P300 based auditory visual brain computer interface

Brain computer interface (BCI) is a developing research area which aims communication between the human user and the computers by way of user´s brainwaves only. This can be accomplished by processing electroencephalography (EEG) data. In this paper the aim is to build a simple BCI in terms of both collection and classification of EEG data using event related potentials (ERP). Proposed approach helps to establish a BCI with no need of training data nor trained subjects. With this purpose, the user is asked to make a drink selection using auditory, visual or both stimuli presented on the user interface. The brain´s response (P300) to highlighting of the targeted drink at the interface, is enhanced by applying wavelet decomposition to raw EEG data, and classified by means of the root mean square (RMS) value during a time interval appropriate for signal characteristics. The performance achieved using auditory and visual stimuli are compared and the use of an auditory BCI for visually impaired subjets is discussed.