EEG signal classification by improved MLPs with new target vectors

This paper proposes the use of new target vectors for MLP learning in EEG signal classification. A large Euclidean distance provided by orthogonal bipolar vectors as new target ones is explored to improve the learning and generalization abilities of MLPs. The data set consisted of EEG signals captured from normal individuals and individuals under brain-death protocol. Experimental results are related to MLP performance comparison by training the networks with three types of target vectors (conventional, orthogonal bipolar and non-orthogonal ones). We have concluded that the use of orthogonal bipolar vectors as target ones has contributed to improve the MLP performance on tasks for EEG signal classification.