolutional neural network (CNN) to extract feature vectors VG images, and 3) solving weights bias between reference BPs ridge regression. Using University California Irvine (UCI) database consisting 348 records, achieves best error performance $0.00\pm 8.46$ mmHg systolic (SBP), notation="LaTeX">$-0.04\pm 5.36$ diastolic (DBP), respectively, in terms mean (ME) standard deviation (SD) error, ranking grade B SBP A DBP under British Hypertension Society (BHS) protocol. Our novel data-driven offers computationally-efficient end-to-end rapid user-friendly PPG-based estimation.

Language: Английский

---

Features from the photoplethysmogram and the electrocardiogram for estimating changes in blood pressure

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: Jan. 18, 2023

There is a growing emphasis being placed on the potential for cuffless blood pressure (BP) estimation through modelling of morphological features from photoplethysmogram (PPG) and electrocardiogram (ECG). However, appropriate models to use remain unclear. We investigated best available PPG ECG BP using both linear non-linear machine learning models. conducted clinical study in which changes ([Formula: see text]BP) were induced by an infusion phenylephrine 30 healthy volunteers (53.8% female, 28.0 (9.0) years old). extracted large diverse set assessed their individual importance estimating [Formula: text]BP Shapley additive explanation values ranking coefficient. trained, tuned, evaluated (ordinary least squares, OLS) (random forest, RF) estimate nested leave-one-subject-out cross-validation framework. reported results as correlation coefficient text]), root mean squared error (RMSE), absolute (MAE). The RF model significantly text]) outperformed OLS signals across all performance metrics. Estimating text]SBP alone text] = 0.86 (0.23), RMSE 5.66 (4.76) mmHg, MAE 4.86 (4.29) mmHg) performed better than 0.69 (0.45), 6.79 5.28 (4.57) mmHg), text]. highest largely modelled increasing reflected wave interference driven arterial stiffness. This finding was supported observed waveform response infusion. number required accurate estimation, highlighting high complexity problem. conclude that may be further explored single source, cuffless, estimator. not justified. Non-linear perform they are able incorporate interactions between feature demographics. demographics adequately account unique individualised relationship BP.

Language: Английский

---

A novel deep learning method for predicting athletes’ health using wearable sensors and recurrent neural networks

Decision Analytics Journal, Journal Year: 2023, Volume and Issue: 7, P. 100213 - 100213

Published: March 31, 2023

Good health is extremely important for athletes who engage in strenuous physical activities, such as football. They must develop a healthy body before participating vigorous activities and competitions. Although researchers have presented wide range of analytical approaches emphasizing athlete health, only small percentage completed studies used neural networks. In this study, we propose novel technique predicting football players' using wearable technology recurrent The proposed system monitors the real-time, making it one first applications sensors athletes' conditioning health. Health prediction results are provided after time-step data entered into network, subsequent deep features obtained from that data. Several trials conducted investigation, outcomes determined by information acquired about simulation illustrate practicality dependability approach. algorithms developed study can serve foundation data-driven monitoring training.

Language: Английский

---