Machine Learning and Electrocardiography Signal-Based Minimum Calculation Time Detection for Blood Pressure Detection

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Cardiovascular Monitoring

Background:

• Hypertension is a major risk factor for cardiovascular diseases and stroke.
• Accurate blood pressure monitoring is crucial for disease prevention and management.
• There is a need for noninvasive, AI-based systems for continuous blood pressure measurement.

Purpose of the Study:

• To determine the minimum Electrocardiography (ECG) signal duration required for accurate systolic and diastolic blood pressure calculation.
• To develop an artificial intelligence-based system for noninvasive blood pressure monitoring using ECG signals.

Main Methods:

• Utilized ECG recordings from five individuals from the IEEE database.
• Segmented ECG signals into epochs ranging from 2 to 20 seconds.
• Extracted 25 features from each epoch and reduced dimensionality using Spearman's algorithm.
• Applied Gaussian Process Regression (GPR) machine learning algorithm for analysis.

Main Results:

• The Mean Absolute Percentage Error (MAPE) for diastolic blood pressure was 2.44 mmHg using 16-second epochs.
• The MAPE for systolic blood pressure was 1.92 mmHg using 16-second epochs.
• High-performance estimation of blood pressure values was achieved with 16-second ECG signals.

Conclusions:

• Systolic and diastolic blood pressure can be accurately calculated using 16-second ECG signals.
• The findings support the development of efficient, noninvasive blood pressure monitoring systems.
• The GPR algorithm's suitability for embedded systems facilitates practical application.