Overview of the experimental design, analytical pipeline, and sulcal definitions. a.

Children were followed longitudinally from approximately 5.5 years of age (Time 1, T1) to age 8.2 (Time 3, T3) with both behavioral assessments and MRI scans, in addition to an intermediate behavior-only assessment at age 7.1 (Time 2, T2). T1-weighted images from T1 and T3 were first processed separately in Freesurfer (cross-sectional processing), and then through the longitudinal pipeline with the creation of a subject template and refinement of T1 and T3 tissue segmentations. b. The occipito-temporal sulcus (OTS, green) and mid-fusiform sulcus (MFS, orange) were defined manually, as shown on 9 sample individual inflated left hemispheres (T1 longitudinal processing). Posterior and anterior sections of ventral temporal cortex were defined in each individual based on a boundary (dotted white line) located at MNI Y=-40 (see projection on Freesurfer average (fsaverage) surface on the bottom left, which roughly aligned with the anterior tip of the MFS). Left hemispheric views from exemplar children with continuous and interrupted left posterior OTS are shown on the top and bottom, respectively. CoS: collateral sulcus; ptCoS/atCoS: posterior/anterior transverse branch of the CoS; IOS: inferior occipital sulcus.

Sample demographics and main neuropsychological scores at the three time points T1-3 (median and inter-quartile range are indicated), for the N=43 children with longitudinal MRI processing.

Cumulative impact of left pOTS interruption on reading scores longitudinally. a.

TOWRE composite standard scores at T3 are significantly higher in children with an interrupted than continuous left pOTS (assessed at T1 with longitudinal processing), after removing residual variance due to age, gender and handedness. Black dots represent mean scores in each group, ± standard error of the mean (s.e.m.). b. Prediction of reading and spelling scores across tests: percentage of variance in different reading and spelling standardized tests explained by left pOTS interruption (≅R2 in models controlling for demographic variables and total brain volume). Left pOTS interruption at T1 was significantly associated with all reading scores at T3, ranging from timed and untimed measures of single word and pseudo-word reading, as well as sentence and passage comprehension or spelling. The bolded TOWRE composite score corresponds to the effect depicted in panel a. c. Percentages of variance in TOWRE T3 reading scores explained by interruption by different sulci in VTC (posterior and anterior OTS (pOTS and aOTS, respectively) and MFS, in each hemisphere). d. Serial mediation model demonstrating the cumulative impact of left pOTS interruption on TOWRE reading scores over time from T1 to T2 to T3. Coefficients are standardized. Reading scores are controlled for age at testing, gender, and handedness, as well as left pOTS interruption by these demographic variables and total brain volume (***: p<0.001; **: p<.01, *: p<.05).

Longitudinal contributions of left pOTS interruption to reading relative to cognitive precursors of reading, as well as their associations. a.

Variance in TOWRE reading scores at Time 3 explained by left pOTS interruption and classical predictors of reading at Time 1 separately, over and above demographic variables, ranked by decreasing contributions (ΔR2 in percentage). b. Associations between left pOTS interruption and predictors of reading at Time 1: percentage of variance (ΔR2) in left pOTS interruption explained by each pre-literacy skill separately, over and above demographic variables and total brain volume. (Stars on each bar denote significant contributions as assessed by nested model comparison using an F-test. **: p<0.01, *: p<.05)

White-matter correlates of left pOTS interruptions. a.

Gyral gaps in the left pOTS were delineated manually (in blue) and contrasted to left pOTS proper (green). Two exemplar left hemispheres are shown in children with a left pOTS interruption. The group-level map shows overlays of left pOTS gaps (projected onto the fsaverage surface) of 15 children with interrupted left pOTS, out of 29 with usable diffusion data. b. Average MD in the cortex of the left pOTS proper and gyral gaps. Top row (from left to right): difference in cortical MD in the left pOTS gap (blue) and left pOTS proper (dark green) in children with an interrupted sulcus, and in left pOTS in children with continuous sulci for comparison (light green). Bottom row, left: example cortical regions in which MD was average in a child with a sulcal interruption. Bottom row, right: MD comparison between children with interrupted and continuous sulci over the left pOTS cortex, including left pOTS gap if applicable. c. Probabilistic tractography from superficial white-matter seeds above the left pOTS cortex and possible left pOTS gap. From top to bottom: example seed in a participant (green: left pOTS proper, blue: left pOTS gap); mean whole-brain connectivity in children with interrupted (n=15) and continuous left pOTS (n=14); statistical differences in connectivity (from left to right): between the two groups not taking into account reading scores, between the two groups when modeling reading scores (no association observed), between connectivity and T3 TOWRE reading scores (p<0.01 TFCE uncorrected threshold).