Dear Anderson,
Just as you said, after checking these images I found that they weren't good, the converter from DICOM didn't work correctly. I've switched to another one and everythings seem ok now. Thanks alots for your advices.
Warm regards,
Kim
On Thu, 5 Mar 2015 09:33:52 +0000, Anderson M. Winkler <[log in to unmask]> wrote:
>Hi Kim,
>
>Could you check (visually or with fslmaths) that one DWI sequence isn't
>producing systematically lower FA values than the other?
>You can also check the result of the _vox_ results (i.e., not TFCE) and see
>if it's all positive or all negative for one of the contrasts.
>
>Another thing, unrelated, is that for a paired t-test in randomise, you
>must define one exchangeability block per subject, and use the option "-e
>design.grp", so that the shufflings will happen only within subject. This
>needs to be done regardless, although it won't affect these unusual results
>you're observing.
>
>All the best,
>
>Anderson
>
>
>On 5 March 2015 at 09:24, Kim TRAN <[log in to unmask]> wrote:
>
>> Hi FSL users,
>>
>> I have a group of 10 subjects who was scanned with 2 DTI different, and I
>> want to find the significative different voxels between these 2 groups of
>> DTI. So I ran randomise with the TFCE option:
>>
>> randomise -i all_FA_skeletonised -o tbss -m mean_FA_skeleton_mask -d
>> design.mat -t design.con -n 500 --T2
>>
>> But I got a following message in the command line:
>>
>> Critical Value for: tbss_tfce_corrp_tstat1 is: 174558
>> 184756 permutations required for exhaustive test of t-test 2
>> Doing 500 random permutations
>> Starting permutation 1 (Unpermuted data)
>> Warning: The unpermuted statistic image for the current image contains no
>> positive values, and cannot be processed with TFCE. A blank output image
>> will be created.
>> Critical Value for: tbss_tfce_corrp_tstat2 is: 0
>>
>> Can you explain to me the error messages?
>>
>> Thanks in advance,
>>
>> Kim
>>
>>
>
|