Identification of chronic heart failure using linear and nonlinear analysis of heart sound

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Chronic heart failure (CHF) is an advanced stage of various cardiac diseases.
• Heart sound (HS) signals contain valuable information about cardiac mechanical activity.
• Distinguishing CHF patients from healthy individuals requires reliable diagnostic markers.

Purpose of the Study:

• To investigate the utility of linear and nonlinear analysis of heart sound (HS) signals for diagnosing chronic heart failure (CHF).
• To identify characteristic features in HS signals that differentiate CHF patients from healthy subjects.
• To explore the potential of HS signal analysis as a supplementary diagnostic tool for CHF.

Main Methods:

• Analysis of heart sound (HS) signals from healthy subjects and CHF patients.
• Application of linear approaches: time and frequency domain analysis.
• Application of nonlinear approaches: largest Lyapunov exponent, correlation dimension, sample entropy, and multifractal spectrum width.
• Statistical testing and Receiver Operating Characteristic (ROC) curve analysis for feature validation.

Main Results:

• Statistically significant differences were observed in both linear and nonlinear HS signal features between healthy and CHF groups.
• CHF patients exhibited decreased chaotic characteristics, complexity, and randomness in their cardiac mechanical activity compared to healthy individuals.
• HS signal features demonstrated potential in distinguishing CHF patients from healthy subjects.

Conclusions:

• Linear and nonlinear analysis of heart sound signals can effectively characterize cardiac mechanical activity.
• The identified HS signal features show promise as supplementary indexes for the diagnosis of chronic heart failure (CHF).
• Heart sound analysis offers an accessible and potentially efficient method for CHF detection.