Steady State Visual Evoked Potentials for a Brain-Computer Interface by Fuzzy Logic

Sensory stimulation of human body causes Evoked potentials in the human brain. These signals have been used for multiple applications over the past few decades. The purpose of this study is to classify the EEG signals extracted from a number of healthy volunteers using a fuzzy classifier and compare the result obtained with several other classifiers.Method In this study, EEG signals are recorded from the subjects. In the method of execution, the EEG signals are initially called and pre-processed. In the second step, appropriate characteristics of these signals are extracted. The extracted features are included in Tabel 1. Then the labels associated with these signals are attributed and eventually the tagged feature matrix is generated. This matrix is used as the input matrix of the classifier. The main classifier that is used in this study is fuzzy classification, while Bayes, SVM, PNN, MLP and KNN classifiers have also been used to compare the results.Results The accuracies of each classifier with different properties are presented in Table 1. In this work, classifier accuracy is defined as prediction capability of new input data for its corresponding class.Conclusions In this paper, it was shown that using fuzzy method for classification of EEG signals results in very good accuracy in comparison with other classifier methods such as KNN, MLP and PNN.