Amodal completion impairs stereoacuity discrimination.

Abstract

Visual stimulus configuration can influence elementary visual processes. We provide empirical evidence to demonstrate this effect in stereoscopic depth discrimination. Two vertically aligned bars were presented in stereo such that one of them was closer to the human observer. Observers discriminated which of the two was closest. In the first, "occluded" condition, a horizontal bar, positioned closest in depth to the observer, was added to the display such that the two vertical bars perceptually completed to form a whole by connecting together behind the horizontal bar. In the second, control condition, the horizontal bar was placed furthest away from the observer such that there was a visible gap between the two vertical bars, which could no longer complete perceptually. We measured observers' psychometric functions using the method of constant stimuli, and found that their discrimination sensitivity d' was smaller when the two vertical bars perceptually completed than when they did not. We used a simple model to illustrate that when the two vertical bars perceptually completed, they also tended to be perceived as coplanar in the fronto-parallel plane. This consequence of completion made it more difficult to discriminate any difference in depth between the two vertical bars.