Objective and non-invasive monitoring of cognitive load through ocular signals has gained increasing attention in human–computer interaction research. This paper presents a controlled experimental protocol and statistical analysis of a structured eye-tracking dataset collected to examine genre-dependent modulation of gaze-derived physiological metrics during interactive digital tasks.

Twenty-seven university students played three video games representing increasing levels of visual dynamics and interaction complexity (Tetris, Sonic Racing, and Fall Guys). Ocular activity was recorded using wearable Tobii Pro Glasses 3 under controlled laboratory conditions. After preprocessing and quality filtering, 78 valid observations were retained for analysis.

One-way ANOVA revealed significant effects of game genre on pupil diameter variability (PDV, p < 0.001), gaze entropy (p = 0.002), blink rate (p = 0.028), and pupil diameter change (PDR, p = 0.036). PDV and gaze entropy showed a progressive increase from structured to highly dynamic environments, indicating higher cognitive and visual demands. Blink-related metrics exhibited differentiated modulation patterns across tasks.

These findings provide preliminary evidence that gaze-derived physiological signals are sensitive to interaction characteristics and visual complexity. The proposed dataset and experimental framework offer a reproducible foundation for future research on gaze-based indicators of cognitive load in interactive environments.