Fig. 7.
Comparison across subjects in the response of individual visual areas to shapes defined by real and illusory contours. The bar graphs in A–G show the average fMRI signal change for individual visual areas across all subjects tested (A–D, n = 11; E,n = 12; F, n = 9;G, n = 7). Data from corresponding visual areas in the left and right hemispheres areas are averaged together. Error bars indicate SEM. Plus signs and asterisks indicate the signal modulations that are significantly different from zero based ont tests at p < 0.05.Asterisks indicate modulations with pvalues that survive Bonferroni correction. A–G, Thebullets with heavy error bars above each bar indicate the increased modulation that could be detected when the regions of interest were restricted to the 3–9o eccentricity representation in the retinotopic areas. A, B,Isoeccentric contours defined by luminance and stereopsis, respectively. C, Comparison between aligned inducers and rotated inducers. D, Grating-based illusory contour versus nondisplaced grating control (lowest spatial frequency case).E, Aligned inducers versus aligned inducers with luminance occluder. F, The locations of the ROIs are shown on the flattened cortical surface of an individual subject in schematic form. The fMRI signals are strongest in higher-tier areas for the stereopsis-defined shape, and the shapes defined by illusory contours.