Testing dynamic correlations and nonlinearity in bivariate time series through information measures and surrogate...

Area of Science:

• Dynamical systems analysis
• Nonlinear time series analysis
• Physiological signal processing

Background:

• Time series data analysis is crucial for understanding system dynamics.
• Surrogate data testing provides robust statistical validation for observed phenomena.
• Distinguishing deterministic from stochastic systems requires advanced analytical tools.

Purpose of the Study:

• To develop and validate novel surrogate methods for assessing statistical properties in random processes.
• To evaluate the presence of autodependencies, nonlinear dynamics, coupling, and nonlinearities in univariate and bivariate processes.
• To apply these methods to physiological time series for insights into cardiorespiratory interactions.

Main Methods:

• Formulating null hypotheses and testing them with surrogate data.
• Utilizing Information Storage as a discriminating statistic for univariate processes.
• Employing Mutual Information Rate for detecting coupling and nonlinearities in bivariate processes.
• Testing methods via simulations and application to human physiological data (RR intervals and RESP).

Main Results:

• Simulations confirmed the effectiveness of proposed methods in identifying dynamical features of stochastic systems.
• Paced breathing at low rates increased the predictability of individual RR interval and RESP dynamics.
• Paced breathing at low rates reduced the nonlinearity in the coupled dynamics of RR and RESP.

Conclusions:

• The developed surrogate methods are effective for analyzing complex time series data.
• Paced breathing significantly alters the dynamics and coupling of cardiorespiratory signals.
• These findings have implications for understanding physiological regulation and developing advanced analytical techniques.