Continuous blood pressure estimation and cardiovascular risk assessment based on wearable PPG signals

In this article, we discuss the validity of noninvasive continuous blood pressure estimation with two different types of peripheral pulse waves. Artery-blocking experiment shows that phase difference of two pulse waves at the same location is well related with blood pressure and blood flow fluctuation. Exercise-recovery experiment resulting from 16 subjects shows that phase difference varies with blood pressure with the correlation from 0.63 to 0.88 when blood pressure changes rapidly. Simulations based on a classic hemodynamic model verify the relationship between phase difference and blood pressure. However, phase difference is strongly correlated with smooth muscle state of the arterial wall as well. If smooth muscle information can be obtained by further study, phase difference can act as a promising approach to portable and wearable device for real-time blood pressure monitoring.

Continuous and noninvasive monitoring of blood pressure is of vital importance in critical care. Its role in management of patients with or at risk of cardiovascular disease is ever increasing. Pulse-transit-time based approaches for estimation of blood pressure hold potential to provide a convenient solution, however there is a need to improve their accuracy. We employ a novel two-sensor technique using photoplethysmogram and piezo sensors for a more accurate measurement of pulse transit time (PTT). PTT is then mapped to arterial blood pressure via calibration process. We implement two different techniques for the PTT measurement. The first technique uses two pulse sensors, while the second technique makes use of one pulse sensor and one piezo element. Clinical data is collected from 46 subjects in the form of systolic, diastolic and mean blood pressure to provide a reference for the calibration step in two techniques. The results show decrease in error of systolic arterial pressure (SAP) and mean arterial pressure (MAP) for the second technique. The root mean square (RMS) error, measured for the calibration equations, reduced from 7.72 mm Hg to 5.26 mm Hg for SAP and from 7.07 mm Hg to 6.30 mm Hg for MAP. RMS error, measured by leave-one-out method, reduced from 8.45 mm Hg to 5.77 mm Hg for SAP and from 7.67 mm Hg to 6.90 mm Hg for MAP.

This study evaluates the impact of capillary refill time (CRT) modulation on photoplethysmography (PPG) signals for improved non-invasive continuous blood pressure (CBP) estimation. Data from 21 healthy participants were collected, applying a standardized 9 N pressure for 15 s to induce CRT during 6-min sessions. PPG signals were segmented into 252 paired 30-s intervals (CRT-modulated and standard). Three machine learning models—ResNetCNN, LSTM, and Transformer—were validated using leave-one-subject-out (LOSO) and non-LOSO methods. CRT modulation significantly enhanced accuracy across all models. ResNetCNN showed substantial improvements, reducing mean absolute error (MAE) by up to 35.6% and mean absolute percentage error (MAPE) by up to 40.6%. LSTM and Transformer models also achieved notable accuracy gains. All models met the Association for the Advancement of Medical Instrumentation (AAMI) criteria (mean error < 5 mmHg; standard deviation < 8 mmHg). The findings suggest CRT modulation’s strong potential to improve wearable CBP monitoring, especially in resource-limited settings.

One of the indirect approaches to obtain blood pressure (BP) values is through the so-called pulse transit time (PTT) that is the time for the cardiac wave to propagate through a length of the arterial tree or between the heart and a peripheral arterial site. In this work, we investigated the performance of four latest cuffless BP estimation methods including the method proposed in our previous work by using the invasive BP as reference. The invasive BP data was collected from 10 patients during surgery. Experimental results showed that the estimation errors from the compared methods are much larger than the 5 ± 8 mmHg limits that is required for any BP measurement method or instrument to be satisfied the requirements of the AAMI Standard. Therefore, the PTT-based BP estimation methods need to be further improved and enhanced for obtaining acceptable performance.

Accurate and ambulatory measurement of blood pressure (BP) is essential for efficient diagnosis, management and prevention of cardiovascular diseases (CVDs). However, traditional cuff-based BP measurement methods provide only intermittent BP readings and can cause discomfort with the occlusive cuff. Although pulse transit time (PTT) method is promising for cuffless and continuous BP measurement, its pervasive use is restricted by its limited accuracy and requirement of placing sensors on multiple body sites. To tackle these issues, we propose a novel dual-modality arterial pulse monitoring system for continuous blood pressure measurement, which simultaneously records the pressure and photoplethysmography (PPG) signals of radial artery. The obtained signals can be used to generate a pressure-volume curve, from which the elasticity index (EI) and viscosity index (VI) can be extracted. Experiments were carried out among 7 healthy subjects with their PPG, ECG, arterial pressure wave and reference BP collected to examine the effectiveness of the proposed indexes. The results of this study demonstrate that a linear regression model combining EI and VI has significantly higher BP tracking correlation coefficient as compared to the PTT method. This suggests that the proposed system and method can potentially be used for convenient and continuous blood pressure estimation with higher accuracy.

Pulse transit time (PTT) has been identified as a potential indicator to estimate blood pressure (BP) continuously and cufflessly in the previous studies, but it is still unsatisfactory for clinical uses. In order to improve the BP estimation accuracy, another candidate signal, photoplethysmogram (PPG) has been used to estimate continuous BP. In this study, a new PPG indicator, the intensity ratio of the first derivative wave of PPG (1st dPIR), was proposed and a novel BP estimation algorithm by using both PTT and 1st dPIR was developed with an attempt to further increase the BP estimation accuracy. The performance of the proposed algorithm was examined and validated using the ECG and PPG recordings from 22 healthy subjects when they performed both mental arithmetic stress and Valsalva's manoeuvre tests that could induce their BP fluctuations. The continuous Finapres® BP were measured with a finger cuff as a referenced BP value. Our results showed that using the proposed BP estimation algorithm, the grand absolute mean and standard deviation (SD) errors between the estimated and referenced BP were 3.22±8.02 mmHg for systolic BP (SBP) and 3.13±4.82mmHg for diastolic BP (DBP). It could achieve a decrease of 0.34±0.27 mmHg in SBP and 0.33±0.44 mmHg in DBP on the grand absolute mean and SD estimation errors in comparison to the state-of-the-art PPT-based method. Findings from this study suggest that the newly proposed PPG indicator would be promising for improving the accuracy of current PTT-based continuous and cuffless BP estimation algorithms.

The aim of this study is to determine if a noninvasive and cuffless blood pressure (BP) estimation method using pulse transit time (PTT) can accurately track the changes in BP after a treadmill exercise. Experiments were conducted on 12 healthy subjects, who were directed to run on a treadmill at 10 kph for 4 minutes. For each subject, 13 trials were recorded in total, including 11 trials obtained immediately after exercise and during the 60-minute recovery period. BP was measured at the finger by the Finometer and at the brachial arteries by a sphygmomanometer respectively. PTT was determined as the time interval between the peak of the electrocardiogram R wave and the foot of the photoplethysmogram. Both PTT-based device and Finometer were calibrated initially and Finometer calibration was also performed after exercise to the sphygmomanometer. Experimental results showed that there was only occasional discrepancy between BPs measured by Finometer and PTT-based device during the recovery period, i.e., at 5 minutes after exercise (p=0.008 for systolic BP and p=0.03 for diastolic BP). At all other post exercise time spots there were no significant differences between BPs measured by Finometer and PTT-based device

Continuous and non-invasive blood pressure (BP) monitoring is crucial for effective health management in patients with arrhythmia. Cuffless BP measurement methods, utilizing electrocardiogram (ECG) and photoplethysmogram (PPG) signals combined with deep learning, have shown great promise. However, existing deep learning algorithms face challenges in representing physiological data due to the position-sensitivity of such data and the need for global context to accurately interpret it, resulting in limited accuracy. This study introduces BM-BPW, a novel deep learning-based model for continuous BP waveform estimation from ECG and PPG signals for arrhythmia patients. The bidirectional Mamba architecture, combined with position embeddings, is employed in the BM-BPW model to enable global context modeling and locally-aware understanding. A multi-scale convolutional neural network (CNN) feature extractor is also employed to capture the local features. The model was evaluated using a dataset of 36,537 beats collected from 48 arrhythmia patients, with invasive arterial waveforms as the reference. Experimental results indicate that BM-BPW effectively manages BP waveform variability caused by arrhythmia, achieving estimation errors of 0.04 ± 5.03 mmHg and -0.05 3.20 mmHg for systolic BP (SBP), and diastolic BP (DBP),±respectively. These results outperform state-of-the-art methods. This study provides an accurate and efficient solution for cuffless continuous BP waveform monitoring for arrhythmia patients, offering substantial potential for improving clinical management in this population.

A mixed attention-gated U-Net for continuous cuffless blood pressure estimation

Mixture of experts framework for continuous contactless dynamic blood pressure estimation from radar signal

This paper proposes a novel sensing system for noninvasive and continuous estimation of blood pressure (BP), which is based on a custom designed wearable 120 GHz Doppler radar sensor. With radar signals detected from brachial artery at the elbow, the moving trajectory of the pulses can be accurately demodulated with micrometer accuracy, and thus the reflective pulse transit time (RPTT) can be obtained for the prediction model to estimate blood pressure. Clinical trials have been launched with invasive blood pressure as reference. The experimental results show a great performance of the designed wearable radar sensor in the detection of weak pulse signatures. Compared to the ground truth of the invasive blood pressure, the estimated blood pressures by radar have mean errors below 3 mmHg for most subjects. The results reveal the possibility for the proposed radar-based BP estimation technique to be employed inside and outside the clinic, such as applications in the vast internet-of-things (loT) and in-home health monitoring.

BackgroundArterial catheters are often used for blood pressure monitoring in the intensive care unit (ICU), but they can cause complications. Non-invasive continuous finger blood pressure monitors could serve as an alternative. However, failure to obtain finger blood pressure signals is reported in up to 12% of ICU patients. ObjectivesOur primary objective was to identify the success rate of finger blood pressure monitoring in ICU patients. Secondary objectives were to assess whether patient admission characteristics could be used to identify patients unsuitable for non-invasive blood pressure monitoring and to determine the quality of non-invasive blood pressure waveforms. MethodsRetrospective observational study conducted in a cohort of 499 ICU patients. When available, the signal quality of the first hour of finger measurement was determined using an open-source waveform algorithm. ResultsFinger blood pressure signals were obtained in 94% of patients. These patients had a high quality blood pressure waveform for 84% of the measurement time. Patients without a finger blood pressure signal significantly more frequently had a history of kidney and vascular disease, were more often treated with inotropic agents, had lower hemoglobin levels, and had higher arterial lactate levels. ConclusionsFinger blood pressure signals were obtained in nearly all ICU patients. Significant differences in baseline characteristics between patients with and without finger blood pressure signals were found, but they were not clinically relevant. The characteristics studied could therefore not be used to identify patients unsuitable for finger blood pressure monitoring.

Compared with diastolic blood pressure (DBP) and systolic blood pressure (SBP), the blood pressure (BP) waveform contains richer physiological information that can be used for disease diagnosis. However, most models based on photoplethysmogram (PPG) signals can only estimate SBP and DBP and are susceptible to noise signals. We focus on estimating the BP waveform rather than discrete BP values. We propose a model based on a generalized regression neural network to estimate the BP waveform, SBP and DBP. This model takes the raw PPG signal as input and BP waveform as output. The SBP and DBP are extracted from the estimated BP waveform. In addition, the model contains encoders and decoders, and their role is to be responsible for the conversion between the time domain and frequency domain of the waveform. The prediction results of our model show that the mean absolute error is 3.96 ± 5.36 mmHg for SBP and 2.39 ± 3.28 mmHg for DBP, the root mean square error is 5.54 for SBP and 3.45 for DBP. These results fulfill the Association for the Advancement of Medical Instrumentation (AAMI) standard and obtain grade A according to the British Hypertension Society (BHS) standard. The results show that the proposed model can effectively estimate the BP waveform only using the raw PPG signal.

BACKGROUND: Non-invasive continuous blood pressure monitoring can provide an important reference and guidance for doctors wishing to analyze the physiological and pathological status of patients and to prevent and diagnose cardiovascular diseases in the clinical setting. Therefore, it is very important to explore a more accurate method of non-invasive continuous blood pressure measurement.OBJECTIVE: To address the shortcomings of existing blood pressure measurement models based on pulse wave transit time or pulse wave parameters, a new method of non-invasive continuous blood pressure measurement – the GA-MIV-BP neural network model – is presented.METHOD: The mean impact value (MIV) method is used to select the factors that greatly influence blood pressure from the extracted pulse wave transit time and pulse wave parameters. These factors are used as inputs, and the actual blood pressure values as outputs, to train the BP neural network model. The individual parameters are then optimized using a genetic algorithm (GA) to establish the GA-MIV-BP neural network model.RESULTS: Bland-Altman consistency analysis indicated that the measured and predicted blood pressure values were consistent and interchangeable.CONCLUSIONS: Therefore, this algorithm is of great significance to promote the clinical application of a non-invasive continuous blood pressure monitoring method.

Continuous and non-invasive monitoring of blood pressure (BP) is important to prevent hypertensive patients from stroke and heart attack. However, most of the prevalent BP devices can provide solely intermittent measurements. In this study, a novel parameter RS <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2 </sub> , defined as the time interval from the R wave of electrocardiographic (ECG) signal to the peak of second heart sound of phonocardiographic (PCG) signal, was used for BP estimation. Experiment was conducted on 66 normal and 19 abnormal subjects during resting condition. The results of this study showed that estimation error of systolic BP (SBP) and diastolic BP (DBP) was 2.0plusmn7.5 mmHg and 0.9plusmn5.1 mmHg respectively. 94.7% of the difference between measured and estimated SBPs was within plusmn2 standard deviations in the Bland Altman plot. 93.4% of the difference between measured and estimated DBFs was within plusmn2 standard deviations in the Bland Altman plot. The preliminary results illustrated that the RS <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> is a promising parameter for continuous and non-invasive estimation of BP in homecare and mobile healthcare