Hi,
If you write 'program' do you mena feat or your script that midifies the
fsf-files?
There might be a general problem, if you specify regressors in the
one-column text format, because this might need the same number of time
points as given in the fsf-file. I use the three-column text format and
it is working fine. So this might be a cause of trouble.
The script I send in a previosus post assumes that there is a fsf-file
($cfl) with a number of volumes specified. It determines the current
functional data set ($dtfl) and its number of volumes (vol=`fslnvols
$dtfl`), greps the current line of volumes from the fsf-file and
replaces it with the new number if different.
The code provided by Martin Monti assumes that your fsf-files uses
'###SUB###' and '###VOL###' as place holder for the entries and
replaces these with valid values as determined bevore the sed call. So
for this option you need to write set fmri(npts) ###VOL###' in your
template fsf file.
Hope this helps a bit,,,
good luck
wolf
Courtney Haswell wrote:
> I could not get the program to determine the number of timepoints
> while executing Feat. Do you just leave the number of volumes as 0 on
> the .fsf? I always get the error for prestats that "Error -
> prefiltered_func_data has a different number of time points to that in
> FEAT setup" if I do not have the exact number of volumes selected in
> my feat file.
> I would be interested in just letting the program determine the number
> of volumes but I could not get it to work that way.
>
> On Fri, Nov 5, 2010 at 3:37 AM, Xue, Feng <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
> Dear folks,
>
> Yes, we absolutely could do in this way, however, I wonder is there
> any particular reason to set npts in fsf script? Why not let the
> program to determind it when exectuting?
>
> On Thu, Nov 4, 2010 at 11:44 PM, Courtney Haswell
> <[log in to unmask] <mailto:[log in to unmask]>> wrote:
> > Thanks everyone, I got it working. I was having issues with the
> Feat GUI
> > allowing me to enter a variable for Total Volumes but I just
> edited the .fsf
> > file in a text editor to add the variable.
> >
> > This is my script in case it will help anyone else in the future:
> > for runs in 1 2 3 4 5 6 7 8; do
> > ANATFILE=$RAWDIR/T1_brain
> > RUNDIR=$RAWDIR/run${runs}
> > OUTPUTDIR=$FSLDATADIR/run${runs}
> > BEHAVDIR=$FSLDATADIR/behavioral/run${runs}
> > VOL_TIMEPT=`ls -l $FUNCDIR/run006_0${runs}/*.img | wc -l`
> > for i in 'FEAT_template_evst.fsf'; do
> > sed -e 's@OUTPUT@'$OUTPUTDIR'@g' \
> > -e 's@ANAT@'$ANATFILE'@g' \
> > -e 's@VOLPOINTS@'$VOL_TIMEPT'@g' \
> > -e 's@RUNDIR@'$RUNDIR'@g' <$i> $FSLDATADIR/FEAT_run${runs}.fsf
> > done
> > feat $FSLDATADIR/FEAT_run${runs}.fsf
> > ##############
> > done
> >
> > On Thu, Nov 4, 2010 at 11:14 AM, Martin M Monti
> <[log in to unmask] <mailto:[log in to unmask]>>
> > wrote:
> >>
> >> Hi Courtney,
> >>
> >> if you look back in the archive there are a couple of posts on
> the topic.
> >> What you want to do is to write a 'sample" script, which you
> can do from the
> >> GUI, then open it and replace all the important variables of
> the .fsf file
> >> you want to change with something unique and easy to spot
> (say.. ###SUB###
> >> for the subject name, ###VOL###, etc). This will now be your
> "master fsf
> >> copy"
> >> Then, the following shows how to create a .fsf file for each
> subject,
> >> changing the values that you have set into a "variable" (for
> example, the
> >> SubjectName and the TRNumber)
> >>
> >> foreach sub (sub1 sub2 sub3) #subject loop
> >> set nvols = `fslnvols ${sub}_run1.*` #pull out the TRs for each sub
> >> echo $sub $nvols #tell me
> subject and TR
> >> number
> >> #Replace values and put in sub's directory
> >> sed -e "s/###SUB###/${sub}/g" \
> >> -e "s/###VOL###/${nvols}/g" \
> >> Run1_Sample.fsf > /imaging/mm03/LingMath/FSL/${sub}/run1.fsf
> >> end
> >>
> >> then you can just make a loop to launch the feat analysis from
> the shell
> >>
> >> Cheers
> >>
> >> Martin
> >>
> >> On 04/11/10 13:24, Courtney Haswell wrote:
> >>>
> >>> Hi FSL users,
> >>>
> >>> For fMRI analysis, I use a Feat template and the sed command
> to create
> >>> all of my .fsf files. I have a bit of a problem now in that
> all of the runs
> >>> and paradigms have a different number of timepoints, so
> therefore a
> >>> different number of total volumes in each run for each
> subject. Is there a
> >>> way to change the total number of volumes through scripting in
> the .fsf file
> >>> after it has been created before I run "feat" on it? It will
> not allow me to
> >>> enter a variable for "Total Volumes" only a number. I would
> prefer not to
> >>> have to create the .fsf files manually for every run of every
> subject.
> >>>
> >>> Thank you,
> >>> Courtney
> >>>
> >
> >
>
>
>
> --
> Best Regards
>
> Xue, Feng Ph.D.
> Major in Developmental Cognitive Neuroscience
>
> National Key Laboratory of Cognitive Neuroscience and Learning
> Beijing Normal University
> Beijing, China. 100875
> Tel: +86-13810154455
> web: http://psychbrain.bnu.edu.cn <http://psychbrain.bnu.edu.cn/>
> ==============================================
> Welcome to MuDuo JinSheng BBS @ Beijing Normal University
> telnet://bbs.mdjs.org <http://bbs.mdjs.org/>
> http://bbs.mdjs.org <http://bbs.mdjs.org/>
>
>
