A Dictionary-Driven SSVEP Speller with a Modified Graphical User Interface

摘要

BCI spellers are mainly composed of an interface, by which alphanumerical characters are presented to users, and a classification system, which identifies the target character online by using the acquired EEG data. In this study, we proposed modifications both to the graphical user interface and to the classification system of the SSVEP-based speller, in order to improve the usability of the Bremen-BCI and to consequently increase the information transfer rate (ITR). The performances of writing three phrases (including every letter of the alphabet at least once) using the standard Bremen-BCI speller and the proposed dictionary-driven SSVEP speller with a modified interface were compared with each other. Seven subjects, in copy spelling mode, used a conventional speller and an improved BCI speller incorporating a custom-built dictionary with a mean ITR of 29.98 ± 5.79 bit/min and of 32.71 ± 9.18 bit/min, respectively. For the vast majority of users, this new dictionary-driven BCI realization resulted in a significant increase in spelling performance.