Motor imagery tasks-based EEG signals classification using tunable-Q wavelet transform

Motor imagery (MI) tasks-based brain–computer interface (BCI) system finds applications for disabled people to communicate with surrounding. The BCI system reliability is relied on how well the different MI tasks are assessed and identified. Electroencephalogram (EEG) recordings provide a noninvasive way for imaging of MI tasks in BCI system. In this framework, tunable-Q wavelet transform (TQWT)-based feature extraction method is proposed for the classification of different MI tasks EEG signals. The TQWT parameters are tuned for the decomposition of EEG signal into sub-bands. Time domain measures of sub-bands are considered as features for MI tasks EEG signals. The TQWT-based features are tested on least-squares support vector machine classifier for the classification of right-hand and right-foot MI tasks. The proposed method provides 96.89% MI tasks classification accuracy, which is the highest as compared to other existing same data set methods. The suggested method can be used for identification of MI tasks in a BCI system designed for controlling robotic arm and wheel chairs, etc.