An Evaluation of Performance for an Independent SSVEP-BCI Based on Compressive Sensing System

This paper aims a comparison of performance on classification between two feature extractors after of a novel process of compression and reconstruction of signals proposed. In this way, Multivariate Synchronization Index (MSI) and Canonical Correlation Analysis (CCA) techniques of feature extraction were used to get the frequency used from visual stimuli. This system is based on detection of visual attention in SSVEP-BCIs for people with severe motor disabilities. Five male subjects (29.8 ± 2.17 years) participated in this study. According to the results, the MSI technique showed better results in terms of accuracy compared to CCA. It was demon- strated that the proposed system based in MSI technique can offer acceptable performance for a high compression ratio compared to CCA technique. Consequently, the power- consumption in wireless systems can be significantly reduced.