Online improvement of CCA performance in BCI spellers

Official URL: https://dx.doi.org/10.1109/SIU59756.2023.10223931

The performance enhancement of Steady-State vi-sually evoked potential (SSVEP) based brain computer studies are mainly based on finding subject-based patterns and combi-nations. Finding person-based patterns and making use of them for the system involves a training process. It may not be easy or possible for people to attend long and tiring training sessions. The training process can be shortened or even eliminated by integrating already prepared patterns regarding the similarity between subject's data and existing patterns. In addition to the already existing templates, using an adaptive algorithm will eventually come up with patterns that have a better handle on the subjects' data. In this conference paper, we describe the combination of an adaptive algorithm with the integration of various filters formed from already existing data sets.