A Biologically Inspired Approach to Frequency Domain Feature Extraction for EEG Classification

Classifcation of electroencephalogram (EEG) signal is important in mental decoding for brain-computer interfaces (BCI). We introduced a feature extraction approach based on frequency domain analysis to improve the classifcation performance on diferent mental tasks using single-channel EEG. Tis biologically inspired method extracts the most discriminative spectral features from power spectral densities (PSDs) of the EEG signals. We applied our method on a dataset of six subjects who performed fve diferent imagination tasks: (i) resting state, (ii) mental arithmetic, (iii) imagination of lef hand movement, (iv) imagination of right hand movement, and (v) imagination of letter “A.” Pairwise and multiclass classifcations were performed in single EEG channel using Linear Discriminant Analysis and Support Vector Machines. Our method produced results (mean classifcation accuracy of 83.06% for binary classifcation and 91.85% for multiclassifcation) that are on par with the state-of-the-art methods, using single-channel EEG with low computational cost. Among all task pairs, mental arithmetic versus letter imagination yielded the best result (mean classifcation accuracy of 90.29%), indicating that this task pair could be the most suitable pair for a binary class BCI. Tis study contributes to the development of single-channel BCI, as well as fnding the best task pair for user defned applications.