Dear SPMers:
I'm a beginner of VBM analysis and thanks for this forum that provided so many useful information for learning and dosing such analysis.
I had read some explanations for the meaning and difference of set-level, cluster-level and voxel-level.
Since our current study regarding the relationship between performance in Wisconsin Card Sorting Test and grey matter volume in healthy subjects was more exploratory one, I planed to use more liberal statistic threshold as voxel-level: uncorrected P<0.001 with some cluster-level corrected for multiple comparisons.
However, the result of my initial analysis revealed since some significant voxels were found, all cluster level results showed non-significant findings, even not corrected for multiple comparisons. I wonder if such significant voxels were meaningless since they belonged to non-significant cluster?
Please see my result:
https://docs.google.com/open?id=0BxYceQvgtXBaZGxfMVQzUTZrQms
Thanks!
KC
|