Alzheimer's Disease Classification Algorithm Based on EEG Microstate

This study developed an analysis method based on Electroencephalography (EEG) microstate features to improve the classification accuracy of Alzheimer's disease patients (AD) and normal controls (NC). Initially, microstate features were extracted from EEG data, and a sliding time window technique was employed to calculate the proportion of microstates within each segment. The microstate feature with the highest proportion was defined as the first component, and its occurrence frequency was statistically weighted to obtain a feature vector for each participant. These features were then input into a support vector machine model for classification training. By leveraging the characteristics of microstates, the method effectively reduced noise interference and decreased computational complexity through segmentation. The use of a sliding time window increased the likelihood of extracting effective features from resting-state EEG data, thereby significantly improving the accuracy of classification. The results of the study indicate that this method has high accuracy and practicality in distinguishing between AD and NC.