A novel Blood Pressure estimation method combing Pulse Wave Transit Time model and neural network model

Area of Science:

• Biomedical Engineering
• Cardiovascular Monitoring
• Artificial Intelligence in Medicine

Background:

• Accurate blood pressure (BP) measurement is crucial for assessing cardiovascular health and diagnosing heart conditions.
• The Pulse Wave Transit Time (PWTT) model is widely used for continuous BP monitoring in clinical settings.
• Individual variations can significantly impact the accuracy of traditional PWTT models.

Purpose of the Study:

• To develop a novel BP estimation method that overcomes the limitations of individual variations in PWTT models.
• To enhance the accuracy and reliability of continuous BP monitoring through a hybrid approach.
• To provide a clinically viable solution for precise diastolic and systolic blood pressure estimation.

Main Methods:

• A five-step methodology integrating a classical PWTT model with a neural network model.
• Includes signal pre-processing, feature extraction, initial PWTT model selection, neural network-based model correction, and final model identification.
• Utilized the Multiparameter Intelligent Monitoring in Intensive Care (MIMIC) database for validation.

Main Results:

• The proposed method demonstrated a minimum mean error of 5 mmHg and a standard deviation of ±8 mmHg in BP estimation.
• Validation experiments on 10 patients confirmed the method's effectiveness.
• Achieved results meeting clinical requirements for both diastolic and systolic blood pressure estimation.

Conclusions:

• The combined PWTT and neural network model offers a significant improvement in BP estimation accuracy.
• This novel method addresses the challenge of individual variations in PWTT measurements.
• The findings suggest this approach is suitable for reliable, continuous blood pressure monitoring in clinical practice.