Systolic blood pressure estimation using PPG and ECG during physical exercise

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Cardiovascular Health

Background:

• Accurate systolic blood pressure (SBP) monitoring is crucial for cardiovascular health management.
• Traditional SBP measurement methods can be invasive or inconvenient during physical activity.
• Non-invasive techniques using photoplethysmography (PPG) and electrocardiography (ECG) offer a promising alternative.

Purpose of the Study:

• To develop and evaluate a novel model for estimating SBP during exercise using PPG and ECG signals.
• To investigate the impact of incorporating pulse arrival time (PAT) and different feature extraction methods on SBP estimation accuracy.
• To assess the influence of PPG measurement site and body posture on the model's performance.

Main Methods:

• A model was developed using 18 features extracted from PPG waveforms, derivatives, and frequency spectra.
• Pulse arrival time (PAT) was calculated from combined PPG and ECG signals.
• Multiple linear regression (MLR) was employed to integrate features for SBP estimation.
• Leave-one-subject-out cross-validation was used to evaluate model performance.

Main Results:

• The inclusion of PAT improved SBP estimation, reducing the standard deviation (SD) of the difference from 14.07 to 13.52 mmHg.
• No significant performance difference was observed between finger- and forehead-derived PPG signals.
• Separate models are required for different body postures (lying, sitting, standing).
• The optimized model achieved a mean difference of 0.43 mmHg and an SD of 13.52 mmHg with a median correlation of 0.86.

Conclusions:

• The proposed model demonstrates promising results for non-invasive SBP estimation during physical exercise.
• Feature groups related to PPG beat morphology and the dicrotic notch significantly contribute to SBP estimation.
• This approach offers potential for continuous and convenient SBP monitoring in various settings.