On Tue, Feb 10, 2009 at 09:43:46AM +0100, Gael Varoquaux wrote:
> Hi Christian,

> Thanks for your answer. I have run a group ICA. I have no
> 'filtered_func_data' obviously around in the .ica directory
> (groupmelodic.ica, under a .gica directory). I guess I should give it a
> list of paths to the individual filtered_func_data nifit files, right ?

Actually, looking at the log file of the commands ran, I discorvered that
their was a '.filelist' hidden file with the absolute path to the input
files. So I ran:

melodic -i .filelist --ICs=groupmelodic.ica/melodic_IC --mix=groupmelodic.ica/melodic_mix --report --Ostats -o tnew_threshold.ica

Melodic is now running. I'll see what comes out.

Thanks for your help,

Gaël

> On Mon, Feb 09, 2009 at 07:13:28PM +0000, Christian F. Beckmann wrote:
> > Hi

> > Yes, there is - if you select the 'output full stats folder' then melodic 
> > will save both the unthresholded probability maps and the thresholded 
> > Z-stats images in the /stats/ subdirectory. If you failed to select that 
> > option then you can get melodic to re-threshold using the following within 
> > the existing .ica directory

> > melodic -i filtered_func_data --ICs=melodic_IC --mix=melodic_mix --report 
> > --Ostats -o tnew_threshold.ica

> > This will create a new sub-dir new_thresh.ica which contains what you want
> > hth
> > Christian




> > On 9 Feb 2009, at 17:26, Gael Varoquaux wrote:

> >> Hi FSL gourous,

> >> I am trying to use the output of a melodic run. I have the
> >> melodic_IC.nii.gz giving me all the maps, but I would like to threshold
> >> them using the threshold estimated by melodic. However I can't figure out
> >> where this threshold as been stored in a format I can reload. I could
> >> (probably) scrape it from the report, but there must be a more clever way
> >> of doing things.

> >> Cheers,

> >> Gaël
-- 
    Gael Varoquaux
    Research Fellow, INRIA
    Laboratoire de Neuro-Imagerie Assistée par Ordinateur
    NeuroSpin/CEA Saclay , Bât 145, 91191 Gif-sur-Yvette France