Classifying ECoG Signals Using Probabilistic Neural Network

Electrocorticograms (ECoG) signals have many potential advantages and gained much attention for use with brain-computer interface (BCI). In this study, feature extraction using band powers was applied to ECoG signals from one subject performing imagined movements of either the left small-finger or the tongue. Probabilistic neural network (PNN) which was very suitable for classification problems was used to classify the two different imaginary movements. The classification accuracy rate for the test data set reached a maximum of 86% when spread of radial basis functions was 0.38. The results of this experiment showed that ECoG signals could be used and proved to be very powerful in BCI system design.