Classification of motor imagery EEG signals using source localization

Brain Computer Interface (BCI) is a communication method which links brain and an external device. Design and implementation of BCIs is an interdisciplinary field of study. The aim is to improve the quality of life of those with neuromuscular diseases or disabilities. There are many commercially available BCIs but, there is still a big demand for user friendly BCIs with higher accuracy. Although EEG signals have high temporal resolution, they have low spatial resolution. In this study, for low spatial resolution is compensated by converting the EEG signals received from the scalp into cortical activations using EEG Source Imaging. Cortical source signals can be classified successfully by adding neuro anatomy and neurophysiology information. This is a multi-disciplinary study with the computer engineering and physiology department of medical science. Here, cortical signals from brain regions corresponding to motor imagery is used to improve the discrimination on the EEG signals by neuroimaging. As a result, successful classification of the EEG cortical signals compared to the signals at the electrode level is accomplished. Minor interpersonal differences in motor imagery regions are also observed.