Multiscale Diffusion Entropy Analysis for the Detection of Crucial Events in Cardiac Pathology

Area of Science:

• Biomedical Engineering
• Physiological Signal Processing
• Complexity Science

Background:

• Crucial events in physiological systems are key to understanding system dynamics but require further research for clinical application.
• Current methods for analyzing physiological signals, particularly ECG, need enhancement for accurate crucial event detection.

Purpose of the Study:

• To introduce modified diffusion entropy (MDEA) and multiscale diffusion entropy analyses (MSDEA) for improved crucial event detection in ECG.
• To analyze the temporal complexity of ECG signals from normal sinus rhythm (NSR), congestive heart failure (CHF), and cardiac arrhythmia (ARR) using MDEA and MSDEA.

Main Methods:

• Applied MDEA with stripes and MSDEA to 30 ECG samples each of NSR, CHF, and ARR datasets from PhysioNet.
• Measured temporal complexity using inverse power law (IPL) and scaling delta (δ) indices.
• Combined multiscale entropy (MSE) with MDEA in MSDEA for multi-time scale analysis.

Main Results:

• Healthy NSR ECGs exhibited approximately 15% higher IPL and scaling δ indices compared to pathological CHF and ARR signals.
• Pathological groups showed higher standard deviations in scaling indices, indicating greater variability.
• MSDEA demonstrated significantly clearer discrimination between healthy and pathological cardiac signals (p<0.0005).
• NSR complexity indices ranged twice as wide as pathological values across twenty temporal scales with reliable trend lines (R²≥0.95).

Conclusions:

• MDEA and MSDEA effectively reveal latent differences in ECG complexity related to crucial events.
• MSDEA offers enhanced diagnostic discrimination between healthy and pathological cardiac conditions.
• This approach holds clinical relevance for improving cardiac condition diagnosis through ECG signal processing.