Spatial location and hyperacuity: the centre/surround localization contribution function has two substrates

Abstract

Vernier acuity and jump detection were investigated using a perturbation technique, in which a flanking line is placed to one side of the target line. The size and direction of vernier displacement, or jump, required for no apparent change of location is strongly influenced by the separation between the flanking line and the test line and by its polarity. For flanks within a zone extending approximately 3'-4' to either side of the target line, the target's location is assigned to a weighted centroid of the complete luminance distribution: The target is pulled towards the flank, when the flank has a positive contrast polarity, and repelled when the polarity is negative. The effects of a dark flank on one side and a bright flank on the other are additive. Outside this central zone repulsion effects are obtained independent of the contrast polarity of the flank and flanks on opposite sides of the target line can cancel each other's influence. Varying the duration of the flank produces maximal effects in the surround with shorter duration than that required for maximal effects in the centre. Thus, while the localization contribution function resembles the popular difference of gaussians receptive field profile, it has two components reflecting differing mechanisms. In the centre the earlier centroid hypothesis can be applied with the addition of distance dependent weights. The surround has characteristics resembling the feature interaction seen in figural after-effects.