Inspection of EEG Signals for Noninvasive Blood Glucose Monitoring in Prediabetes Diagnosis

Prediabetes is a metabolic disorder where the blood glucose (BG) level is higher than normal but not high as diabetes; early diagnosis can prevent health complications and death. However, to determine the BG level, it is required to prick the finger, which causes pain and discomfort. To eliminate this problem, there is a need to investigate noninvasive techniques to estimate BG values and continuous BG monitoring. In this paper, we investigated the changes in electroencephalogram (EEG) frequency parameters that have been scarcely considered for prediabetes diagnosis. We analyzed EEG signals after carrying out an oral glucose tolerance test on 25 participants. Five frequency bands of EEG signals were obtained continuously in 3 positions; frontal (F), occipital (O), and parental (P). The analysis is performed using a boxplot to examine the pattern of the recorded signals. The result shows that the EEG signal from the O position has a sensitivity of 95.3% in the left hemisphere, the P location has 90.3% in the right hemisphere, and F has 93% in the left hemisphere. This observation shows the appropriate location and the combination of EEG frequency parameters, such as the alpha and beta mean power from O and P, which can be integrated into a wearable device to provide a promising clinical solution for noninvasive blood glucose monitoring and prediabetes diagnosis.