Hello Zilong
have a look here
http://fsl.fmrib.ox.ac.uk/fslcourse/lectures/practicals/melodic/index.html
Also every option in the Melodic GUI has a tool tip describing its
usage. Concatenation will be done and motion correction/brain extraction
of the functional image can be by the gui.
For further analysis you need to be able to decide if a resulting
independent component is noise or data. There 2 ways in general: visual
inspection of timecourses and spatial extension or automatic
classification using a lot of different tools around (FSL has Fix, which
needs training data, and fsl_regfilt; there are matlab and stand-alone
tools).
It's also useful to know the influence of each (pre-)processing step to
decide what to choose. There you need to have a look at the literature
and ask your people in the lab. It's also dependend on your fmri setup.
Greetings,
Matthias
> Hi FSL experts, I am a beginner in resting state fMRI studies. Now I
> have a dataset with 16 subjects, and for each subject, there are four
> runs. I have brain extracted and motion corrected the structural images
> and BOLD images. I am not sure what should I do next with the data in
> the Melodic gui. Should I first concatenate the four runs for each
> subject like the first level analysis in FEAT gui? After that, should
> I do a group level like analysis as in FEAT gui?
>
> Could you provide a practical guideline (including the what options
> should I use in the Melodic gui) for me to follow?
>
> Thank you very much!
>
> Best,
> Zilong
|