Skip to main content
. Author manuscript; available in PMC: 2022 Jun 11.
Published in final edited form as: Brain Connect. 2021 Aug 23;12(5):402–416. doi: 10.1089/brain.2021.0058

Table 2.

Validation of whole-tract FA correlations, corresponding to Figure 7. Columns show the parameters of the best-fit line y = mx + c and the correlation coefficient, r, between tract FA values from subject-specific fibre bundles and those from each atlas, measured using a repeated measures correlation (Bland and Altman, 1995). Also shown is the difference between the z-transform of the correlation coefficients for the age-specific atlas and the JHU atlas, and the 95% confidence intervals (CI) for this difference. Positive differences indicate a higher correlation with the age-specific atlas. These are shown for the same-site validation data, the HBN data with FOD-based tractography, the HBN data with tensor-based tractography, and the HBN data with minimal processing and tensor-based tractography. *p < 10-10. HBN = Healthy Brain Network; FOD = fibre orientation distribution; JHU = Johns Hopkins University.

Age-specific Atlas JHU Atlas Difference between z-transformed correlation coefficients (95% CI)
Dataset m c r m c r
Same-site 0.88 0.13 0.715* 0.57 0.22 0.536* +=0.298 (+0.115, +0.300)
HBN (FOD) 0.84 0.15 0.781* 0.59 0.25 0.617* +0.328 (+0.231, +0.412)
HBN (Tensor) 0.51 0.27 0.697* 0.39 0.32 0.595* +0.176 (+0.087, +0.281)
HBN (Minimal) 0.72 0.17 0.661* 0.55 0.22 0.563* +0.158 (+0.031, +0.268)