A pilot case study on the implementation of a motor imagery-based brain-computer interface speller for Vietnamese, a character dense language

By translating brain signals into commands, Brain-Computer Interfaces (BCIs) offer an innovative alternative to traditional input methods like keyboards and mice, establishing a direct communication link between the brain and external devices. One significant application of this technology is BCI spellers, which allow users to type text using mental signals to control an interface. This is particularly beneficial for individuals with severe speech and mobility impairments. This study introduces BCI-vSpeller, a novel Vietnamese spelling system. The BCI-vSpeller system makes two main contributions: (i) A complete BCI system for Vietnamese: This system uses Motor Imagery-Electroencephalography (MI-EEG) signals to control text input and integrate a language model to improve efficiency. (ii) A custom Vietnamese virtual keyboard: The keyboard is specifically designed for the Vietnamese language. It includes a full range of characters, tone marks, and special symbols, with over 160 selectable keys. The BCI-vSpeller system uses MI-EEG signals as input, which are passed through a Linear Discriminant Analysis (LDA) algorithm to classify the MI tasks for its effectiveness in processing subject-dependent data. To enhance typing speed, a lightweight language model is integrated to provide real-time word suggestions. In our initial experiments, the system achieved average typing speeds of 4.95 characters per minute for one participant and 3.45 characters per minute for another. This research is a step forward in the development of BCI applications, showing potential in improving communication accessibility for Vietnamese users.