Hi,
Yes, this appears to be a bug.
In the meantime, running it on the cut-down data will give you results
which will be almost identical to what the full spectra would be.
All the best,
Mark
On 24 Sep 2008, at 05:34, Xinian Zuo wrote:
> Hi, Mark and Christian
>
> Just coming back for reporting a thing during my testing on the
> codes in
> fslpspec. I am still wandering there were some bugs in the codes
> regarding
> the test. The test is described below.
>
> I have a 4D bold data with 197 vols. The previous reported error was
> met
> when this data computed by using
>
> fslpspec bold-4d.nii.gz bold-4d-ps.nii.gz.
>
> Further, I used the same command to calc the psd of the bold data
> with less
> vols which is achieved by using command below:
>
> fslroi bold-4d.nii.gz bold-4d196.nii.gz 0 196
>
> Then the fslpspec works on the vols-reduced data (The speed of this
> command
> is really improved than that before!):
>
> fslpspec bold-4d196.nii.gz bold-4d-ps.nii.gz.
>
> I am confused by that. I read the source codes in fslpspec and found
> potential problems in following lines (line 82, 'Mat.Nrows()/2')
> when the
> number of vols is odd.
>
> 80 ReturnMatrix calcFFT(const Matrix& Mat)
> 81 {
> 82 Matrix res(Mat.Nrows()/2,Mat.Ncols()), empty(1,1);
> 83 empty=0;
> 84 ColumnVector tmpCol;
>
> In a word, there could be an error relevant to an incompatible
> dimension of
> processing 4d bold-data with odd vols. The even sounds ok. Hope a
> reply
> about this, thanks!
>
> All the best
> Xinian
>
> On Fri, 19 Sep 2008 08:55:51 +0100, Mark Jenkinson <[log in to unmask]
> > wrote:
>
>> Hi,
>>
>> It looks to me like your syntax is wrong too.
>> The version of fslpsec here does not have a -m option.
>> In fact it doesn't an option for a mask at all.
>>
>> So you should try:
>> fslpspec 4d-bold.nii.gz 4d-bold-ps.nii.gz
>> and then if you want to mask it, do it afterwards with fslmaths.
>>
>> As for you compiling problems, I am not familiar enough with
>> the altix compilers to know how to deal with your specific errors.
>> Maybe someone else on the list can help. Here are some sample
>> errors just in case:
>>
>> standard input}: Assembler messages:
>> {standard input}:2257: Warning: Use of 'ld8' may violate WAW
>> dependency 'GR%, % in 1 - 127' (impliedf), specific resource number
>> is
>> 77
>> {standard input}:2257: Warning: Only the first path encountering the
>> conflict is reported
>> {standard input}:2254: Warning: This is the location of the
>> conflicting usage
>>
>>
>> /tmp/ccDniEw0.s: Assembler messages:
>> /tmp/ccDniEw0.s:1242: Warning: Additional NOP may be necessary to
>> workaround Itanium processor A/B step errata
>> /tmp/ccDniEw0.s:1551: Warning: Additional NOP may be necessary to
>> workaround Itanium processor A/B step errata
>>
>>
>> /tmp/ccjI4bFi.s: Assembler messages:
>> /tmp/ccjI4bFi.s:6084: Warning: Use of 'mov' may violate WAW
>> dependency
>> 'GR%, % in 1 - 127' (impliedf), specific resource number is 14
>> /tmp/ccjI4bFi.s:6083: Warning: This is the location of the
>> conflicting
>> usage
>>
>> As these also occur in third-party code (not FSL code) then I think
>> it
>> is a more
>> general problem that needs to be fixed with your compiler and
>> possibly
>> the options you are using for it.
>>
>> All the best,
>> Mark
>>
>>
>>
>>
>>
>> On 19 Sep 2008, at 08:18, Christian F. Beckmann wrote:
>>
>>> Hi Xinian,
>>>
>>> did you check that the original 4d data and the mask have the same
>>> dimensions? What do you get from fslsize on these two files?
>>> ta
>>> Christian
>>>
>>> _______________________________________________
>>> Christian F. Beckmann, DPhil
>>>
>>> Senior Lecturer, Clinical Neuroscience Department
>>> Division of Neuroscience and Mental Health
>>> Imperial College London
>>> Hammersmith Hospital - London W12 0NN
>>> Tel.: +44 (0) 208 383 3722 --- Fax: +44 (0) 208 383 2029
>>> Email: [log in to unmask]
>>> http://www.imperial.ac.uk/medicine/people/c.beckmann/
>>>
>>> Senior Research Fellow, FMRIB Centre
>>> University of Oxford
>>> JR Hospital - Oxford OX3 9DU
>>> Tel.: +44 (0) 1865 222551 --- Fax: +44 (0) 1865 222717
>>> Email: [log in to unmask]
>>> http://www.fmrib.ox.ac.uk/~beckmann
>>>
>>>
>>>
>>>
>>> On 19 Sep 2008, at 02:37, Xinian Zuo wrote:
>>>
>>>> Hi, Mark
>>>>
>>>> Thanks for the response and zealous helps at all times. I am
>>>> using a
>>>> pre-compiled version of FSL (CentOS4.0) under RedHat-9.0 and
>>>> employing
>>>> fslpspec cmd for computation of the PSD of a 4D BOLD fMRI data with
>>>> around
>>>> 190 time points. I retested this using another pre-compiled vesrion
>>>> (CentOS5.0) under RedHat-AS 3.4 and got the same matter. As you
>>>> suggested,
>>>> it would be a good choice to compile it from source release. The
>>>> cmd I used
>>>> like below,
>>>> #---------------------------------------
>>>> fslpspec 4d-bold.nii.gz -m example_func_brain_mask.nii.gz 4d-bold-
>>>> ps.nii.gz
>>>> #---------------------------------------
>>>> and the output is only a message written in previous post. If you
>>>> need
>>>> further information or data transfer for a diagnostic usage, please
>>>> let me know.
>>>>
>>>> btw: I compiled source under altix platform using intel C++ 9.0 and
>>>> failed.
>>>> I am wandering all this may be related with the specific
>>>> architecture of our
>>>> machine...
>>>>
>>>> All the best
>>>> Xinian
>>>>
>>>> On Thu, 18 Sep 2008 11:56:44 +0100, Mark Jenkinson <[log in to unmask]
>>>>> wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I cannot replicate this here.
>>>>>
>>>>> What exactly are you doing?
>>>>> And is this a pre-compiled version of FSL or not?
>>>>> Probably best to send us the result of running
>>>>> fslerrorreport (on the input file).
>>>>>
>>>>> All the best,
>>>>> Mark
>>>>>
>>>>>
>>>>>
>>>>> Xinian Zuo wrote:
>>>>>> Hi, folks
>>>>>>
>>>>>> Is there anyone to test the work of fslpspec in 4.1 version
>>>>>> right? I just
>>>>>> let it calculate a 4D fmri data and got an error below:
>>>>>>
>>>>>> terminate called after throwing an instance of
>>>>>> 'NEWMAT::IncompatibleDimensionsException'
>>>>>>
>>>>>> and exit...
>>>>>>
>>>>>> btw: I displayed my bold timeseries by using fslview and checked
>>>>>> potential
>>>>>> data flaw, but they looks good.
>>>>>>
>>>>>> Is this a bug? Hope any words about this.
>>>>>>
>>>>>> Thanks
>>>>>> Xinian
>>>>>>
>>>>>>
>>>
>
>
>
|