Wrist pulse signal based vascular age calculation using mixed Gaussian model and support vector regression

Area of Science:

• Biomedical Engineering
• Cardiovascular Health
• Signal Processing

Background:

• Vascular age (VA) is a critical indicator of vascular aging and public health.
• Current methods for assessing VA are often inconvenient and costly.
• Developing accessible and affordable VA assessment tools is a significant research focus.

Purpose of the Study:

• To propose and validate a novel method for evaluating vascular age (VA) using wrist pulse signals.
• To establish a convenient and cost-effective approach for assessing vascular aging.

Main Methods:

• Pulse signals were analyzed using a mixed Gaussian model (MGM) to extract shape features.
• Principal Component Analysis (PCA) was employed for feature dimension optimization.
• A Support Vector Regression (SVR) model was developed using principal components and chronological age (CA) to predict VA.

Main Results:

• The developed VA prediction model demonstrated strong correlations with various vascular indices, including pulse width (PW), inflection point area ratio (IPA), Ratio b/a (RBA), augmentation index (AIx), diastolic augmentation index (DAI), and pulse transit time (PTT).
• Vascular age (VA) showed a closer correlation with these indices compared to chronological age (CA), indicating its superior representativeness of vascular aging.
• Significant correlations were observed across both male and female participants, highlighting the method's generalizability.

Conclusions:

• The proposed method effectively estimates vascular age (VA) from wrist pulse signals.
• Vascular age (VA) is a more accurate reflection of vascular aging than chronological age (CA).
• This approach offers a direct, objective, and potentially widespread tool for public health assessments of vascular aging.