PCA-Based Multi-Wavelength Photoplethysmography Algorithm for Cuffless Blood Pressure Measurement on Elderly Subjects

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Cardiovascular Monitoring

Background:

• Hypertension prevalence necessitates convenient blood pressure (BP) monitoring via wearable devices.
• Cuffless BP monitoring commonly uses pulse transit time (PTT) derived from electrocardiography and photoplethysmography (PPG).
• Existing multi-wavelength PPG (MWPPG) methods for BP estimation face challenges in signal decoding and motion artifact (MA) removal.

Purpose of the Study:

• To develop an improved MWPPG algorithm for accurate cuffless BP estimation using a single sensing node.
• To enhance the robustness of arteriolar PTT calculation and effectively mitigate motion artifacts in wearable BP monitoring.

Main Methods:

• Proposed an enhanced MWPPG algorithm utilizing principal component analysis (PCA) for signal decomposition.
• Matched PCA statistical decomposition with arterial and capillary pulses to calculate arteriolar PTT from the entire waveform phase shift.
• Employed PCA-derived MA components to identify and exclude contaminated signal segments.
• Conducted a comparative experiment (N=22) using a cuffless MWPPG device against auscultatory BP measurement.

Main Results:

• The PCA-based MWPPG algorithm demonstrated improved accuracy in cuffless BP measurement.
• Achieved mean errors of 1.44 ± 6.89 mmHg for systolic BP and -1.00 ± 6.71 mmHg for diastolic BP.
• The method effectively improved the performance of MWPPG signals by robustly calculating PTT and filtering motion artifacts.

Conclusions:

• The proposed PCA-based algorithm significantly enhances the performance of MWPPG for cuffless BP measurement in wearable medical devices.
• This approach offers a more robust and accurate method for self-monitoring of blood pressure, addressing key limitations of current technologies.