Table 5 shows previous work based on graph theory features and the application of machine learning classification using EEG in different scenarios. Our classification result shows improved classification accuracy, which also corroborates the efficacy of the features. In our work, the KNN algorithm achieved the highest accuracy compared to the previous studies in which SVM was applied. To distinguish between younger and older adult brain networks, using functional connectivity in a resting state, SVM obtained an accuracy of 94% in classifying the brain by age group [75]. In the present work, the highest classification accuracy was 98.89% with KNN during a visual WM task state when classifying middle-aged and elderly EEG brain networks (Table 3). We also obtained a higher classification accuracy during the WM task than in the resting state. It is clear from the classification performance measures that the KNN classifier performed better compared to SVM and RF. The best sensitivity and specificity achieved by KNN confirm the validity of the classification model. Our classification results clearly exhibit increased accuracy in the WM task. To the best of
