Comparison of EEG signals of brain tumor patients between two cognitive skills “working memory“ and “focused attention“ using machine learning tools

Using the EEG signal is very helpful for detecting emotional and cognitive states. Feature selection and extraction from the recorded signal is an important part of this article and machine learning tools are used. The data used in this study is related to the task of working memory and focused attention from 5 subjects with brain tumor. In this way, after performing the famous Sternberg Task (SWMT) and listening to a speech, the EEG signals were received and stored by the device. This experiment was conducted to detect 5 different classes and the data is stored in 14 different indexes for each sample. In the set of methods used in this study, KNN and SVM methods performed better than others. Also, in terms of class separation, these two methods were the best. In terms of index selection, there is no improvement in classification accuracy, which of course could be due to the number of data indices not being very large. In terms of execution speed, the KNN method is much faster than SVM, so it may be the best method in terms of accuracy and speed.

Comparison of EEG signals of brain tumor patients between two cognitive skills “working memory“ and “focused attention“ using machine learning tools Keywords:

Working memory , focused attention , machine learning , electroencephalography , Brain tumor