Second derivative of the finger arterial pressure waveform: an insight into dynamics of the peripheral arterial pressure pulse

A finger on the pulse of cardiovascular health: estimating blood pressure with smartphone photoplethysmography-based pulse waveform analysis

Smartphone photoplethysmography (PPG) offers a cost-effective and accessible method for continuous blood pressure (BP) monitoring, but faces persistent challenges with accuracy and interpretability. This study addresses these limitations through a series of strategies. Data quality was enhanced to improve the performance of traditional statistical models, while SHapley Additive exPlanations (SHAP) analysis ensured transparency in machine learning models. Waveform features were analyzed to establish theoretical connections with BP measures, and feature engineering techniques were applied to enhance prediction accuracy and model interpretability. Bland-Altman analysis was conducted, and the results were compared against reference devices using multiple international standards to evaluate the method's feasibility. Data collected from 127 participants demonstrated strong correlations between smartphone-derived digital waveform features and those from reference BP devices. The mean absolute errors (MAE) for systolic BP (SBP), diastolic BP (DBP), and pulse pressure (PP) using multiple linear regression models were 7.75, 6.35, and 4.49 mmHg, respectively. Random forest models further improved these values to 7.34, 5.79, and 4.45 mmHg. Feature importance analysis identified key contributions from time-domain, frequency-domain, curvature-domain, and demographic features. However, Bland-Altman analysis revealed systematic biases, and the models barely meet established accuracy standards. These findings suggest that while smartphone PPG technology shows promise, significant advancements are required before it can replace traditional BP measurement devices.

A new methodology for determining the central pressure waveform from peripheral measurement using Fourier-based machine learning

Radial applanation tonometry is a well-established technique for hemodynamic monitoring and is becoming popular in affordable non-invasive wearable healthcare electronics. To assess the central aortic pressure using radial-based measurements, there is an essential need to develop mathematical approaches to estimate the central pressure waveform. In this study, we propose a new Fourier-based machine learning (F-ML) methodology to transfer non-invasive radial pressure measurements to the central pressure waveform. To test the method, collection of tonometry recordings of the radial and carotid pressure measurements are used from the Framingham Heart Study (2640 individuals, 55 % women) with mean (range) age of 66 (40-91) years. Method-derived estimates are significantly correlated with the measured ones for three major features of the pressure waveform (systolic blood pressure, r=0.97, p < 0.001; diastolic blood pressure, r=0.99, p < 0.001; and mean blood pressure, r=0.99, p < 0.001). In all cases, the Bland-Altman analysis shows negligible bias in the estimations and error is bounded to 5.4 mmHg. Findings also suggest that the F-ML approach reconstructs the shape of the central pressure waveform accurately with the average normalized root mean square error of 5.5 % in the testing population which is blinded to all stages of machine learning development. The results show that the F-ML transfer function outperforms the conventional generalized transfer function, particularly in terms of reconstructing the shape of the central pressure waveform morphology. The proposed F-ML transfer function can provide accurate estimates for the central pressure waveform, and ultimately expand the usage of non-invasive devices for central hemodynamic assessment.

Cardiovascular system modeling using windkessel segmentation model based on photoplethysmography measurements of fingers and toes

Background:

Photoplethysmography (PPG) contains information about the health condition of the heart and blood vessels. Cardiovascular system modeling using PPG signal measurements can represent, analyze, and predict the cardiovascular system.

Methods:

This study aims to make a cardiovascular system model using a Windkessel model by dividing the blood vessels into seven segments. This process involves the PPG signal of the fingertips and toes for further analysis to obtain the condition of the elasticity of the blood vessels as the main parameter. The method is to find the Resistance, Inductance, and Capacitance (RLC) value of each segment of the body through the equivalent equation between the electronic unit and the cardiovascular unit. The modeling made is focused on PPG parameters in the form of stiffness index, the time delay (△t), and augmentation index.

Results:

The results of the model simulation using PSpice were then compared with the results of measuring the PPG signal to analyze changes in the behavior of the PPG signal taken from ten healthy people with an average age of 46 years, compared to ten cardiac patients with an average age of 48 years. It is found that decreasing 20% of capacitance value and the arterial stiffness parameter will close to cardiac patients' data. Compared with the measurement results, the correlation of the PPG signal in the simulation model is more than 0.9.

Conclusions:

The proposed model is expected to be used in the early detection of arterial stiffness. It can also be used to study the dynamics of the cardiovascular system, including changes in blood flow velocity and blood pressure.

Local Crack-Programmed Gold Nanowire Electronic Skin Tattoos for In-Plane Multisensor Integration

Sensitive, specific, yet multifunctional tattoo-like electronics are ideal wearable systems for "any time, any where" health monitoring because they can virtually become parts of the human skin, offering a burdenless "unfeelable" wearing experience. A skin-like, multifunctional electronic tattoo made entirely from gold using a standing enokitake-mushroom-like vertically aligned nanowire membrane in conjunction with a programmable local cracking technology is reported. Unlike previous multifunctional systems, only a single material type is needed for the integrated gold circuits involved in interconnects and multiplexed specific sensors, thereby avoiding the use of complex multimaterials interfaces. This is possiblebecause the programmable local cracking technology allows for the arbitrary fine-tuning of the properties of elastic gold conductors from strain-insensitive to highly strain-sensitive simply by adjusting localized crack size, shape, and orientations-a capability impossible to achieve with previous bulk cracking technology. Furthermore, in-plane integration of strain/pressure sensors, anisotropic orientation-specific sensors, strain-insensitive stretchable interconnects, temperature sensors, glucose sensors, and lactate sensors without the need of soldering or gluing are demonstrated. This strategy opens a new general route for the design of next-generation wearable electronic tattoos.

Continuous non-invasive determination of nocturnal blood pressure variation using photoplethysmographic pulse wave signals: comparison of pulse propagation time, pulse transit time and RR-interval

OBJECTIVE

Cardiovascular diseases are the leading cause of death, whereas nocturnal ambulatory blood pressure (BP) is the most potent predictor for cardiovascular risk. The volume clamp and pulse transit time (PTT) are common methods for continuous non-invasive BP measurement, but have drawbacks during unsupervised ambulatory use and undisturbed sleep. The pulse propagation time (PPT), defined as the time between pulse wave systolic peak and diastolic peak, provides valid information about the pressure pulse waveform. However, the use of PPT for nocturnal BP variation determination and whether such variation is affected by BP or heart rate (i.e. RR-interval or RRI) has not been investigated.

APPROACH

To assess whether the PPT method is suitable for ubiquitous nocturnal BP monitoring, we compared systolic blood pressure (SBP) estimates derived from PPT, PTT, and RRI signals with parallel recorded BP measurements. The RRI-derived SBP signals were used as a baseline for testing a potential heart rate dependency. This work provides an overview of BP measurements, presents the developed real-time signal analysis, and describes the performance assessment. The signal analysis was validated with data records from 42 subjects acquired from an ergometry and sleep laboratory in equal parts.

MAIN RESULTS

The algorithms applied to the ergometry laboratory database achieved a correlation coefficient between reference SBP and estimated SBP_PPT of 0.89 (p  <  0.001) with bias 0.1 mmHg and limits of agreement (LoA)  -29.8 to 30.0 mmHg, SBP_PTT of 0.97 (p  <  0.001) with bias 0.0 mmHg and LoA  -15.2 to 15.3 mmHg, and SBP_RRI of 0.96 (p  <  0.001) with bias 0.0 mmHg and LoA  -19.5 to 19.5 mmHg. For the sleep laboratory database, the correlation coefficient was 0.95 (p  <  0.001) with bias 0.2 mmHg and LoA  -18.3 to 18.8 mmHg for SBP_PPT, 0.88 (p  <  0.001) with bias 0.0 mmHg and LoA  -25.0 to 24.9 mmHg for SBP_PTT, and 0.88 (p  <  0.001) with bias of 0.1 mmHg and LoA  -23.6 to 23.7 mmHg for SBP_RRI. A heart rate dependency of PPT or PTT could not be found. The analysis of variance shows no significant differences between the reference SBP values and the estimated values for either the ergometry (F(3, 627)  =  2.27, p  =  0.08) or the sleep laboratory (F(3, 327)  =  2.28, p  =  0.08).

SIGNIFICANCE

In conclusion, the PPT method seems to be an interesting alternative for continuous determination of SBP during simplified cardiovascular monitoring and sleep screening compared to more expensive devices based on volume clamp or PTT methods.

Evaluation of Cardiorespiratory Function During Cardiopulmonary Exercise Testing in Untreated Hypertensive and Healthy Subjects

Objective: This study aimed to compare differences in cardiorespiratory function between untreated hypertensive subjects (UHS) and healthy subjects (HS) during cardiopulmonary exercise testing (CPET). Additionally, it also aimed to explore the potential mechanisms of different exercise responses in cardiorespiratory function before, during and after CPET. Methods: Thirty subjects (15 UHS and 15 HS) were enrolled. Photoplethysmography (PPG), respiratory signal, and ECG were simultaneously collected while subjects were performing CPET. Fiducial points (a, b, c, d, e) were extracted from the second derivative of the PPG (SDPPG), and the ratios b/a, c/a, d/a, e/a, and (b-c-d-e)/a (named Aging Index, AGI) were calculated as markers of systolic and diastolic function. Additionally, respiratory rate was calculated and analyzed. Results:Before CPET, there were no significant differences in b/a, d/a, and AGI between two groups. However, after CPET, b/a (-0.9 ± 0.19 vs. -1.06 ± 0.19, p-value = 0.03) and AGI (-0.49 ± 0.75 vs. -1.15 ± 0.59, p-value = 0.011) of the UHS group were significantly higher than those of the HS. The d/a (-0.32 ± 0.24 vs. -0.14 ± 0.17, p-value = 0.024), and c/a (-0.33 ± 0.26 vs. -0.07 ± 0.19, p-value = 0.004) were significantly lower in UHS than those in HS. In contrast, before CPET, e/a (0.22 ± 0.11 vs. 0.32 ± 0.09, p-value = 0.007) in UHS was significantly lower than that in HS, while after CPET there was no significant difference between the two groups in this variable. In addition, during CPET, AGI (p-value = 0.003), and respiratory rate (p-value = 0.000) in UHS were significantly higher in comparison with before CPET. Conclusions: Different exercise responses showed the differences of cardiorespiratory function between UHS and HS. These differences not only can highlight the CV risk of UHS, but also can predict the appearance of arterial stiffness in UHS. Additionally, during CPET, significant differences in AGI, autonomic nervous function and respiratory activity assessed by respiratory rate were found between the two groups in comparison with before CPET.

Hypersampling of pseudo-periodic signals by analytic phase projection

A method to upsample insufficiently sampled experimental time series of pseudo-periodic signals is proposed. The result is an estimate of the pseudo-periodic cycle underlying the signal. This "hypersampling" requires a sufficiently sampled reference signal that defines the pseudo-periodic dynamics. The time series and reference signal are combined by projecting the time series values to the analytic phase of the reference signal. The resulting estimate of the pseudo-periodic cycle has a considerably higher effective sampling rate than the time series. The procedure is applied to time series of MRI images of the human brain. As a result, the effective sampling rate could be increased by three orders of magnitude. This allows for capturing the waveforms of the very fast cerebral pulse waves traversing the brain. Hypersampling is numerically compared to the more commonly used retrospective gating. An outlook regarding EEG and optical recordings of brain activity as the reference signal is provided.