Theoretical analysis of xDAWN algorithm: Application to an efficient sensor selection in a p300 BCI

A Brain-Computer Interface (BCI) is a specific type of human-machine interface that enables communication between a subject/patient and a computer by a direct control from the decoding of brain activity. To improve the ergonomics and to minimize the cost of such a BCI, reducing the number of electrodes is mandatory. A theoretical analysis of the subjacent model induced by the BCI paradigm leads to derive a closed form theoretical expression of the spatial filters which maximize the signal to signal-plus-noise ratio. Moreover, this new formulation is useful to improve a previously introduced method to automatically select relevant sensors. Experimental results on 20 subjects show that the proposed method is efficient to select the most relevant sensors: from 32 down to 8 sensors, the loss in classification accuracy is less than 2%. Furethermore, the computational time required to rank the 32 sensors is reduced by a 4.6 speed up factor allowing dynamical monitoring of sensor relevance as a marker of the user's mental state.