Table 2.
Summary of ANCOVA results for DTI parameters in white matter tracts (a) and perisylvian cortical regions (b).
| Structural maturation |
|---|
| a. White matter tracts |
| ANCOVA with all tracts averaged over hemispheres |
| age : F (1,20)= 62.6, p < 0.001 |
| tract : F (2,40)= 64.5, p < 0.001 |
| age x tract : F (2,40) < 1, p = 0.734 |
| ANCOVA with only AR and AF |
| age : F (1,20)= 47.6, p < 0.001 |
| tract : F (1,20)= 59.2, p < 0.001 |
| hemisphere : (1,20)= 14.8, p = 0.001 |
| age x tract : F (1,20) <1, p = 0.457 |
| age x hemisphere : F (1,20) <1, p = 0.914 |
| tract x hemisphere : F (1,20) <1, p = 0.673 |
| age x tract x hemisphere : F (1,20) <1, p = 0.796 |
| Posthoc comparisons |
| Left Hemisphere |
| AF vs. AR, t(1,21)= 7.9, p < 0.001 |
| AF vs. ACC, t(1,21)= 9.4, p < 0.001 |
| AR vs. ACC, t(1,21) = 1.7, p = 0.096 |
| Right Hemisphere |
| AF vs. AR, t(1,21)= 5.6, p < 0.001 |
| AF vs. ACC, t(1,21)= 15.9, p < 0.001 |
| AR vs. ACC, t(1,21)= 4.5, p < 0.001 |
| Left vs Right tracts |
| AF, t(1,21)= -2.7, p = 0.014 |
| AR, t(1,21)= -3.7, p = 0.002 |
| b. Perisylvian regions |
|---|
| age : F (1,19)= 51.5, p < 0.001 |
| region : F (5,95)= 101.5, p < 0.001 |
| age x hemisphere : F (1,19) <1, p = 0.379 |
| region x hemisphere : F (5,95)= 7.5, p < 0.001 |
| age x region x hemisphere : F (5,95) = 1.4, p = 0.233 |
| Posthoc comparisons |
| Left Hemisphere |
| IFG vs. MTG, t(1,20)= 2.3, p = 0.031 |
| MTG vs. STG, t(1,20)= 3.6, p = 0.002 |
| STG vs. AG, t(1,20)= 2.3, p = 0.033 |
| AG vs. PT, t(1,20)= 3.7, p = 0.002 |
| PT vs. HG, t(1,20)= 3.2, p = 0.005 |
| Right Hemisphere |
| MTG vs. IFG, t(1,20)= 2.2, p = 0.050 |
| MTG vs. AG, t(1,20) = 1.5, p = 0.182 |
| AG vs. IFG, t(1,20) = 1.2, p = 0.265 |
| IFG vs. STG, t(1,20)= 3.7, p = 0.002 |
| AG vs. STG, t(1,20)= 6.0, p < 0.001 |
| STG vs. PT, t(1,20)= 7.0, p < 0.001 |
| STG vs. HG, t(1,20)= 6.6, p < 0.001 |
| PT vs. HG, t(1,20) <1, p = 0.668 |
| Left vs Right regions |
| IFG, t(1,20)= 3.9, p = 0.004 |
| MTG, t(1,20)= -2.1, p = 0.067 |
| STG, t(1,20)= 2.5, p = 0.044 |
| AG, t(1,20)= -4.6, p < 0.001 |
| PT, t(1,20) = 1.3, p = 0.252 |
| HG, t(1,20) <1, p = 0.461 |
Transverse and parallel diffusivities were considered as the dependent variables in the ANCOVA models, with age, tract/region, hemisphere (left/right) entered as independent variables (1 covariate and 2 within-subject factors for the models of bilateral tracts and perisylvian regions). An additional ANCOVA is presented for combined left and right tracts (i.e. averaged transverse diffusivity over left and right acoustic radiations AR and arcuate fascicles AF) to include auditory fibers of the corpus callosum ACC without the hemisphere factor.
We further selected post hoc analyses and performed comparisons between selected pairs of tracts/regions using paired t-tests: within each hemisphere (c = 3 for tracts, c = 15 for regions), and across hemispheres for homologous tracts/regions (c = 2 for tracts, c = 6 for regions). Abbreviations used for the tracts/regions names are identical to those of Fig. 3. P-values were corrected using FDR approach for the number of comparisons within each hemisphere condition. For the sake of simplicity, we present only the pairs of regions with the least significant results (all the other comparisons show more significance and can be derived from these pairs). Significant (p < 0.05) comparisons after FDR correction are marked in bold, while those significant comparisons before correction that did not reach the significance threshold of p < 0.05 after correction, are highlighted in italic bold. For the comparisons within each hemisphere, the pairs of tracts/regions were ordered to highlight the gradient of λ┴ / λ// values (from higher to lower ones).