Item

Abstract

To interpret complex and ambiguous visual input, the visual system uses prior knowledge, or assumptions about the world. These 'priors' could be hard-wired, or learnt in response to statistical regularities in the environment. Here, we consider the 'light from above' prior used by the visual system to extract shape from shading. Observers viewed monocular disks with shading gradients at various orientations. Reported shape (convex or concave) as a function of stimulus orientation was used to recover each observer's assumed light position. During training, observers also 'touched' the disks. The haptic (felt) shape of the training stimuli was consistent with a light source shifted by ±1 30° from the observer's original assumed light position. After training, observers again judged the stimulus shape from purely visual information. Additionally, observers made lightness judgments of a Mach-card type stimulus, before and after haptic training with the concave/convex disk stimuli. Initially, our observers assumed a light position that was roughly overhead. However, after haptic feedback, observers learned to use a shifted light direction for their prior. Importantly, this learning was not specific to the trained task, but generalised such that it affected visual perception in a separate lightness judgment task.