Abstract
The fovea is a specialized region of the primate retina whose high density of photoreceptors underlies the high acuity and color sensitivity of our central vision. Despite this, much fMRI research into the basic responses of early visual cortex has used peripherally presented stimuli, often accompanied by an irrelevant task at fixation. Here we test the extent to which fMRI responses of a range of visual areas were dominated by the foveal part of the stimulus. We used radial sinusoidal gratings of four types, designed to target different mechanisms that vary in their distribution across the retina: L-M isolating (red-green), S-cone isolating (blue-yellow), achromatic P-type (high SF, low TF), and achromatic M-type (low SF, high TF). In Experiment 1, these stimuli were presented at each of 3 sizes: 1.4, 4.1 and 15.0 degrees. In Experiment 2, each stimulus was composed of different stimulus types in the center (1.4 degrees) and annular surround (1.4–15.0 degrees). Across runs, participants attended to either the center or the surround stimulus. We used a series of classification analyses to measure how well each region’s pattern of activity could be used to decode the type of stimulus and its size (Experiment 1) or the type of stimulus in the center and surround (Experiment 2). Early visual areas (V1, V2, V3) had relatively more information about stimulus size than stimulus type, whereas for later visual areas (V3A/B, LO, hMT, hV4, VO1 and VO2) this was reduced or reversed. Across stimulus types, there was no evidence for different patterns of spatial summation across stimuli targeting mechanisms with differing retinal distributions. In Experiment 2, decoding of the of the surround stimulus increased when attended across all cortical areas. Decoding of the center stimulus only increased with attention in VO1/VO2, suggesting a foveal/peripheral asymmetry in attentional enhancement
Acknowledgement: CIHR grants 153277 and 10819 to Kathy T. Mullen