The 2 Streams Hypothesis of Vision
Most popularized by Milner and Goodale in 1992, this hypothesis presents the argument that human neural processing of visual information splits into two distinct streams after reaching the primary visual cortex. Though the model faces criticism for overemphasizing the independence of these pathways— it is more likely that there is interaction between them—, it provides a solid framework for explaining optical phenomena such as the jitteriness of the object borders seen in the illusion above, which feature equiluminant colors (they would be the same shade if you changed the image to grayscale). Considerable evidence gives credence to this model wherein the ventral stream (referred to as the “what” stream) primarily guides object recognition, is better at coding complex visual details, sees in color, has lower contrast sensitivity, and is characterized by slower response time. In contrast, the dorsal stream (“where”) is evolutionarily older, codes for motion and space, is more sensitive to contrast, has shorter response times, but sacrifices resolution. This disparity can be explained by the fact that the two streams are fed by distinct inputs from different populations of retinal cells, which specialize in coding different visual information. The border where the turquoise meets equiluminant red seems to pulsate because the object is seen by the “what” stream, which is fed by retinal cells that take in color information, but not— or not as well— by the “where” stream, since the latter pathway is fed colorblind inputs. Also complicating the perception of the border is the fact that the “where” system’s sensitivity to contrast and quicker neuronal responses causes the brain to constantly re-evaluate the figure-ground relationship. So, moment by moment, your brain changes its mind about which color is darker than the other, and the image seems to pulsate in response to these quick recalculations.
Guest post written by Rachel
Hey guyyssss, my submission to the awesome science blog, sciencesoup, got posted! \o/


