Classification and Prediction of Human Cognitive Skills Using EEG Signals

In this paper, a novel method to perform binary classification of human cognitive skill using electroencephalography (EEG) signals was developed. The classification was done by assessing activity in the brain, at different frequencies while the subjects were solving a series of arithmetic questions of varying complexity. Recording of brain activity was done through EEG signals with the aid of the EMOTIV EPOC+ neural headset. The acquired data was analyzed and the classifications were accomplished with the help of supervised learning classifiers, namely K Nearest Neighbors (KNN), Support Vector Machine (SVM), Regression, Discriminant, Tree and Ensemble classifiers. The mental skill of the test subject was classified in binary terms of a state of solving/not solving in a manner similar to emotional classification using the valence/arousal model. The results are presented and discussed in detail in this paper. It is believed that this work would lead to the development of accurate methods to predict cognitive skills of human beings for a better understanding of their learning capabilities.