Poincaré Plot Nonextensive Distribution Entropy: A New Method for Electroencephalography (EEG) Time Series

Area of Science:

• Neuroscience
• Data Analysis
• Signal Processing

Background:

• Poincaré plots are visual analysis tools for time series.
• Traditional methods struggle with complex correlation patterns in time series.
• Poincaré plot distribution entropy (DE) has limitations in analyzing fractional Brownian motion.

Purpose of the Study:

• To propose Poincaré plot nonextensive distribution entropy (NDE) for improved EEG time series analysis.
• To address the insufficient discrimination ability of DE for fractional Brownian motion with varying Hurst indices.
• To enhance the analysis of single-channel EEG data.

Main Methods:

• Analysis of the reasons for DE's failure in fractional Brownian motion analysis.
• Introduction of a nonextensive parameter based on the distance of sector ring subintervals from the origin.
• Application of Poincaré plot NDE to simulated fractional Brownian motion and sleep EEG datasets.

Main Results:

• NDE demonstrates superior discrimination ability for fractional Brownian motion time series with different Hurst indices compared to DE.
• The method effectively distinguishes between different sleep stages in a published sleep EEG dataset.
• Parameter determination for Poincaré plot NDE is explained.

Conclusions:

• Poincaré plot NDE is a novel and effective tool for single-channel EEG time series analysis.
• The proposed method offers a prospective approach for understanding complex time series data.
• NDE enhances the analytical capabilities of Poincaré plots for physiological signals.