Print

Print


Hi Julia,

What you are seeing is the internal representation when using MIGP. If you want the associated time courses you should now be running dual regression of the spatial maps. That way you will get the least-squares time courses per subject as dr_satge1 outputs.

hth
Christian

> was changed as it’s not meaningful to look at theOn 28 Oct 2015, at 19:00, Julia Schumacher <[log in to unmask]> wrote:
> 
> Hi,
> 
> I am trying to perform a group concatenation ICA in Melodic with FSL version 5.0.9.
> 
> I am concatenating data from 17 subjects with 128 volumes each which should give me 128*17 points on the timecourse plot in the Melodic report. However, the resulting timecourse plot only shows 255 values (this value is independent of the number of subjects that are entered in the group ICA) and looks very abnormal. So does the powerspectrum plot. I attached an example for each of them for illustration. 
> 
> When using the command line function of Melodic on the same data (which does not perform the registration, however) it works fine.
> 
> I also tried it with the Melodic GUI on another computer which is running FSL version 5.0.8 and it worked without problems.
> 
> Could there be a problem with the newest version?
> 
> Is there an easy way to install an older version of FSL (I am using Ubuntu)?
> 
> Thank you very much for your help!
> 
> Cheers,
> Julia
> 
> <t1.png><f1.png>