Classification of mental states of human concentration based on EEG signal

This paper provides a suitable method for classifying the EEG signal. In this article, a number of features are extracted from the EEG signal and by using these different features and networks, these signals are classified into three categories: relaxation, moderate concentration and high concentration. In this case, based on the amount of mental activity that has a direct effect on the EEG signal, the state of attention can be categorized. In this paper, four sensors (electrodes) are used to collect the voltage of the brain signals, then the Large Laplacian Filter is used to localize the signals, and by this method, the signals of the four sensors are converted into one signal, then the frequency of 50 Hz (City frequency) is removed using a Notch passive filter and then a wavelet filter is used to remove noise and artifacts. In this article, the diagnosis of mental states in the time domain is examined. Then, a window is determined on the measured signal and in these windows, various features are extracted and by using these features and machine learning methods, different mental states are categorized. Finally, the method used is tested on the data set and the results of the method is checked. One of the advantages of the proposed method is to reduce the number of network inputs based on PCA feature reduction method, which leads to a reduction in network volume, which is especially important in neural networks. In this article, we have tried to increase the accuracy of classification by using various features. Finally, we use to-fold cross validation.