Pulse wave velocity (PWV) measured at a specific artery location is called local PWV, which provides the elastic characteristics of arteries and indicates the degree of arterial stiffness. However, the large and cumbersome ultrasound probes require an appropriate sensor position and pressure maintenance, introducing usability constraints. In this paper, we developed a light (0.5 g) and thin (400 $\mathsf{\mu }$m) flexible ultrasound array by encapsulating 1–3 composite piezoelectric transducers with a silicone elastomer. It can capture the distension waveforms of four arterial positions with a spacing of 10 mm and calculate the local PWV by multi-point fitting. This is illustrated by in vivo experiments, where the local PWV value of five normal subjects ranged from 3.07 to 4.82 m/s, in agreement with earlier studies. The beat-to-beat coefficient of variation (CV) is 12.0% ± 3.5%, showing high reliability. High reproducibility is shown by the results of two groups of independent measurements of three subjects (the error between the mean values is less than 0.3 m/s). These properties of the developed flexible ultrasound array enable the bandage-like application of local PWV monitoring to skin surfaces.

Multi-label electrocardiogram (ECG) classification is of great significance for the diagnosis of cardiovascular diseases, and it is also a challenging task because of the various of pathological patterns across different cardiac abnormalities. This paper analyzes the complex spatial-temporal variations information contained in the 12-lead ECG from a new perspective, and proposes an advanced deep neural network framework to accurately identify various multi-label ECG abnormalities. Firstly, the 12-lead ECG was modeled as 3-dimensional (3-D) spatial-temporal variations involving the intra-period, inter-period temporal dimensions, and cross-lead spatial dimension. Then, a 3-D spatial-temporal learning network (3D-STLNet) was carefully designed to extract the complex pathological features from the 3-D spatial-temporal variations. Secondly, the multi-head attention (MHA) mechanism was introduced into the proposed network to extract global temporal dependencies in ECG segments, which further improved the performance of the proposed network. By integrating a residual backbone network, a 3D-STLNet, and the MHA module, our proposed network allows end-to-end learning and prediction in a unified framework. The influence of several classic threshold selection strategies on the performance of multi-label ECG classification algorithm was also analyzed, which provides a reference for subsequent research on optimal threshold selection for multi-label ECG classification. Finally, our proposed network was evaluated on multiple large multi-label ECG public datasets, and the performance of our network outperforms that of current state-of-the-art networks, which demonstrates that our network has great competitiveness in identifying a wider range of multi-label ECG abnormalities.

Accurate and portable Blood Pressure (BP) monitoring is vital for managing cardiovascular diseases. However, existing wearable continuous BP monitoring technologies are often inaccurate and rely on external calibration, limiting their practical application in continuous BP monitoring. To address this challenge, we have developed a Wearable continuous non-invasive BP Monitor (WeBPM) equipped with a finger cuff sensor, capable of monitoring BP continuously and accurately within medical-grade precision. WeBPM integrates advanced finger oscillographic BP measurement technology to provide reliable self-calibration functionality. Moreover, Pulsatile Cycle Volume Adjustment Method (PCVAM) we proposed for the closed-loop phase can continuously track changes in vasomotor tone under a controlled frequency based on pulsatile cycles, thereby enabling continuous BP measurement. In comparative experiments with the Nexfin monitor, WeBPM demonstrates excellent performance in induced dynamic BP experiments, with measurement errors of (-1.4 ± 6.24) mmHg for Systolic BP (SBP) and (-0.82 ± 4.83) mmHg for Diastolic BP (DBP). Additionally, compared to clinical invasive reference measurements, WeBPM's SBP and DBP measurement errors are (-1.74 ± 4.9) mmHg and (0.37 ± 3.28) mmHg, respectively, further proving its outstanding performance. These results highlight WeBPM's potential in personalized health management and remote monitoring, offering a new solution for continuous non-invasive BP monitoring.

Blood pressure estimation through pulse wave analysis (PWA) aims to establish the relationship between features of pulse waveforms and blood pressure. This study is the first to investigate the connection between features of carotid artery diameter waveforms and variations in blood pressure, as well as to develop a blood pressure estimation model based on these features. A dataset was constructed from 14 subjects, with data collected across various physiological states and time points. For each subject, carotid artery diameter waveforms were measured using ultrasound, while synchronous blood pressure data were recorded with a reference device. A total of 52 morphological features were extracted from the diameter waveforms and their first and second derivatives. The influence of different models and feature combinations on blood pressure estimation was analyzed using various machine learning approaches. Ultimately, optimal models were developed for each subject to dynamic blood pressure fluctuations. On independent test data where blood pressure fluctuations exceeded 25 mmHg, the mean absolute error (MAE) of the estimates was 3.3 ± 4.1 mmHg. Even after a period of two days or more, the models remained effective, yielding a MAE of 4.2 ± 5.3 mmHg.