Feature extraction of visual evoked potentials using state-space model

Recently, there are many studies on brain computer interface (BCI) system and some use EEG response at oddball paradigms. The aim of this study is to extract the features from the EEG signal speedy and safety. In this paper, we construct the state-space model of the EEG signal, and report the results to extract the features from the EEG signal included (or not included) VEP. It is confirmed that a significant difference about signal-to-noise ratio (SNR) between the measured EEG signal included VEP and one not included VEP. Effective information is obtained from the EEG signal near of the occipital area.