BCI Implementation with Short-Time SSVEP Using Canonical Correlation Analysis

Recently developed effective methods for detection commands of steady-state visual evoked potential (SSVEP)-based brain–computer interface (BCI) that need calibration for visual stimuli, which cause more time and fatigue prior to the use, as the number of commands increases. This paper develops a novel unsupervised method based on canonical correlation analysis (CCA) for accurate detection of stimulus frequency. Canonical correlation analysis (CCA) is commonly used to recognize the frequency of steady state visual evoked potential (SSVEP) for the implementation of brain computer interface (BCI). The performance of CCA is degraded when lower data length is used. On the other hand, BCI implementation becomes more effective when it uses lower data length i.e. lower calibration time. This paper presents a CCA based approach to enhance the frequency recognition accuracy of short-time SSVEP signal. To decrease the calibration time, a shorter data is concatenated to increase the data length for better fit of using CCA. The multiset CCA (MsetCCA) is employed to derive the reference signal from the training set and then traditional CCA is used to recognize the frequency of short-time SSVEP. The performance of the proposed method is evaluated using publicly available dataset. The experimental results show that the newly introduced method performs better than the recently developed algorithms.