Evaluation and Classification of Dementia Using EEG Indicators During Brain-Computer Interface Tasks

The rapid increase in the number of patients with dementia is currently a concern. According to a survey, the number of patients with dementia in Japan would exceed 10 million by 2060. Thus, there is a need to develop simple techniques for early diagnosis of dementia to suppress the increase in patients with dementia. In our laboratory, we are developing a dementia-screening tool using character input-type Brain-Computer Interface. In this study the electroencephalogram (EEG) data obtained using the tool were analyzed in the frequency band. The purpose is to find the difference in EEG between healthy people, patients with mild cognitive impairment (MCI), and patients with Alzheimer's disease (AD). The results show that the mean value of the ratio of beta to alpha wave (beta/alpha) significantly differs between healthy subjects and MCI patients. The mean value of beta/alpha was lower in the MCI patients than in the healthy subjects. In addition, there was also a significant difference in the range of beta/alpha between beta/alpha for patients with MCI and that for the patients with AD; that of AD patients was higher. From the results, it is considered that the degree of concentration decreases, and its variation becomes remarkable as the cognitive function declines. With these indicators, the three states are expected to be classified. In future studies, we shall verify whether the classification accuracy can be improved by using these indicators in machine learning.