Intelligent heart sound diagnosis based on Convolutional Neural Networks (CNN) has been attracting increasing attention due to its accuracy and efficiency, which have been improved by recent studies. However, the performance of CNN models, heavily influenced by their parameters and structures, still has room for improvement. In this paper, we propose a heart sound classification model named CAFusionNet, which fuses features from different layers with varying resolution ratios and receptive field sizes. Key features related to heart valve diseases are weighted by a channel attention block at each layer. To address the issue of limited dataset size, we apply a homogeneous transfer learning approach. CAFusionNet outperforms existing models on a dataset comprising public data combined with our proprietary dataset, achieving an accuracy of 0.9323. Compared to traditional deep learning methods, the transfer learning algorithm achieves an accuracy of 0.9665 in the triple classification task. Output data and visualized heat maps highlight the significance of feature fusion from different layers. The proposed methods significantly enhanced the performance of heart sound classification and demonstrated the importance of feature fusion, as interpreted through visualized heat maps.

Cuffless continuous blood pressure (BP) monitoring is essential for personal health management. However, its accuracy is challenged by the diversity and heterogeneity of physiological data sources. We propose a multi-source feature selection framework based on Markov blanket theory and the concept of causal invariance. We extracted 218 BP-related photoplethysmography (PPG) features from three heterogeneous datasets (differing in subject population, acquisition devices, and methods) and constructed a causal feature set using the Multi-Dataset Stable Feature Selection via Ensemble Markov Blanket (MDSFS-EMB) algorithm. BP estimation was then performed using four machine learning models. The MDSFS-EMB algorithm integrated PPFS and HITON-MB, enabling adaptability to different data scales and distribution scenarios. It employed Gaussian Copula Mutual Information, which was robust to outliers and capable of modeling nonlinear relationships. To validate the effectiveness of the selected feature set, we conducted experiments using an independent external validation dataset and explored the impact of data segmentation strategies on model prediction outcomes. The results demonstrated that the MDSFS-EMB algorithm has advantages in feature selection efficiency, prediction accuracy, and generalization capability. This study innovatively explores the causal relationships between PPG features and BP across multiple data sources, providing a clinically applicable approach for cuffless BP estimation.

The high incidence of cardiovascular disease and the early failure of bioprosthetic valves due to calcification have driven the development of anti-calcification technologies. As a new storage technology, drying treatment is expected to delay the calcification process by reducing glutaraldehyde residues. However, the effects of drying treatment on the mechanical properties and valve functions of bovine pericardial materials are still unclear. The objective of this study is to evaluate the influence of drying and rehydration treatments on the mechanical integrity and geometric properties of bovine pericardium and the hemodynamic performance of bioprosthetic valves made with these tissues. Cross-linked bovine pericardial samples (n = 15) were divided into three groups—wet (control group progressed with normal glutaraldehyde), dehydrated (ethanol–glycerol dehydration), and rehydration (saline immersion) groups—and the geometric stability and nonlinear mechanical behaviors of the materials were analyzed via thickness measurements and uniaxial and biaxial tensile tests. Quantitative results showed that thickness remained stable across groups (wet: 0.356 ± 0.052 mm; dry: 0.361 ± 0.053 mm; rehydrated: 0.361 ± 0.053 mm, p > 0.05). Elastic modulus values were preserved (wet: 12.5 ± 1.8 MPa; dry: 13.1 ± 2.0 MPa; rehydrated: 12.7 ± 1.9 MPa, p > 0.05), and anisotropy ratio showed no significant changes (1.53 ± 0.06 vs. 1.57 ± 0.07, p > 0.05). The hemodynamic performance of bioprosthetic valves made with these materials was evaluated in vitro using a pulsating flow simulation. Hemodynamic parameters demonstrated excellent preservation: effective orifice area (wet: 2.625 ± 0.11 cm²; rehydrated: 2.585 ± 0.12 cm², Δ = 1.5%, p = 0.32) and regurgitation fraction (wet: 39.35 ± 2.9%; rehydrated: 42.78 ± 3.2%, p = 0.15) showed no statistically significant differences. The geometric properties of the material were not significantly changed by the drying treatment, and the material maintained its nonlinear viscoelastic characteristics and anisotropy. The rehydrated bioprosthetic valves did not differ significantly from those in the wet group in terms of the effective orifice area, regurgitation fraction, and transvalvular pressure difference, and the hemodynamic performance remained stable.