While color and brightness of an isolated light are closely related to the light's physical properties -- its energy and wavelengths -- this is a misleading fact for understanding normal viewing. The color and brightness of a light in a complex scene are not directly related to the light's physical features. See the two examples on the left.
Color and brightness are not in light. What we see depends directly on a pattern of neural responses, not on the wavelength or energy of light that enters the eye. The simple relation between a physical stimulus and how we perceive it breaks down when the light is part of a complex scene. In natural viewing, the whole visual stimulus is a patchwork of different lights from many objects. The neural response to a particular light, and therefore our perception of it, is affected by the context of the other lights also in view.
Resolution of visual ambiguity
When multiple ambiguous objects are presented in a visual scene, they are often grouped and perceived identically. The neural processes that lead to these grouped percepts can be investigated by presenting conflicting information to each eye. Recent results suggest that grouping is a process driven by binocularly-tuned neurons and that objects with multiple visual features are grouped at the object level, i.e. when all the features have been bound together.
Top: Images presented to each eye are rapidly switched in a presentation method called Interocular Switch Rivalry. The possible stable percepts are shown to each eye, thus the ambiguity created by Interocular Switch Rivalry can be resolved by one eye dominating over the other. Bottom: The image shown to each eye is a patchwork of colored dots, and the ambiguity cannot be resolved by eye dominance. Instead, stimulus information from each eye are combined, such that the emerging stable percepts are of 16 dots of the same color.
Visually ambiguous stimuli are stimuli that can be perceived in multiple ways, either due to the nature of their presentation or inherent properties of the images themselves. Two common types of stimulus ambiguity are binocular rivalry, in which different images are presented to each eye, and image multistability, in which an image has two or more equally valid interpretations. A common example of the second is the Necker cube, which can be interpreted as facing front-upwards or front-downwards. We currently investigate the connections between these two types of visual ambiguity, and how these connections affect perceptual grouping of multiple ambiguous stimuli.
Neural representation of illusions
Chromatic induction is an illusion where the chromaticity (physical measure of color) of a pattern surrounding an area changes the hue of that area. In the figure, the area in question is the "pink" rings connected by the "T" shape. The Flash-Lag effect is an illusion where an object that is flashed very briefly at the same location as an object in motion is perceived to have been flashed at a location lagging behind the object in motion. By combining these two illusions, whether the neural representation color changing over time incorporates the physical chromaticity or the illusory induced color can be investigated.
Top: An example of chromatic induction. The color of the "pink" ring appears different based on the chromaticity of the pattern surrounding the ring. Bottom: Combination of chromatic induction and the Flash-Lag effect. The surrounding pattern changes over time but are the same chromaticity when the pink ring is flashed, yet the pink rings are perceived as different
Top-down effects on perception
Top-down effects on perception are the influences higher cognitive information and stored contextual knowledge have on the resulting percept. One such effect is the Face Lightness Distortion Effect, where a Caucasian face and African-American face will be perceived to be lighter even when the two faces are matched for luminance and contrast. Using a powerful statistical technique called Maximum Likelihood Difference Estimation, the effect can be investigated while controlling for low level differences and demand characteristics of participants.