Independence of color and shape processing in the ventral visual pathway of humans and macaques
26.4009, Saturday, 16-May, 2:45 pm - 6:45 pm, Pavilion
Rosa Lafer-Sousa1,2, Nancy Kanwisher1,3, Bevil Conway1,2,3; 1Brain and Cognitive Science Department, MIT, 2Neuroscience Program, Wellesley College, 3McGovern Institute for Brain Research
To what extent are color and shape processed independently of each other in the ventral visual pathway? Color is sometimes assumed to be just another stimulus feature for object recognition. On this view, there is a computational advantage to multiplexing color and shape processing. But the importance of color can sometimes be distinguished from shape: color signals a person’s emotional state, independent of face features; similarly, color provides independent information about fruit’s edibility (bananas ripen, green to yellow, without changing shape). We asked whether distinct regions are sensitive to color and shape by scanning 13 humans and 2 rhesus monkeys with fMRI while they viewed full-color or achromatic movie clips of faces, objects, scenes, bodies, or scrambled objects. We defined color, face, place, and shape-selective regions within each subject, and measured the response magnitude of each region to the ten conditions in independently acquired data. We confirmed the existence of three color-biased regions in humans, and determined that none of these regions showed an interaction of color preference with stimulus category; moreover, two of the color-biased regions showed no higher response to intact versus scrambled objects, suggesting these regions code color independent of shape or stimulus category. Furthermore, shape-biased cortex (LO) did not show a bias for colored objects, providing a double dissociation of color and shape processing. Consistent with our previous report on responses to static stimuli in monkeys, a preliminary analysis of monkey data obtained using movie clips suggests that color-biased regions in inferior temporal cortex (IT) are no more responsive to intact than scrambled shapes, and that IT shape-selective regions are less sensitive to color compared to the color-biased regions. Taken together, these data support a dissociation of color and shape processing at certain stages of the visual-processing hierarchy in both humans and monkeys.