Independent bilateral-eye stimulation for gaze pattern recognition based on steady-state pupil light reflex

Area of Science:

• Biomedical Engineering
• Neuroscience
• Human-Computer Interaction

Background:

• Brain-computer interfaces (BCIs) offer non-contact interaction methods.
• Pupil oscillations synchronized with visual stimuli can serve as a BCI input.
• Current pupil oscillation BCIs are limited by low stimulus frequencies, restricting information transfer rate (ITR).

Purpose of the Study:

• To investigate methods for increasing the ITR of pupil oscillation-based BCIs.
• To compare the efficacy of monocular-single, monocular-superposed, and binocular-independent stimulation conditions.
• To optimize stimulus parameters for enhanced BCI performance.

Main Methods:

• Compared monocular-single, monocular-superposed, and binocular-independent stimulation conditions.
• Selected a square wave stimulus for efficient pupil response.
• Evaluated spectral leakage and power spectral density for different conditions.
• Classified stimulus patterns at 5, 10, and 15 stimuli levels.

Main Results:

• Square wave stimuli efficiently elicited pupil responses.
• The monocular-single condition showed the greatest power spectral density at stimulus frequencies.
• The monocular-single condition achieved the highest ITR (23.7 bits min⁻¹) with 15 stimuli.
• No significant ITR difference was found between binocular-independent and monocular-superposed conditions.

Conclusions:

• The monocular-single stimulation approach significantly increases ITR in pupil oscillation BCIs.
• This method allows displaying more simultaneous stimuli without compromising performance.
• Findings suggest broader applicability of pupil oscillation BCIs for diverse users.