[Complexity Analysis of Physiological Signals Using Encoding Lempel-Ziv Algorithm].

Area of Science:

• Physiological signal analysis
• Complexity science
• Biomedical engineering

Background:

• Distinguishing randomness from chaos in physiological signals is crucial for complexity analysis.
• Existing complexity algorithms often struggle with performance consistency across different signal lengths and parameters.
• Accurate complexity quantification is vital for understanding physiological system dynamics.

Purpose of the Study:

• To develop and evaluate an encoding Lempel-Ziv (ELZ) complexity algorithm for discerning randomness and chaos in signals.
• To assess the performance of ELZ compared to classic LZ (CLZ), multi-state LZ (MLZ), sample entropy (SampEn), and permutation entropy (PE).
• To determine if ELZ is independent of signal length and sensitive to dynamical property changes.

Main Methods:

• Proposed an encoding Lempel-Ziv (ELZ) complexity algorithm.
• Compared ELZ with CLZ, MLZ, SampEn, and PE using simulated data (Gaussian and 1/f pink noise) and real physiological data.
• Evaluated performance across various time series lengths (100, 500, 5000) and sensitivity to dynamical changes.

Main Results:

• ELZ successfully distinguished randomness and chaos across all tested time series lengths.
• ELZ accurately reflected that Gaussian noise has lower complexity than 1/f pink noise.
• ELZ demonstrated that congestive heart failure (CHF) exhibits lower complexity than normal sinus rhythm (NSR) in RR interval data (P<0.01).

Conclusions:

• The proposed ELZ algorithm effectively discerns randomness and chaos characteristics of time series.
• ELZ exhibits superior generalization ability and independence from time series length compared to other methods.
• ELZ provides a robust tool for analyzing physiological signal complexity, applicable to conditions like CHF.