Self-Paced Brain-Computer Interface on Sensorimotor Rhythms for Controlling Virtual Objects

Non-invasive electroencephalogram (EEG) based Brain-Computer Interface (BCI) systems have become a fascinating study area in many research fields. A majority of the research in this area is conducted synchronously. Hence, at the time of the experiment, the user’s state of mind tends to be different from its natural behavior. As a solution to this problem, self-paced BCI systems started gaining popularity in recent years. However, certain challenges remain to be addressed even with this method. Most of the research on self-paced BCI systems is focused on motor-imagery control, whereas research on non-motor imagery mental tasks is limited. However, individuals with severe paralysis may face challenges in performing motor imagery tasks. In this research, we explore the possibility of using the techniques used in the motor-imagery method for non-motor imagery mental tasks. The intention here is to use them in virtual object-controlling applications. The research was done with five different classification models using features from the Fast Fourier Transform (FFT) and Wavelet Transform (WT). The K-nearest neighbor model with features obtained with FFT continuously sustained its performance with a high 55% true positive rate.