After a three letter sequence is presented, volunteers must respond as quickly as possible with regard to whether or not the word that appeared matched the word spelled by the letter sequence that came before it.

ICA and dipole fitting yielded multiple regions related to reading. Within each horizontal slice, individual participants’ ICs are represented by blue dots. Red dots are the cluster centroids; for each cluster the average residual variance of included ICs is less than 6.5% (see ). The indicated z-coordinates are those of the cluster centroids (red dots). vOT ventral occipito-temporal cortex; AG angular gyrus; STG superior temporal gyrus; IFG inferior frontal gyrus; d dorsal; v ventral.

(a) Topographic maps of group average ICs for the ROIs shown in . (b) Topographic maps of the individual participants’ ICs for the vOT illustrate their resemblance to the group average IC. Results are similar for other ROIs. vOT ventral occipito-temporal cortex; AG angular gyrus; STG superior temporal gyrus; IFG inferior frontal gyrus; d dorsal; v ventral.

Power in dB with respect to a baseline from −250 to −50 before presentation of the first letter. All ROIs show a prominent burst in theta (3–7 Hz) power, the earliest activity occurring immediately after word presentation (<50 ms), and the latest occurring ∼600 ms after the onset of the word. vOT ventral occipito-temporal cortex; AG angular gyrus; STG superior temporal gyrus; IFG inferior frontal gyrus; d dorsal; v ventral.

(a) vOT shows distinct patterns of functional connectivity among the regions in the reading network. PLVs are masked for significance at p = 0.001 by permutation test for individual participants and then again at p = 0.000001 for the group (see Methods). (b) Example of theta-band NBTE (30–50 ms lag) between vOT and vIFG from a single subject. The plot depicts the flow of information between areas, where the blue line represents the transfer of information from vOT to vIFG, and the red line represents information flow from vIFG to vOT. In this case, across all lags, vOT sends information from ∼250–470 ms, followed by vIFG sending information back at ∼500–550 ms, indicating an initial feed-forward response, followed by feedback to the original source. Also shown here are instances in which both NBTE patterns overlap, signifying a bi-directional relationship (e.g., ∼300–400 ms). NBTE results for individual participants were masked for significance at p = 0.05 by permutation test (see Methods). vOT ventral occipito-temporal cortex; AG angular gyrus; STG superior temporal gyrus; IFG inferior frontal gyrus; d dorsal; v ventral.

All connectivity analyses were computed over the span of the entire 6(a) Theta-band PLVs (significant functional connectivity, black lines – see Methods) among all ROIs, evolving over the course of viewing a word. (b) Theta-band PLVs relative to vOT. These depictions of connectivity are derived from (a), isolated to bring focus to vOT’s widespread involvement in the reading network. (c) Theta-band NBTE (causal connectivity, black arrows mean that NBTE was significantly different from 0 in both 50-ms bins in 100-ms interval – see Methods) relative to vOT, showing a comparable pattern of network engagement to PLVs (b). This also highlights some of the uni- and bi-directional relationships that exist in the time-course of word reading. (d) Gamma PLVs show similarities to their theta counterparts, albeit with sparser connectivity overall. This is further highlighted when vOT gamma interactions are isolated in (e), illustrating their confinement to posterior cortices. (f) Gamma-band NBTE (black arrows) relative to vOT, showing a comparable, and perhaps even more robust, pattern of network engagement relative to PLVs (e).