Dear FSL experts
I want to perform multiple comparison to MRI data (such as VBM) in a voxel-wise manner using the 'randomise' option
implemented in FSL.
I think that permutation method is based on a non-parametric statistics and that I do not need to smoothing those MRI
brain images before performing multiple comparison in the TFCE permutation method.
I detected some statistically significant regions when adopting permutation multiple comparison method to
those whole brain MRI data after smoothing. However, there are no statistically significant regions when performing permutation multiple comparison to those MRI data without smoothing process.
In this case, how should I interpret this result?
I think that the result derived from the data without smoothing process (that is, the fact that there is no statistically significant regions) is more appropriate.
Am I correct?
Any comments and helps are greatly appreciated.
Thank you very much.
Takuya
|