Figure 6
From: Tracing the Flow of Perceptual Features in an Algorithmic Brain Network
Common Feedforward Architecture.
DFI network schematics represent DFI with color-coded arrows that denote the transfer of the color-framed stimulus features underlying both perceptions in each observer (from Fig. 5 and Supplemental Figures 2–5). Pie charts in the network schematics summarize proportional coding of stimulus features at each node, where blue and red color-code the left and right nuns’ face, respectively, whereas green codes the more idiosyncratic, observer-specific Voltaire features. Color-coded time courses adjacent to the occipito-temporal hubs plot the association (MI) between stimulus feature and network node activity; shaded grey region indicates threshold of statistical significance (p = 0.01, multiple comparison corrected over ICA sources, time points and component filters). Arrowhead labels indicate the receiving time of highlighted DFI. Color-coded numbers near each node indicate the sending time of the features and color-coded numbers near each arrow indicate the corresponding transfer lag to the hub.
