Hello FSL users
I have processed my data using tbss_2_reg -t & -n options. All things being
equal the statistics come out dramatically different. I have a control
(n=14) and patient group (n=15) and both are men aged 20-49. My questions are:
1. Is it typical to expect large differences in the statistical results
(e.g., increased and decreased FA in very different locations) when the -t
vs. -n options are used especially in men (who I would predict would match
well to the MNI template)?
2. Seeing as the results are so different which set of data do I trust? I am
assuming in this case that I would trust the data generated with the -n
option as this should be most representative of my sample.
3. As a side, is it appropriate to trust the atlas labels in the atlas
toolbox if your sample has not been normalized to that atlas (e.g., JHU
labels when the default MNI template was used as the target)?
Cheers,
Deryk
|