An optimal spatial filtering electrode for brain computer interface

Abstract

There are millions of people in the U.S. and many more worldwide who could benefit from a noninvasive-based electroencephalography (EEG) brain computer interface (BCI). A BCI is an alternative or augmentative communication method for people with severe motor disabilities. However, EEG suffers from poor spatial resolution and signal-to-noise ratio (SNR). To improve the spatial resolution and SNR many researchers have turned to implantable electrodes. We have previously reported on significant improvements in BCI recognition rates using tripolar concentric ring electrodes compared to disc electrodes. We now report on a optimal method for combining the outputs from the independent elements of the tripolar concentric ring electrodes to improve the spatial resolution further. We used minimum variance distortionless look (MVDL), a beamformer, on simulated data to compare the spatial sensitivity of the optimal combination to disc electrodes and the tripolar concentric ring electrode surface Laplacian. The optimal combination shows the highest spatial sensitivity with the Laplacian a close second and disc electrodes resulting in a distant third. Further analysis is necessary with a more realistic computer model and then real signals. however it appears that the optimal combination may improve the spatial resolution of EEG further which in turn can be utilized to improve noninvasive EEG-based BCIs.