When geometry constrains vision: Systematic misperceptions within geometrical configurations

How accurate are we in reproducing a point within a simple shape? This is the empirical question we addressed in this work. Participants were presented with a tiny disk embedded in an empty circle (Experiment 1 and 3) or in a square (Experiment 2). Shortly afterwards the disk vanished and they had to reproduce the previously seen disk position within the empty shape by means of the mouse cursor, as accurately as possible. Several loci inside each shape were tested. We found that the space delimited by a circle and by a square is not homogeneous and the observed distortion appears to be consistent across observers and specific for the two tested shapes. However, a common pattern can be identified when reproducing geometrical loci enclosed in a shape: errors are shifted toward the periphery in the region around the center and toward the center in the region nearby the edges. The error absolute value declines progressively as we approach an equilibrium contour line between the center and the outline of the shape where the error is null. These results suggest that enclosing an empty space within a shape imposes an organization to it and warps its metrics: not only the perceived loci inside a shape are not the same as the geometrical loci, but they are misperceived in a systematic way that is functional to the correct identification of the center of the shape. Eye movements recordings (Experiment 3) are consistent with this interpretation of the data.