The graph indices show significant differences in the efficiency and clustering of networks in the resting state, as well as in the WM state. We further used these network features for the classification of middle-aged and elderly subjects. Three classifiers (KNN, RF, SVM) were used to evaluate the classification problem of middle-aged vs. elderly individuals. In an eyes-closed resting condition, to classify the young vs. middle-aged subjects, SVM achieved an accuracy greater than 82% using graph theory features. In our study, we obtained an accuracy of 87.80% in the eyes-open state and an accuracy of 93.33% in the eyes-closed state using KNN. Our study shows excellent values of sensitivity, specificity, Kappa statistics, and F-measure values of KNN (Table 1 and Table 2).
