I have also found the use of the -T or -S flags can make a big difference, this global cluster may be a side effect of how TFCE works, assuming you used that. It may even be a true result (there is a global effect in patients). However, I think you are right to be concerned. I would manually look at each participant's data to make sure it looks OK in the postereg output. Maybe something went wrong in a few participants and the data came out wonky.
It is never a bad idea to manually go through the registered data for all participants. Sorry I forget the specific file name, it's been a while since I ran TBSS. But they exist in a single 4D file.
Best of luck.
Colin Hawco, PhD
Neuranalysis Consulting
Neuroimaging analysis and consultation
www.neuranalysis.com
[log in to unmask]
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of John anderson
Sent: November-05-16 5:25 PM
To: [log in to unmask]
Subject: [FSL] TBSS global effect
Dear FSL experts,
I am working on a TBSS analysis (40 patients and 20 healthy controls).
I followed the steps included in FSL wiki
For "TBSS_3_postreg" when I use the flag "-T" I get results of multiple significant clusters.
When I use "TBSS_3_postreg" with the flag "-S" including a mean template built from all the subjects. I get global effect (one big significant cluster at a very small p value). The global effect is exist if I permute 5000 or 500.
The mean template was created by registering all the FA maps for all the subjects to the best FA map that I have then using the command :
fslmaths image1 -add image2 -add image3 -add ..... -div N template -odt float
In this case which flag (-n or -T) I must use with the script "TBSS_3_postreg "?
I tried to investigate global effect by dividing the images into two groups depending on age and the global effect persists when I used the flag -S with the script TBSS_3_postreg
I highly appreciate any advice
Jon
|