Multi-modal EEG online visualization and neuro-feedback

Abstract

Brain computer interface (BCI) is a communication pathway between brain and peripheral devices, which is promising in the field of rehabilitation and helps to improve the life quality of physically challenged people. Analysis of EEG signal is essential in non-invasive BCI system. However, because of EEG signal's low signal-to-noise ratio and huge amount of data, signal analysis function in current BCI systems is rarely available online, which is inconvenient for system adaptation and calibration, as well as comprehension of data's characteristics. To address the problem, this paper presents two features that are suitable for online visualization. Rhythm power indicates active brain region, and filtered ERSP (Event related spectrum power) is a substitute for original ERSP which provides information in signal's frequency domain. Moreover, visualization of CSP (Common Spatial Pattern) feature is also realized which serves as an indicator of epochs' quality. © 2010 Springer-Verlag.