Fig. 4. Increased feedback influences during the preferred percept and increased feedforward influences during the non-preferred percept across large-scale cortical networks.

A For each inter-lobe electrode pair, Granger causality was calculated in each direction, separately for trials during preferred and non-preferred percept. Trials were dropped from the preferred percept to yield an equal number of trials between percepts. To assess significance, the difference in Granger causality between percepts was then compared to a null distribution created by shuffling the trial labels 1000 times. B Pairwise connections between lobes that showed significantly (p < 0.002, uncorrected) increased Granger causality during the preferred (left) or non-preferred (right) percept. Color indicates the direction of the connection on the Sagittal plane coded by the combination of anterior ↔ posterior and dorsal ↔ ventral directions (see center color wheel). Thus, feedforward connections are primarily red and feedback connections are primarily blue. Dorsal→ventral and ventral→dorsal connections are coded as black and white, respectively; medial-lateral dimension is not color-coded. C To aggregate the results across many electrode pairs, a bias measure was calculated by comparing the number of significant inter-lobe connections in each direction using a two-sided binomial test, separately for each perceptual preference. D Significant (p < 0.05, uncorrected) biases (as calculated in C) in inter-lobe connections for those that are stronger during the preferred percept (left) or stronger during the non-preferred percept (right). Line width indicates significance of binomial test. Source data are provided as a Source Data file.