A direct visualization of the surface boundary of subjective visual space, the intrinsic bias
Poster Presentation 23.420: Saturday, May 16, 2026, 8:30 am – 12:30 pm, Pavilion
Session: 3D Shape and Space Perception: Surfaces, objects
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Lizhu Yan1 (lizhu.yan@louisville.edu), Teng Leng Ooi2, Zijiang He1; 1University of Louisville, 2The Ohio State University
A stationary observer perceives a single monocular target as if located on an implicit quarter-elliptical surface in the dark, which we coin the visual system’s intrinsic bias (Ooi et al., 2001; Bai et al., VSS 2024). We hypothesized the intrinsic bias, serving as a priori, represents the farthest surface boundary of visual space. Should this be true, multiple monocular targets, even when arranged linearly, would be perceived along the curved intrinsic bias. Five LED targets were presented as: (1) vertical line LED-display ~1.5m (eye-height dependent) from the observer, or (2) horizontal LED-display about the floor 1.5–7.0m from the observer. The angular declination of both the vertical and horizontal targets was 5–45 deg. The 5 LEDs (0.23 or 0.16 cd/m2) had the same angular size (0.23 deg) and inter-target angular separation (10 deg). Observers were tested both with monocular and binocular viewing and responded using drawing-on-paper and blind-walking-gesturing tasks. Confirming predictions, observers with monocular viewing drew the perceived linearly arranged vertical and horizontal LED-displays with curved profiles. This was corroborated with the blind-walking-gesturing task that measured the judged locations of single targets within the displays, and without the displays. The average (n=8) judged target location profiles with the vertical and horizontal displays overlapped (ps>0.05) and were well fitted by the quarter-elliptical function, reflecting the intrinsic bias’ shape. We further found with binocular viewing, judged target locations in the two displays deviated away from the intrinsic bias (ps<0.05) and were more veridical. And with the drawing task, observers drew the vertical and horizontal LED-displays as curves that were respectively, stepper and flatter than the intrinsic bias. Overall, the vertical and horizontal LED-displays provided direct visualization of the quarter-elliptical shape of the intrinsic bias. With binocular viewing, the visual system represented target locations by integrating the binocular depth cue with the intrinsic bias.
Acknowledgements: NIH R01EY033190