Neural Prediction of the Target "to BUY" or "NOT to BUY" by the ERP- based Cognitive BMI.

抄録

We have been developing an EEG-based cognitive brain-machine interface (BMI) system, Neurocommunicator® to support patients with severe motor deficits in communication. This system uses an event-related potential (ERP) to decode the message (pictogram) that users want to convey. In addition to the communication aid, it would be expected for the Neurocommunicator to improve the QOL of the users in a variety of daily-life situations. In this study, we focused on the brain activity that reflected the decision-making process of customers. We used the virtual shopping task as a model of daily shopping, in which a variety of product images were used as visual stimuli for the sequential delayed matching-to-sample paradigm. We recorded EEG data from 12 normal subjects and examined the ERPs in two conditions; the subject selected one (target) out of 8 products (nontargets) either “to BUY” (positive condition) or “NOT to BUY” (negative condition). We observed the strong ERPs not only to the target “to BUY” but also to the target “NOT to BUY” compared to the nontarget. The magnitude of the response and the decoding accuracy of the former were, however, greater than the latter. These results suggest that the ERP to a product is enhanced by the intention to buy it, and that the ERP- based Neurocommunicator has potential to become a “body-free” shopping tool for patients with severe motor deficits as well as a real-time neuromarketing/neuroconsulting tool for any person.