Table 1.
Network-based statistics results
| Group t-test | PreHD < Controls | Cont < PreHD |
|---|---|---|
| Functional (114 ROIs) | 0.672 | 0.6046 |
| Structural (114 ROIsr) | 0.3447 | 0.2384 |
| Functional (structural constrained) 114 ROIsr | 0.6072 | 0.5792 |
| Functional (514 ROIsr) | 0.4296 | 0.5976 |
| Structural (514 ROIsr) | 0.1504 | 0.4613 |
| Functional (structural constrained) 514 ROIsr | 0.2603 | 0.7998 |
| Correlations | Positive | Negative |
| NFL–fMRI correlation (whole group) 114 ROIsr | 0.0304* | 1 |
| NFL–fMRI correlation (preHD) 114 ROIsr | 0.019* | 0.6633 |
| NFL–structural correlation (whole group) 114 ROIsr | 0.191 | 0.391 |
| NFL–structural correlation (preHD) 114 ROIsr | 0.2523 | 0.1818 |
| NFL–fMRI correlation (whole group) 514 ROIsr | 0.0398* | 0.6615 |
| NFL–fMRI correlation (preHD) 514 ROIsr | 0.027* | 0.6478 |
| NFL–structural correlation (whole group) 514 ROIsr | 0.3177 | 0.0284* |
| NFL–structural correlation (preHD) 514 ROIsr | 0.2919 | 0.023* |
For these analyses, permutation testing using unpaired t-tests and 5000 permutations was performed on a general linear model that included age and sex as covariates. A test statistic was then computed for each connection and a default threshold applied (t = 3.1) to produce a set of suprathreshold connections that displayed significant between-group connectivity differences. *FWE-correction was applied at P = 0.05. Due to the high false positive rates in fMRI connectivity analyses73 the functional connectivity analysis was repeated, constraining the functional connectome by the structural. Here, the functional matrix was simply multiplied by the structural matrix to remove any functional connections that do not have supporting structural connections, and the NBS analysis repeated. ROIs = regions of interest.