Which Multivariate Multi-Scale Entropy Algorithm Is More Suitable for Analyzing the EEG Characteristics of Mild...

Area of Science:

• Complexity Science
• Biomedical Signal Processing
• Neuroscience

Background:

• Multivariate multi-scale entropy (MMSE) algorithms are widely used to analyze signal complexity, but their measured characteristics and influencing factors are often unclear.
• Understanding these aspects is crucial for accurate signal analysis, particularly in clinical applications like mild cognitive impairment (MCI).

Purpose of the Study:

• To analyze six common MMSE algorithms, clarifying the signal characteristics they measure and the factors influencing them.
• To determine the most suitable MMSE algorithm for analyzing electroencephalograph (EEG) signals in patients with MCI.
• To investigate the application of the chosen algorithm in distinguishing MCI from control subjects.

Main Methods:

• Comparative analysis of six MMSE algorithms using simulations to assess their measurement of intra- and inter-channel correlation and signal complexity.
• Evaluation of algorithm performance based on signal complexity, coupling strength, noise resistance, and data length.
• Application of the selected algorithm (RCmvMFE) to analyze EEG data from healthy controls and individuals with MCI.

Main Results:

• MMSE algorithms like mvMSE, mvMFE, and RCmvMFE measure intra- and inter-channel correlation and signal complexity, decreasing with reduced complexity and coupling strength.
• The RCmvMFE algorithm demonstrated superior ability in distinguishing signal complexity and inter-channel correlations, alongside robust anti-noise and length analysis capabilities.
• Analysis of EEG data revealed lower entropy in the MCI group compared to controls on short scales, with the opposite trend on long scales. Frontal entropy positively correlated with cognitive assessment scores on short scales.

Conclusions:

• The RCmvMFE algorithm is highly effective for analyzing multi-channel signals, particularly EEG data in the context of MCI.
• MMSE algorithms provide valuable insights into brain complexity and connectivity, with potential for diagnosing and monitoring neurological conditions like MCI.
• Frontal EEG entropy, analyzed via RCmvMFE, shows promise as a biomarker for cognitive function in MCI.