Dear Arash,
The problem is with this design.
Since the origin of the coordinate system is arbitrary and we are
only looking for variations around the mean shape, the design
matrices used in FIRST cannot model the mean. That is, they
cannot include a constant column or any combination of columns
that would add up to a constant column. If you add your two
columns together you get just such a constant column. So
if you just *remove the second column* your design will be fine
and will test the difference between the two groups (it will be
demeaned internally, and so be like a column of +1's and -1's).
I'm not sure about how you've installed your FSL version.
It doesn't sound correct to me. You should definitely not need
a prefix to every command. I would sort that out first before
trying anything else. If you have installed the new FSL in a
different place then you should update your FSLDIR in your
setup files (normally .profile or .bash_profile in your home
directory). Check to see if they are correct (using a text editor)
and fix them if not. If that isn't it and you are still stuck then
please give us some more information about your setup and
we will try to help.
All the best,
Mark
On 9 Jun 2011, at 14:15, Arash Nazeri wrote:
> Dear Mark,
>
> I have 2 groups, 11 subjects in each. This the design matrix that I used lately. I created it with design_ttest2 DES 11 11.
>
> /NumWaves 2
> /NumPoints 22
> /PPheights 1 1
> /Matrix
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
>
>
> First off, I concatenated all L_Hipp.bvars with
>
> concat_bvars all.bvars *L_Hipp*.bvars
>
> and then I used the following command for vertex analysis:
>
> first_utils --usebvars --vertexAnalysis -i all.bvars -o diff_c -d DES.mat --useReconMNI
>
> After I updated the FSL now I need to put fsl4.1- as as prefix to every command I want to use. Does this have anything to do with fsl installation?
>
> Thank you very much,
> Arash
>
>
> ________________________________________
> From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Mark Jenkinson [[log in to unmask]]
> Sent: Thursday, June 09, 2011 4:05 AM
> To: [log in to unmask]
> Subject: Re: [FSL] Problem with FIRST vetex analysis
>
> Dear Arash,
>
> The conversion of the individual vtk files with the command you use
> is absolutely fine and nothing to worry about. It does not affect the
> accuracy or reliability of the results in any way, so do not worry about that.
> It is purely a viewing issue.
>
> The abort of your vertex analysis is a concern though. It is possible that
> your matrix is badly formed. Can you please include a copy of the
> design matrix and some more details on the analysis (number of subjects,
> exact command you ran, etc.). With this information we should be able
> to figure out what is wrong.
>
> All the best,
> Mark
>
>
> On 9 Jun 2011, at 04:12, Arash Nazeri wrote:
>
>> Dear all,
>>
>> We've been trying to run first_utils on our run_first_all results. But no matter what flag or GLM (either with design_ttest2 or GLM) we use, we end up with the following error:
>> .
>> .
>> .
>> set shape Mesh
>> 1 0 0
>> 0 0 1
>> do mvglm
>> terminate called after throwing an instance of 'NEWMAT::SingularException'
>> Aborted
>>
>> Nevertheless, first_utils works perfectly fine with fsl course data. Also, fslview crashes while opening the original *.vtk files generated by our own run_first_all, unless we reformat the vtk files using the following command:
>> run_mesh_utils --doConvert_Binary_To_ASCII -m original_file.vtk -o new_file.vtk
>>
>> Can anyone help us solve this issue?
>>
>> Cheers,
>> Arash
>>
>>
>> ______________________________________________________________________
>> This email has been scanned by the CAMH Email Security System.
>> ______________________________________________________________________
>>
>
> ______________________________________________________________________
> This email has been scanned by the CAMH Email Security System.
> ______________________________________________________________________
>
>
>
> ______________________________________________________________________
> This email has been scanned by the CAMH Email Security System.
> ______________________________________________________________________
>
|