Resting-state EEG microstate features for Alzheimer's disease classification.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Resting-state electroencephalogram (EEG) microstate analysis identifies distinct network activations.
• Alzheimer's disease (AD) is associated with global microstate disorganization in EEG patterns.
• Microstate parameters may serve as potential biomarkers for neurological conditions like AD.

Purpose of the Study:

• To evaluate the classification efficacy of EEG microstate parameters for identifying Alzheimer's disease (AD).
• To compare the diagnostic performance of microstate features against conventional EEG features in AD detection.

Main Methods:

• Extracted raw EEG data from the OpenNeuro EEG database.
• Preprocessed EEG data and filtered into five frequency bands.
• Utilized Support Vector Machine (SVM) for initial microstate feature selection and classification, followed by comparison with KNN, RF, and LR classifiers using both microstate and conventional EEG features.

Main Results:

• The microstate feature set achieved an optimal classification accuracy of 99.22% for AD recognition in the Alpha (8-13 Hz) band using SVM.
• Across four classifiers, the microstate feature set yielded an average accuracy of 98.61% in the Alpha band.
• Conventional EEG features achieved an average accuracy of 91.19% in the Alpha band across the same classifiers.

Conclusions:

• Microstate parameters are effective biomarkers for classifying EEG data in Alzheimer's disease patients.
• Microstate EEG features demonstrate superior performance over conventional EEG features for AD classification, independent of the machine learning classifier used.