Entropy in the natural time domain.

Area of Science:

• Complex systems analysis
• Nonlinear dynamics
• Biophysics

Background:

• Traditional entropy measures like Shannon entropy are static and may not fully capture dynamic system behaviors.
• Understanding non-Markovianity is crucial for analyzing complex time series data in various scientific fields.
• Distinguishing between healthy and pathological physiological signals can improve diagnostic accuracy.

Purpose of the Study:

• To introduce and validate a surrogate data analysis based on natural time entropy fluctuations.
• To investigate the non-Markovian properties of seismic electric signals, artificial noises, and electrocardiograms (ECGs).
• To assess the potential of this method in differentiating healthy ECGs from those indicative of sudden cardiac death.

Main Methods:

• Calculation of entropy (S) fluctuations in the natural time domain using a sliding window approach.
• Application of surrogate data analysis by comparing entropy fluctuations in original data (deltaS) with shuffled (randomized) data (deltaSshuf).
• Analysis of the ratio deltaSshuf /deltaS to identify non-Markovian characteristics and differentiate signal types.

Main Results:

• The natural time entropy analysis successfully identified non-Markovianity in seismic electric signals, artificial noises, and ECGs.
• A significant difference was observed in the deltaSshuf /deltaS ratio between healthy and sudden cardiac death ECGs, enabling differentiation.
• The analysis provides insights into the physical meaning of deltaSshuf in the context of signal dynamics.

Conclusions:

• Natural time entropy analysis is a powerful tool for detecting non-Markovianity in diverse complex time series.
• This method demonstrates potential as a non-invasive diagnostic tool for identifying individuals at risk of sudden cardiac death.
• Further investigation into the physical interpretation of deltaSshuf can enhance understanding of complex system dynamics.