Algorithm based on the short-term Rényi entropy and IF estimation for noisy EEG signals analysis

Area of Science:

• Neuroscience
• Signal Processing
• Biomedical Engineering

Background:

• Stochastic electroencephalogram (EEG) signals are inherently nonstationary and multicomponential.
• Movement-related cortical activities manifest as spectral EEG changes, crucial for understanding motor control disorders.
• Accurate detection of EEG signal components is vital for localizing neurological dysfunction.

Purpose of the Study:

• To introduce a novel algorithm for detecting and extracting components from EEG time-frequency distributions (TFDs).
• To enhance the diagnostic capabilities for neurological disorders, particularly those affecting motor control.
• To improve the spectral description of brain activities for better clinical interpretation.

Main Methods:

• Utilized a modified Rényi entropy-based technique for estimating the number of components, termed short-term Rényi entropy (STRE).
• Incorporated an iterative algorithm to enhance the performance of component detection.
• Combined STRE with instantaneous frequency (IF) estimation for detailed spectral analysis.

Main Results:

• The proposed algorithm efficiently detects EEG signal components in both noise-free and noisy conditions.
• Demonstrated effectiveness in limb movement EEG signal analysis, providing spectral descriptions at each electrode.
• The method proved robust up to moderate levels of additive noise.

Conclusions:

• The developed algorithm offers an efficient approach for EEG signal component detection and analysis.
• Extracted information on the number of components and their IFs shows significant potential for improving diagnostics and treatment of motor control disorders.
• This technique can enhance the localization of brain neurological dysfunctionalities in patients.