Evidence for the inhibition hypothesis in expanded angle illusion

Abstract

THE illusory expansion of acute angles has long been suspected to be an important contributory factor in a number of well known illusions, such as the Zollner, the Wundt–Hering and the Poggendorf^1,2. Recently Blakemore, Carpenter and Georgeson³ proposed that angle expansion is a side effect of an inhibitory process that improves the orientation resolution of the visual system. Neural line and edge detectors considered in isolation, are assumed to respond to a wide range of orientations, but when incorporated into a functional system inhibitory interactions occur between them which sharpen the specification of a line or edge⁴. When two spatially contiguous lines of neighbouring orientations are exposed simultaneously, however, the activity peaks in the population of orientation detectors are shifted away from each other because of the inhibitory interactions. Consequently, the orientations of the lines comprising the angle are perceived wrongly. This process of central lateral inhibition is thought to be similar to that which operates in peripheral sensory structures³. It seems natural to ask how far this similarity extends.