Fig. 3: Impact of V4 inactivation on contour-induced modulations in V1.

a Spatial profile of contour-induced modulations in V1 of monkey MA. Gray dots represent individual V1 sites, with pre-cooling d^′ values (9-bar contour vs. noise pattern) plotted against distance from the RF center to the contour path (n = 681 sites pooled across 8 contour positions). Binned averages (open circles) were fitted with a difference-of-Gaussians (DoG) curve. Its intersection with the horizontal dashed line (d^′ = 0) defined the boundary between near and far site groups (vertical dashed line). b Comparison of V4-cooling and sham-cooling effects on V1 near and far sites in MA. Mean neuronal d^′ is shown as a function of contour length for cooling (left column: near sites, n = 340; far sites, n = 341) and sham-cooling (middle column: near sites, n = 328; far sites, n = 335; light shading: SEM). For each site group, the reduction in contour modulation strength (Δd^′) was computed by pooling data across contour lengths. Half-violin plots (right column) show distributions of Δd^′ for cooling (ΔCool) and sham-cooling (ΔSham), with medians (dots) and interquartile ranges (horizontal colored lines). Median ΔCool and ΔSham values were compared for near sites (upper panel; p = 0.054) and far sites (lower panel; p = 4.5 × 10⁻⁵, two-tailed Wilcoxon rank-sum test). c As in a, but for monkey MB (n = 619). Compared with MA, the spatial scale of contour-induced modulations differed substantially (note x-axis range), likely due to differences in RF eccentricities between animals (see Fig. 1c). d Similar to b, but for MB. Left column: cooling experiment (near sites, n = 211; far sites, n = 408). Middle column: sham-cooling (near sites, n = 248; far sites, n = 372). Right column: Δd^′ distributions (upper panel, p = 0.053; lower panel, p < 10⁻⁶). Significance: n.s., p ≥ 0.05; ***, p < 0.001. Source data are provided as a Source Data file.