Local and global integration in spatial localization

Abstract

E. Herring (1899) proposed that spatial integration is important for precise spatial localization. To investigate the role of spatial integration in foveal spatial hyperacuity, we measured three line bisection threshold. The center horizontal test line was also continuous (100% sampled), while the outer reference lines were randomly undersampled by different percentages. We varied the interline separation, line length, sample size, sampling percentage and line contrast. Our results suggest that there are two independent forms of spatial integration in spatial localization: local integration, this was revealed by varying the percentage of samples; and global integration, this is seen by varying the line length. These two forms of integration show the following properties: (1) local integration has a steeper slope (on a log-log plot of threshold vs sampling percentage) and is much more efficient than global integration; (2) integration occurs even when the effect of reduced sampling on stimulus visibility is accounted for; (3) integration saturates; the saturation point depends on interline separation. Our results suggest that there exist two independent local and global integration mechanisms and both contribute to the achievement of precise spatial localization in the human visual system.