Multimedia remote interactive operations based on EEG signals constructed BCI with convolutional neural network

In order to construct the multimedia remote interactive operations for the locked-in patients, this paper proposed a novel BCI system that constructed by the EEG signals with convolutional neural network. To constructing the remote interactive operations, the temporal and spatial features of electroencephalogram (EEG) signals were extracted by longitudinal convolution kernel and transverse convolution kernel respectively. The classification of motion imagination was completed by using two fully-connected layers. Finally, the classification results will be transferred as the multimedia remote interactive results in response to the multimedia remote interactive operations. The experimental results showed that the proposed EEG-BCI system had a good performance and efficiency for the multimedia remote interactive operations, which will facilitate the fields of multimedia remote interactive operations.