Table 6.
Overall comparison of FEAT and dHCP pipelines using t-statistics from the GLM output and significantly activated voxel counts for all 15 subjects. T-statistics: using the dHCP pipeline, all grey matter ROIs showed an increase in mean t-statistic (maximum t-statistic in parentheses). This increase in t-statistics was revealed to be statistically significant (voxel-based thresholding, 5% FWER corrected) using the constrained-region analysis, localised to the PoCG. # active voxels: the dHCP pipeline also resulted in an increase in the number of significantly active voxels using all three thresholding approaches. % white matter: voxel-based thresholding was not valid due to zero voxels being activated. Cluster-based thresholding showed a clear reduction in mislocalised activity using the dHCP pipeline. TFCE-based thresholding showed a very modest increase in mislocalised activity using the dHCP pipeline, possibly due to unavoidable ‘enhancement’ of white matter voxels by the neighbouring grey matter. Voxel is FDR corrected; cluster and TFCE are FWER corrected. ACC = anterior cingulate cortex; PoCG = postcentral gyrus; TFCE = threshold free cluster enhancement; * = statistically significant difference.
| FEAT | dHCP | ||
|---|---|---|---|
| T-statistics | ACC | 1.529 (3.518) | 2.104 (4.030) |
| Insula | 2.651 (5.244) | 2.977 (5.224) | |
| PoCG | 2.822 (6.568) | 2.959* (6.987) | |
| Thalamus | 2.168 (6.197) | 2.565 (6.797) | |
| # active voxels | Voxel | 0 | 66,984 |
| Cluster | 60,175 | 82,837 | |
| TFCE | 16,823 | 66,756 | |
| % white matter | Voxel | – | 17.598 |
| Cluster | 10.051 | 5.400 | |
| TFCE | 4.412 | 4.687 | |