Convolutional Neural Networks with 3D Input for P300 Identification in Auditory Brain-Computer Interfaces

<div>The different steps of input construction: (a) The experimental EEG channel layout. (b) EEG channel matrix disposition to form 2D and 3D inputs (upper and lower images, resp.). Gray cells contain no information. (c) Usual 2D input shape and proposed 3D input shape following our considerations (upper and lower images, resp.).</div>

Computational Intelligence and Neuroscience

fig1

Figure 1

Figure 1: Convolutional Neural Networks with 3D Input for P300 Identification in Auditory Brain-Computer Interfaces 