Thank you Anderson for your replies. I think I understand what you mean.

Sincerely,

Virendra

From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Anderson M. Winkler [[log in to unmask]]
Sent: Wednesday, April 10, 2013 1:13 AM
To: [log in to unmask]
Subject: Re: [FSL] cluster-based thresholding v/s tfce

Hi Virendra,

Are you sure the files you want are called tbss_maxc_*? You can have tbss_corrp_clustere_*, which is precisely what you get with the -c option as you are using, and I think the option -N outputs the maximum statistic (here cluster size then) per permutation in a text file. So this isn't an image then, and cannot possibly ever be the same as the image results from TFCE.

If you are interested in the threshold that corrects the image with the voxels marked with the cluster size of the cluster that they belong to, then use the numbers that are printed in the screen at the end of randomise. I'm not sure if this was called "maxc" in some earlier version, but I don't think it's used like that anymore.

And in case you want the largest cluster of your results, check the outputs of the command "cluster".

All the best,

Anderson



2013/4/10 Virendra Mishra <[log in to unmask]>
Hi Anderson,

Many thanks again. I am really interested in tbss_maxc_* result. How do I get tbss_maxc_* results then using randomise?

Thanks.

Sincerely,

Virendra

From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Anderson M. Winkler [[log in to unmask]]
Sent: Tuesday, April 09, 2013 8:08 PM

To: [log in to unmask]
Subject: Re: [FSL] cluster-based thresholding v/s tfce

Hi Virendra,
These are different things and won't be equal.
Anderson

Am Mittwoch, 10. April 2013 schrieb Virendra Mishra :
Thank you for your quick reply. What I really meant to type was without the --T2 option and with -c 3 option whether the results of tbss_tfce_corrp_tstat_* will be equal to tbss_maxc_tstat_* ?

Thanks again for your time.

Sincerely,

Virendra

From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Anderson M. Winkler [[log in to unmask]]
Sent: Tuesday, April 09, 2013 7:09 PM
To: [log in to unmask]
Subject: Re: [FSL] cluster-based thresholding v/s tfce

Hi Virendra,

No no, they are different things and adding extra options will give you additional results, but one result doesn't interfere with another.
So, having --T2 and -c 3 in the same randomise call will give produce TFCE and cluster extent results, each on their own right.

All the best,

Anderson


Am Mittwoch, 10. April 2013 schrieb Virendra R Mishra :
Hello FSL users,

I have a quick question on cluster-based thresholding v/s tfce in randomise.

If I use "randomise -i all_FA_skeletonised -o tbss -m mean_FA_skeleton_mask -d design.mat -t design.con -n 500 --T2 -V", I understand that the output is tfce based but if I use "randomise -i all_FA_skeletonised -o tbss -m mean_FA_skeleton_mask -d design.mat -t design.con -n 500 --T2 -V -c 3" then the results are cluster based, right?

What I mean to ask is, whether tbss_tfce_corrp_tstat1 and tbss_tfce_corrp_tstat2 are equivalent to tbss_maxc_tstat1 and tbss_maxc_tstat2 if I use " -c 3" option?

Thank you for your help.

Sincerely,

Virendra



UT Southwestern Medical Center
The future of medicine, today.