“Functional Micro-Architecture for Shape and Color in Primate Primary Visual Cortex”
Faculty Host: Elizabeth Gould
Color vision is probably the most compelling example of the artificial relationship between perception and reality. For example, we know that violet and red monochromatic light are at the furthest extremes of the visible spectrum, yet we perceive hues as a circular continuum that links blue to red through violet. And mixtures of wavelengths do not create a mixed sensation like they do for sound – the 3 notes of a chord sound distinct rather than blending together to sound like the intermediate tone. Instead, red and green when mixed is perceived as yellow and adding blue to yellow creates white. These discontinuities between the physical world and perception have fascinated physicists and psychologists for centuries. How do we create the perception of thousands of distinct hues from the relatively impoverished visual representation provided by just three cone types? And how do the neural representations of color interact with mechanisms for the perception of shape and motion? We achieved stable expression of GCaMP6 in V1 of macaque monkeys and used two-photon calcium imaging to reveal the functional micro-architecture of color tuning. We characterized responses of more than 4,000 neurons across a very large visual stimulus space to reveal the joint representations of color, orientation, spatial frequency and ocular dominance. We directly related color responses to maps of the cone inputs to individual neurons and to cytochrome oxidase histology. Our results demonstrate functional micro-architecture for color tuning within V1, a surprisingly high proportion of color-preferring neurons, and some unexpected relationships to the anatomical cytochrome-oxidase “blob” landmarks. Despite the presence of systematic color mapping, as well as a strong inverse relationship between color selectivity and orientation tuning, we demonstrate that there are many neurons that jointly represent both color and shape. This is contrary to the cartoon view presently emphasized in most textbooks, claiming that shape and color are extracted independently within V1 and sent off to separate higher-level visual areas.