Abstract
The blind spot’s V1 representation receives monocular input only. Previous studies suggest that illusory contour (IC) interpolation through the blind spot is impaired compared to intact regions of the retina (Maertens & Pollmann, 2007) or scotomata (De Stefani et al., 2011). One explanation is that ICs require local signals to spread within monocular channels in V1. However, these studies used a shape discrimination task which might be less sensitive than tasks probing the contour directly, for example a dot localization paradigm (Guttman & Kellman, 2004). Here we used dot localization to test curvature perception of ICs and real contours (RC) across the blind spot.
We presented curved ICs induced by Kanizsa-type figures or RCs, either through the blind spot or occluders. Additionally, ICs were shown uninterrupted (Control), while RCs were interrupted by a gap (“Deleted” condition). A dot briefly flashed near the contour midpoint. Participants judged whether it appeared inside or outside the shape. Psychometric functions for the proportion of “outside” responses as a function of dot position measured underestimation of curvature and precision of contour interpolation.
Generally, ICs were flatter than RCs. Crucially, blind spot ICs had the same precision and flatness as Control, suggesting intact representations of ICs across the blind spot. Occluded ICs had low precision and were completely flat. For RC, curvature was significantly flattened in the blind spot, but not in Occluded or Deleted conditions, suggesting modal contours are flatter than amodal.
Using this more sensitive dot localisation task, our experiment found that ICs are represented across the blind spot in the same way as in other parts of the visual field, suggesting a monocular V1 region does not disrupt IC interpolation. Results also corroborate previous findings on modal and amodal curvature differences (Singh, 2004; Guttman & Kellman, 2004).