Skip to main content
. 2019 Feb 1;186:286–300. doi: 10.1016/j.neuroimage.2018.11.006

Table 4.

Comparison of the effects of spatial smoothing on t-statistics from the GLM output and significantly activated voxel counts for all 15 subjects. T-statistics: as spatial smoothing extent increased, peak t-statistics (in parentheses) decreased and statistically significant (voxel-based thresholding, 5% FWER corrected) increases in t-statistics (* and † symbols) were observed across all grey matter ROIs. The mean t-statistic in all ROI increased after 3 mm smoothing compared to no smoothing, and increased in PoCG and thalamus after 5 mm smoothing compared to 3 mm # active voxels and % white matter: as smoothing extent increased, the number of statistically significant voxels increased, measured using three thresholding approaches. Voxel is FDR corrected; cluster and TFCE are FWER corrected. ACC = anterior cingulate cortex; PoCG = postcentral gyrus; TFCE = threshold free cluster enhancement; * and † = statistically significant difference.

0 mm FWHM 3 mm FWHM 5 mm FWHM
T-statistics ACC 1.956 (5.392) 2.104* (4.030) 2.066 (3.102)
Insula 2.674 (6.222) 2.977* (5.224) 2.968 (4.317)
PoCG 2.423 (8.778) 2.959* (6.987) 3.265 (5.874)
Thalamus 1.928 (7.467) 2.565* (6.797) 2.927 (5.389)



# active voxels Voxel 14,484 66,984 140,389
Cluster 48,332 82,837 109, 137
TFCE 25,226 66,756 93,852



% white matter Voxel 13,353 17.598 21.309
Cluster 4.616 5.400 7.192
TFCE 2.319 4.687 6.148