Visualization of Attention Concentration Area of P300-Speller BCI

Currently, the number of patients with dementia is increasing worldwide. The incidence of mild cognitive impairment (MCI), a pre-dementia stage, is also on the rise. It has been reported that 50% of patients with MCI will transition to dementia within 5 years. Therefore, early detection is important to prevent the progression of dementia. For the purpose, we are developing an inexpensive and simple cognitive screening method using the P300-Speller, a spelling-BCI based on event-related potential P300. It is a mechanism whereby P300 is elicited when the subject gazes at a certain letter in the matrix. Subsequently, the character is estimated using a discrimination scores (DSs) calculated based on the P300 component. As cognitive function declines, the subject’s ability to pay attention concentration is impaired, thereby increasing the chance of incorrect input. By focusing on these DSs, we think that this method can be used to confirm differences and screen patients. In this study, we used the sigmoid function to represent the probability of these DSs. We created a probability distribution map. As a result, we confirmed that with each progression of dementia, the scatter of the data increased with respect to the character that was the focus of attention. In addition, statistical processing was applied to the variance values combined with SEDV. The results of Bonferroni’s multiple comparisons confirmed a significant difference of 1% between all groups (i.e., normal control, MCI, and Alzheimer’s disease). The differences in probability distribution maps suggest that this method is useful for screening for dementia.