A Novel EEG-BCI Dependent on Discrimination of Imagined Stimuli in Visual Field Quadrants

Stojic, Filip 1, 2 ; Chau, Tom 1, 2

1. Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, ON, Canada; 2. Bloorview Research Institute, Holland Bloorview Kids Rehabilitation Hospital, Toronto, ON, Canada

Brain computer interfaces (BCIs) can provide those living with severe motor disorders a means of communication not dependent on motor actions or verbalization. By measuring brain activity elicited from different mental tasks, communicative intent can be identified in these individuals. However, many of these mental tasks do not come naturally, are unintuitive, demand too much effort, or rely on neural pathways that have been compromised by motor control disorders. Brain activity during visual mental imagery is detectable and does not rely on potentially compromised neural pathways. This study aims to determine whether imagined stimuli in the four quadrants of the visual field can be used to signify intent in both an offline and online electoencephalography (EEG)-based BCI. Additionally, the extent to which performance on this task relies on individual visual cognitive ability will be assessed. Participants will have their visual cognitive capabilities assessed with the Rey-Osterrieth Complex Figure (ROCF) test and Vividness of Visual Imagery Questionnaire (VVIQ2) prior to the BCI visual mental imagery task. Over the course of 3 sessions, participants will be trained to imagine checkerboard arrow stimuli in the four quadrants of the visual field while having their brain activity recorded with EEG. On the 4th and 5th sessions, they will be asked to navigate through an online navigation task by imagining the arrow stimuli in the quadrant they wish to move towards. It is expected that classification accuracies of the offline and online sessions will correlate to ROCF and VVIQ2 scores, and that user satisfaction will be high. Furthermore, it is hypothesized that both offline and online classification accuracies will achieve above-chance levels. This study will shed light on the potential of determining individual BCI performance using standardized cognitive assessments, and importantly, validate a practical and novel approach to identifying user mental state using visual mental imagery.