Bivariate global frequency analysis versus chaos theory. A comparison for sleep EEG data

Area of Science:

• Neuroscience
• Signal Processing

Background:

• Electroencephalogram (EEG) data analysis is crucial for understanding sleep physiology.
• Quantitative descriptors are increasingly used to analyze complex EEG signals.
• Hjorth's parameters and fractal dimension are established methods for EEG complexity assessment.

Purpose of the Study:

• To compare the efficacy of various quantitative descriptors for classifying sleep stages using EEG data.
• To evaluate Hjorth's mobility and complexity measures against dimensionality analysis techniques.
• To assess the discriminative power of correlation exponent and bivariate global frequency analysis for sleep staging.

Main Methods:

• Application of Hjorth's mobility and complexity measures to EEG data.
• Utilizing dimensionality analysis techniques for comparison.
• Calculating the correlation exponent to describe EEG complexity.
• Performing bivariate global frequency analysis on EEG signals.

Main Results:

• Hjorth's parameters were used to classify sleep stages.
• The correlation exponent showed significant distinction between sleep stages, except for REM sleep.
• Bivariate global frequency analysis demonstrated superior discriminative power compared to the correlation exponent.
• A high statistical correlation was found between fractal dimension estimator and Hjorth's mobility.

Conclusions:

• Bivariate global frequency analysis is a powerful tool for sleep stage classification from EEG.
• Hjorth's parameters offer a viable method for sleep stage analysis, with notable correlation to fractal dimension.
• Further research can refine these quantitative descriptors for enhanced sleep EEG analysis.