HemoSC-P: A Hemodynamic Semantic Channel Paradigm for Cardiovascular Parameter Estimation.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiovascular Physiology

Background:

• Cardiovascular disease is a leading global cause of death, necessitating improved monitoring and early warning systems.
• Current systems suffer from a
• black box dilemma
• leading to unreliable results and limited robustness due to physiological heterogeneity.

Purpose of the Study:

• To propose a novel hemodynamic channel-guided cardiovascular parameter estimation paradigm (HemoSC-P).
• To address the limitations of interpretability and robustness in existing cardiovascular monitoring systems.
• To leverage semantic communication principles for enhanced cardiovascular signal analysis.

Main Methods:

• Developed a dual-pillar modeling framework: a semantic pillar using multi-scale convolutional and phase-aware attention, and a cardiovascular channel guided by the Windkessel model.
• Modeled non-stationary cardiovascular signal dynamics and aligned features to the physiological domain.
• Employed physiologically deformable attention at the channel level for inverse parameter estimation.

Main Results:

• Validated the HemoSC-P paradigm using non-invasive blood pressure estimation on three public datasets (UCI-BP, MIMIC-III, PPG-BP).
• Achieved high accuracy on MIMIC-III with mean absolute errors of 3.04 ± 3.24 mmHg for systolic and 2.57 ± 2.70 mmHg for diastolic blood pressure.
• Demonstrated superior accuracy, stability, and scalability compared to benchmark methods across multiple datasets.

Conclusions:

• The HemoSC-P paradigm offers a significant advancement in cardiovascular parameter estimation, enhancing accuracy and interpretability.
• This approach effectively overcomes the limitations of current cardiovascular early warning systems.
• The findings suggest broad applicability and scalability for improving cardiovascular disease management.