Relevance. Brain-computer interface (BCI) technology is widely employed in studies focused on possibility of substitution of human verbal and motor communication channels when such channels are lost due to stroke or trauma. One of the most promising examples of BCI technology is P300 BCI – a system based on P300 component of visual evoked potential. Visual stimulation plays a key role in P300 BCI concept, and it is thus important to explore the influence of visual stimuli environment on user’s eye movements and BCI performance.

Objective. The objective was to explore the influence of the main components of P300 BCI visual environment, namely matrix stimuli size and spacing, on the movements of user's eyes during typing, the accuracy of typing, and the amplitude and latency of P300 component.

Methods. Sixteen healthy subjects were recruited for the study. Each subject participated in five experimental sessions, where P300 BCI matrix stimuli size and spacing varied from 1.22 and 0.73 to 2.43 and 1.45 degrees of visual angle. During each experimental session, subjects were typing in a certain text sequence using the BCI while their eye movements and EEG were being recorded.

Results. We found that the stimuli size significantly affected certain characteristics of user’s eye movements. The stimuli size and spacing, however, had no significant effect on the accuracy of the typing and the amplitude and latency of P300 component.

Conclusions. The results of the study can prove to be beneficial to the development of efficient visual stimuli environments for P300 BCI.

This paper presents a new class of noninvasive brain-computer interfaces based on user’s silent counting of homogeneous visual stimuli, without defining separate classes of target and non-target stimuli (the “single-stimulus” paradigm), at an average (asynchronous) rate of about 2 stimuli per second and with counting that started immediately after moving the gaze to the stimulus presentation area. Interface’s operation was modeled by applying a linear classifier to EEG signals obtained from healthy participants, in rest condition and after moving their gaze to the stimulus presentation position. A response time of 2 second after the beginning of saccade was demonstrated, which is exceptionally fast comparing to typical results known for the existing noninvasive brain-computer interfaces. This result creates a basis for the development of efficient hybrid eye-brain-computer interfaces, considering that integration of the proposed EEG-based interfaces with the analysis of visual fixations and saccades can further improve recognition of the user’s intentional commands.