Temporal Convolutional Learning: A New Sequence-based Structure to Promote the Performance of Convolutional Neural Networks in Recognizing P300 Signals

Distinguishing P300 signals from other components of the EEG is one of the mostchallenging issues in Brain Computer Interface (BCI) applications, and machine learningmethods have vastly been utilized as effective tools to perform such separation. Althoughin recent years deep neural networks have significantly improved the quality of the abovedetection, the significant similarity between P300 and other components of EEG in parallelwith their unrepeatable nature have led to P300 detection, which are still an open problemin BCI domain. In this study, a novel architecture is proposed in order to detect P300 signalamong EEG, in which the temporal learning concept is engaged as a new substructureinside the main Convolutional Neural Network (CNN). The above Temporal ConvolutionalNetwork (TCN) may better address the problem of P300 detection, thanks to its potentialin involving time sequence properties in modelling of these signals. The performance ofthe proposed method is evaluated on the EPFL BCI dataset, and the obtained results arecompared in two inter-subject and intra-subject scenarios with the results of classical CNNin which temporal properties of input are not considered. Increased True Positive Rate ofthe proposed method (an average of 4 percent) and its accuracy (an average of 2.9 percent)in parallel with the decrease in its False Positive Rate (averagely 3.1 percent) shows theeffectiveness of the TCN structure in promoting the detection procedure of P300 signals inBCI applications