A Physiological-Model-Based Neural Network Framework for Blood Pressure Estimation from Photoplethysmography Signals.

Area of Science:

• Biomedical Engineering
• Cardiovascular Physiology
• Artificial Intelligence in Medicine

Background:

• Continuous blood pressure (BP) monitoring is crucial for managing hypertension.
• Photoplethysmography (PPG) offers a non-invasive method for BP estimation but faces challenges in accuracy and physiological insight.
• Current methods often lack interpretability regarding underlying cardiovascular parameters.

Purpose of the Study:

• To develop and validate a novel physiological model-based neural network (PMB-NN) for continuous BP estimation from PPG signals.
• To integrate the estimation of total peripheral resistance (TPR) and arterial compliance (AC) within the PMB-NN framework.
• To enhance the physiological interpretability of BP estimation from PPG signals for cardiovascular applications.

Main Methods:

• Development of a PMB-NN framework utilizing PPG signals.
• Incorporation of physiological parameters like TPR and AC into the neural network model.
• Validation of the model using data from a single healthy participant under varying activity intensities.

Main Results:

• The PMB-NN framework demonstrated promising accuracy in BP estimation.
• Median standard deviations were 6.88 mmHg for systolic BP and 3.72 mmHg for diastolic BP.
• Estimated TPR and AC showed expected reductions with increased physical activity intensity.

Conclusions:

• The novel PMB-NN framework offers a physiologically interpretable approach to continuous BP estimation from PPG.
• The integration of TPR and AC enhances the understanding of cardiovascular dynamics during BP monitoring.
• Further research is warranted to validate the model's performance across diverse populations and conditions.