Extraction of qEEG Variables to Diagnose Early Dementia

Although it is difficult to definitely diagnose dementia with noninvasive methods such as EEG and MEG, early detection of subjects with probable dementia is of importance to apply currently available treatment strategies. In the present study, we have extracted featured qEEG variables, which can be potentially used to diagnose early dementia, from resting eyes-closed continuous EEGs of 22 early dementia patients and 27 age-matched healthy controls. Among 1680 candidate qEEG variables, about 160 variables showed statistically significant difference (p < 0.01) between early dementia group and control subject group, when mean phase coherence as well as EEG coherence was evaluated for all possible combinations of electrode pairs. Some preliminary trials to discriminate the two groups with the extracted qEEG variables demonstrated that the use of mean phase coherence as a supplementary or alternative measure for the EEG coherence may enhance the accuracy of diagnosis of early dementia.