One of the goals of vision research is to understand how neurons at various levels in the visual processing stream (the eyes, the primary visual cortex-V1, and the higher cortical areas) code progressively more complex aspects of an image. This project concerns the coding of shape. Previous research established that initial shape coding occurs in the primary visual cortex (V1) where the retinal image is analyzed by "feature detectors" which determine the dominant orientations and spatial frequencies at each position on the image. The next level of coding involves shape. Since we are able to see whole shapes (e.g., see a triangle as "triangle" rather than as "a collection of 3 oriented lines"), global shape attributes must also be coded in the brain. Because we can recognize shapes regardless of where they are located, the coding must achieve some level of position-independence. Some neurons in the higher cortical areas, such as inferotemporal cortex, have been shown to be selectively activated by various geometric shapes, almost regardless of where those shapes are presented. To determine what perceptual qualities these neurons are coding, a new behavioral paradigm will be used that examines what are called "shape contrast effects." Two shapes are presented successively, a prime shape followed by a visual mask and a test shape. When these stimuli are presented at optimum timing, the test shape appears distorted in such a way that it seems to be more dissimilar to the prime shape than it actually is. For example, a tall prime shape makes a test shape appear flatter, and a prime shape tapered in one direction makes a symmetric test shape appear tapered in the opposite direction. Importantly, such global shape-contrast effects occur even if the two shapes are presented at separate locations, suggesting that the effects are tapping into position-independent coding of global shape attributes. Perhaps, overall aspect ratio (taller/flatter) and taper may be two such global shape attributes systematically coded by the neurons in the higher cortical areas. One goal of this research is to understand the distinction between the type of shapes that produce shape-contrast effects and the type of shapes that do not. Such an effort will help determine what global aspects of shapes are selectively coded by the visual system. To complement this exploratory approach, experiments will be conducted to test the hypothesis that shape-contrast effects may tap into a specialized processing "module" that encodes a specific family of image transformations. Coding of such transformations may be a part of the mechanism underlying our ability to recognize objects from different points of view. Additional experiments will be conducted to determine how shape-contrast effects are modulated by various stimulus and cognitive factors such as visual context, quadrants of the visual field, various forms of attention, experience, conscious awareness, central vs peripheral vision, and the degree of shape similarity between the prime and the test item. The proposed research will evaluate this new technique and use it to explore the image-coding structure of a global shape representation
