Threshold distribution of equal states for quantitative amplitude fluctuations

Area of Science:

• Biomedical Signal Processing
• Quantitative Physiology
• Sleep Medicine

Background:

• The distribution of equal states (DES) is a metric for quantifying amplitude fluctuations in biomedical signals.
• Traditional DES can fail when data resolution is high or processing techniques yield rare equal states, limiting its utility.
• Accurate quantification of amplitude fluctuations is crucial for analyzing physiological signals like EEG.

Purpose of the Study:

• To develop a novel threshold DES (tDES) method to overcome limitations of traditional DES in measuring amplitude fluctuations.
• To evaluate the tDES method using synthetic signals and real-world sleep electroencephalography (EEG) data.
• To assess the tDES method's utility in characterizing sleep stages based on EEG amplitude fluctuations.

Main Methods:

• Developed a threshold DES (tDES) algorithm to measure differential states within a specified threshold.
• Validated tDES on five sets of synthetic signals across different frequency bands.
• Applied tDES to sleep EEG datasets from the public PhysioNet database.

Main Results:

• tDES effectively quantified amplitude fluctuations in synthetic signals and detrend-filtered sleep EEGs, where traditional DES failed due to lack of equal states.
• A significant increase in tDES was observed with advancing sleep stages in EEG data, indicating reduced amplitude fluctuations.
• A general inverse relationship was found: more low-frequency components correlated with smaller amplitude fluctuations and larger DES.

Conclusions:

• The tDES method offers a robust and conceptually simple approach for quantifying amplitude fluctuations in biomedical signals.
• tDES expands the applicability of amplitude fluctuation analysis, particularly for high-resolution or processed physiological data.
• Findings support tDES as a valuable tool for sleep stage classification using EEG data.