High-speed noninvasive brain-computer interfaces

Abstract

Brain computer interface (BCI) systems allow interaction with machines through a channel that does not involve the traditional motor pathways of the human nervous system. Thus they can be used by people with severe motor disabilities or those whose limbs are occupied with other tasks. In BCI systems that recently showed greatest interest of researchers, electrical brain activity is measured on the scalp, thus basically they are noninvasive. Using the EEG measurements as the input to the BCI offers the advantages of low cost and high time resolution. However, due to small amplitude of the signal components, relatively high power of noise and poor spatial resolution, achieving large speed, accuracy and the number of targets is a challenge. At present, the steady-state visual evoked potential (SSVEP) BCI paradigm is believed to provide the most promising way of optimizing the BCI performance in that sense. A review of the SSVEP BCI projects is presented, including studies of biodiversity of human EEG response to visual excitation, as well as the design of techniques for visual stimulation, EEG signal acquisition and analysis for best BCI performance. The review is based both on the literature and results of own teamwork.