Boggle : an SSVEP-based BCI web browser

Abstract

Brain-Computer Interfaces (BCIs) have led to significant enhancements in the lives of physically-restricted individuals. BCIs based on Steady State Visually Evoked Potentials (SSVEPs) are robust and rely on a neuronal response evoked when a person focuses attention onto a flickering visual stimulus. Our first study, which provided empirical insights on web technologies’ applicability for SSVEP stimuli-generation, demonstrated that both Cascading Style Sheets (CSS) and Web Graphics Library (WebGL) can produce effective stimuli via square wave approximations, using Google Chrome and Mozilla Firefox. Building upon these findings, this work explores the feasibility of adopting these technologies to implement an SSVEP-driven web browser, supporting online and asynchronous BCI-based control. Informed by a systematic review of literature and a succession of user-centred studies, this paper discusses results produced throughout the development of Boggle - a novel SSVEP-based BCI web-browser. As for in-browser stimuli-generation, enhanced stimuli efficacy was observed when adopting a custom-developed CSS-based stimuli-generator on Chrome, particularly in high-load rendering conditions. In turn, this contributed to increased classification accuracy and Information Transfer Rates (ITRs), compared to other BCI-based browsers. When evaluated within an online, asynchronous BCI context, participants achieved a global mean classification accuracy and ITR of 90.98% and 29.58 Bits Per Minute (BPM) respectively. Moreover, various usability tests were adopted to gauge progress throughout the different iterations. Boggle is the first cross-platform, SSVEP-based BCI browser that is fully developed using web-native technologies, and which exploits approximation techniques for stimuli-generation. Feedback provided by domain experts further highlights Boggle’s suitability as a primary assistive technology.