Brain-Computer Interface based Web Browser using Steady State Visual Evoked Potentials

Abstract

A Brain-Computer Interface (BCI) is a communication and control interface, used to transfer commands from the human brain to the computer allowing disabled patients to interact with their surroundings. Applications of BCI systems have spread in multiple directions, including the control of assistive devices, communications applications and even computer gaming. Since using the internet shall be a right for every human being, disabled patients need special web browsers that are capable of providing web navigation using their brain signals. BCI web browsing has become a challenge from multiple aspects, as it needs to assure the least time of processing brain commands, keeping the highest detection accuracy, as well as having an easy to use user interface. In this thesis, a BCI-based web browser software is introduced. It utilizes the Steady State Visual Evoked Potentials (SSVEP) as the main paradigm of processing the electroencephalography (EEG) signals of the brain. In addition to the simplicity of the user interface, the software uses an enhanced SSVEP processing algorithm named Principle Component Analysis-based Spectral Recognition (PCA-SR) which is introduced in this thesis that achieves high levels of detection accuracy and low response time. The introduced web browser has been demonstrated to be functional by 3 healthy users using the wireless Emotiv EPOC headset performing various web navigation tasks including searching, navigating hyperlinks in addition to saving bookmarks, with a web browsing accuracy of 86.08±15.46%. Key words: Brain-Computer Interface, EEG, SSVEP, Web Browsing