Hi,
If you search the email list for your error message you will see Matthew Webster's previous posts about this which indicate that it reflects either a problem with your setup or a result which is highly unlikely to give any significant voxels.
All the best,
Mark
Alejandro Cosa Liñán <[log in to unmask]> wrote:
It still don't work. I meant another conditions that was evaluated worked fine but this not (for this reason, incorrect data preprocessing was discarded). I checked other thresholding options (with flags -c and -C) and it worked correctly. The issue had to be related with the Threshold-Free Cluster Enhancement but I haven't found the reason yet.
Alejandro
2013/7/12 Rosalia Dacosta Aguayo <[log in to unmask]<mailto:[log in to unmask]>>
I am glad that it is finally working. Although I am not sure about the design... Jeanette Munford or Mark Jenkinson could give you a better advice. I mean that if it is statistically correct to split these time points....depending on what is your question.
Best,
Rosalia.
El 12/07/2013 00:23, "Alejandro Cosa Liñán" <[log in to unmask]<mailto:[log in to unmask]>> escribió:
Briefly, we are evaluating the effect of alcoholism in a model animal.
Timepoint 1: 18 subjects before alcohol intake
Timepoint 2: 18 subjects after alcohol intake
Timepoint 3: 9 of the 18 subjects treated + 9 of the 18 subjects without any treatment.
Currently, I am checking the differences induced in the third timepoint, Treatment vs Placebo. In the remainder of conditions evaluated (Timepoint 1 vs Timepoint 2 and Timepoint 2 vs Timepoint 3), randomise works correctly (without warnings).
Alejandro
2013/7/12 Rosalia Dacosta Aguayo <[log in to unmask]<mailto:[log in to unmask]>>
Three points of treatment? But in your design you are not checking this....
Rosalia.
El 11/07/2013 23:53, "Alejandro Cosa Liñán" <[log in to unmask]<mailto:[log in to unmask]>> escribió:
Yes, I am using a skeleton derived from all the subjects (actually, it is derived from 18 subjects acquired in three timepoints of the treatment) as a mask.
Cheers,
Alejandro Cosa Liñán
2013/7/11 Rosalia Dacosta Aguayo <[log in to unmask]<mailto:[log in to unmask]>>
Umm, ok...
Only to be sure....are you using a group mask?
Best,
Rosalia.
El 11/07/2013 23:28, "Alejandro Cosa Liñán" <[log in to unmask]<mailto:[log in to unmask]>> escribió:
Hi Rosalia,
Thanks for your advice.
The command line was an example. I forgot to include the -m flag in this example but it was included in other tries and it didn't work.
Best regards,
Alex
2013/7/11 Rosalia Dacosta Aguayo <[log in to unmask]<mailto:[log in to unmask]>>
Hi Alejandro, as Gwenaelle says, the design looks fine for me ton. But, if you are using a mask, as far as I knox you should use the -m flag in the randomise comand.
Best,
Rosalia.
El 11/07/2013 20:50, "Alejandro Cosa Liñán" <[log in to unmask]<mailto:[log in to unmask]>> escribió:
Hi Gwenaëlle,
I tried to use a mask (binarized skeleton of mean FA map ) and the same warning was thrown. Also, I tried to use an incorrect design matrix (5 subjects in group A and 13 in B) for check the data and it worked! Unfortunately it doesn't work with the correct design.
Thanks for your help!
Alex
El 11/07/2013 20:40, "Gwenaëlle DOUAUD" <[log in to unmask]<mailto:[log in to unmask]>> escribió:
Hi Alejandro,
I am not sure what this error message means and I suspect Matt (Webster) would have a much better clue than me, but in the mean time, I just would like to double-check: have you used a mask in your randomise command? The rest looks ok to me.
Cheers,
Gwenaëlle
------------------------------------------------------------------------------------
Gwenaëlle Douaud, PhD
MRC Career Development Fellow
FMRIB Centre, University of Oxford
John Radcliffe Hospital, Headington
OX3 9DU Oxford UK
Tel: +44 (0) 1865 222 523<tel:%2B44%20%280%29%201865%20222%20523>
Fax: +44 (0) 1865 222 717<tel:%2B44%20%280%29%201865%20222%20717>
www.fmrib.ox.ac.uk/team/principal-investigators/gwenaelle-douaud
-----------------------------------------------------------------------------------
________________________________
De : Alejandro Cosa Liñán <[log in to unmask]<mailto:[log in to unmask]>>
À : [log in to unmask]<mailto:[log in to unmask]>
Envoyé le : Jeudi 11 juillet 2013 19h31
Objet : Re: [FSL] TBSS - Randomise
Thanks for the suggestion.
-d refers to the next argument, design matrix file. Is that correct?
Cheers,
Alex
El 11/07/2013 20:27, "Rosalia Dacosta Aguayo" <[log in to unmask]<mailto:[log in to unmask]>> escribió:
Hi,
It is just a suggestion, but have you tried without the -d flag?
Bert regards,
Rosalia.
El 11/07/2013 20:19, "Alejandro" <[log in to unmask]<mailto:[log in to unmask]>> escribió:
Dear all,
I am trying to perform an statistical analysis from FA Maps of two groups with 9 subjects each one using randomise function.
randomise is throwing an error at me after stalling at permutation 2. "Warning: tfce has detected a large number of integral steps. This operation may require a great deal of time to complete.
See attached more information.
Any suggestion about the issue?
Thanks in advance,
Alex
The call to the function is:
randomise -i <input> -o <output> -d <design.mat> -t <design.con> -n 5000 --T2 -V
The design matrix is:
/NumWaves 2
/NumPoints 18
/PPheights 1.000000e+00 1.000000e+00
/Matrix
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 1.000000e+00
and contrast
/ContrastName1 group A > group B
/ContrastName2 group B > group A
/NumWaves 2
/NumContrasts 2
/PPheights 1.000000e+00 1.000000e+00
/RequiredEffect 2.801 2.801
/Matrix
1.000000e+00 -1.000000e+00
-1.000000e+00 1.000000e+00
