Yes - it sounds like one of the upsampled inputs to higher-leve
(stored in reg_standard within the first-level FEAT directories) were
corrupted - possibly because you ran out of disk space during the
original upsampling. Glad you got it sorted.
Cheers, Steve.
On 29 Apr 2006, at 07:11, Nicholas P Holmes wrote:
> thanks Mark,
>
> so, i have learnt a little more about the feat output, but i
> couldn't see
> anything obviously wrong with the dof/tdof/var/cope/pe/neff(?)
> files either at
> second level or at the lower level. overlaying these stats images
> onto the final
> z-stat images showed no relation between the very high z-scores in
> the second
> level cope and the other stats. images
>
> however, i repeated the second level analysis a few times, trying
> different
> things, like freeing up space or increasing virtual memory, but the
> only thing
> that seemed to work was removing the 'reg_standard' directories
> from the
> lower-level feats, and/or turning on the 'clean up standard space'
> option.
> Alternatively, it could just have been 7th time lucky...
>
> could that make any sense?
>
> thanks,
> nick
>
> on 28/4/06 11:19 AM, FSL - FMRIB's Software Library
> <[log in to unmask]> wrote:
>
>> Nick,
>>
>> First thing to check is if the DOF information that is being passed
>> into flame is OK. Take a look at tdof_filtered_func_data and check
>> for zero or negative values in brain voxels (particularly those which
>> ended up with odd z-stats in the output). Similarly, check that
>> var_filtered_func_data has sensible looking (positive) values.
>>
>> Cheers, Mark.
>>
>> Dr Mark Woolrich.
>>
>> Oxford University Centre for Functional MRI of the Brain (FMRIB),
>> John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK.
>>
>> Tel: (+44)1865-222782 Homepage: http://www.fmrib.ox.ac.uk/~woolrich
>>
>> On 28 Apr 2006, at 08:04, Nicholas P Holmes wrote:
>>
>>>> Dear Imagers,
>>>>
>>>> I'm having problems with a second-level Fixed Effects analysis
>>>> (combining
>>>> the two runs of each subject). Most subjects and most COPEs work
>>>> fine, but
>>>> occasionally, one of the COPEs produces Z-scores into the hundreds
>>>> and the
>>>> .png image goes all yellow-spotty. The report is coped below &
>>>> attached.
>>>> This has happened several times, but usually re-running the
>>>> analysis for
>>>> that cope 'fixed' the problem. This time it didn't get fixed after
>>>> four
>>>> repeats
>>>>
>>>> any ideas?
>>>>
>>>> system = cygwin, PC, windows XP, 1.6GHz, 712MB, with 2-3GB space
>>>> on HD,
>>>> and pictures of goldfish on my desktop.
>>>>
>>>> many thanks in advance,
>>>>
>>>> nick
>>>>
>>>> Started FEAT at Fri Apr 28 08:49:07 RDT 2006 on nick-cbo2xux66h
>>>> FEAT output is C:/cygwin/home/Nick/H00_funcs/func_06_L1_1B.gfeat/
>>>> cope8.feat
>>>> /bin/cp C:/cygwin/home/Nick/H00_funcs/func_06_L1_1B.gfeat/
>>>> design.fsf
>>>> design.fsf
>>>> /usr/local/fsl/bin/feat_model design
>>>> cat ../design.lcon | awk '{ print }' > design.lcon
>>>> /usr/local/fsl/bin/flame --cope=filtered_func_data
>>>> --vc=var_filtered_func_data --dvc=tdof_filtered_func_data --
>>>> mask=mask
>>>> --ld=stats --dm=design.mat --cs=design.grp --tc=design.con --fe
>>>> Log directory is: stats
>>>> Setting up:
>>>> No f contrasts
>>>> nevs=1
>>>> ntpts=2
>>>> ngs=1
>>>> nvoxels=264928
>>>> Running:
>>>> DOF cannot be zero or negative! [repeated26 times]
>>>> stdtr domain error [repeated 7 times]
>>>> ndtri
>>>>
>>>>
>>>>
>>
>>
>>
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|