In this work, a middle-aged group was compared with an elderly group in eyes-open and eyes-closed states, as well as during a visual WM task, to determine network matrices. Several network features were calculated from all subjects, and a statistical analysis was performed. The statistical analysis showed differences in a resting state and in working memory state networks. We further extended our described technique by combining the machine learning classification with three well-known classifiers (RF, KNN, and SVM). Our classification model achieved 98.89% accuracy with KNN during a WM task, thus corroborating the efficacy of our technique using EEG network features. In a resting state, the eyes-closed state of KNN achieved 93.33% accuracy, which was higher than in the eyes-open state.
