Recognition of Motor Imagery Using Spatial filters and Adaptive Neural-Fuzzy Classifier

in this study, Recognition of motor imagery EEG is performed using spatial Filters-based feature extraction method. The dataset used, contains data from BCI competition IIIa. EEG signal analysis system consists of three steps: preprocessing, feature extraction and AFNN clustering. In preprocessing, continuous wavelet transform (CWT), the 2D anisotropic Gaussian filter and Student’s two-sample t-statistics are used to select active segments of the signal. Common spatial pattern (CSP) is applied to extract data features and Adaptive Neural-Fuzzy clustering (AFNN) is used for recognition of left and right MI data. The system is tested on available EEG datasets and compared with previous popular approach. The experimental results demonstrate that active segment selection can greatly improve the performance and the method of CSP is a good choice to combine with AFNN clustering in terms of achieving high classification accuracy. So, it has a splendid potential in the applications of brain-computer interface (BCI) work.

کلیدواژه ها:

Electroencephalogram ، Brain–computer interface ، Common spatial pattern ، Adaptive Neural-Fuzzy Classification