Selecting and Extracting Effective Features of SSVEP-based Brain-Computer Interface

User interfaces are always one of the most important applied and study fields of information technology. The development and expansion of cognitive science studies and functionalization of its tools such as BCI1, as well as popularization of methods such as SSVEP2 to stimulate brain waves, have led to using these techniques every day, especially in appropriate solutions for physically and mentally handicapped people. Computer- brain interfaces enable users to communicate without involvement of their lateral muscles and nerves and since these interfaces are not dependent on muscular nervous control, they enable people with muscle neuromuscular control disorders (such as amyotrophic lateral sclerosis, brain stroke, cerebral palsy, etc.) to control and communicate. The main idea in this research is the implementation of a proposed system to provide the best options to the user due to the limitations of the simultaneous options in the SSVEP method and the specific user conditions for disabled people. To do this, we present a new implementation method based on extracting the wavelet feature and then dimension reduction by PCA3 and after the extraction step, the features are classified by the SVM4 and KNN5 classifiers. It has been observed in this project that99.3% accuracy can be achieved by KNN classifier.