Single-Trial Detection of Error-Related Potential by One-Unit SOBI-R in SSVEP-Based BCI

Abstract

Error-related potential (ErrP) is a form of event-related potential (ERP) that is triggered in the brain when a user either makes a mistake or the application behaves differently from his/her intend. Unfortunately, due to its short-time duration, low signal-to-noise ratio, non-stationarity and transient characteristic, a single-trial extraction of ErrP remains a difficult task. In this study, we propose the use of one-unit second order blind identification with reference (SOBI-R) for extraction of ErrP in the context of steady-state visual evoked potentials based brain-computer interface (SSVEP-based BCI). Fractal features are extracted from the one-unit SOBI-R data by means of Katz fractal dimensional. At last, the ErrP classification is obtained using a regularized version of the linear discriminant analysis (LDA). The proposed method was tested on 6 subjects data and achieved an average recognition rate of correct and erroneous single-trials of $87.03\%$ and $80.7\%$, respectively. These results show that single-trial detection of ErrP is feasible for SSVEP-based BCI.