Hi,

Yes, you can, this is fine.

However, you should have at least 200 participants (or ideally more) otherwise the model can't be determined.

Two other comments:

- If the regions are discrete (not fuzzy), they should not overlap.

- Drop he option --demean from the first run of fsl_glm (you need to edit the dual_regression script, which is in ${FSLDIR}/bin

All the best,

Anderson

On Thu, 19 Jul 2018 at 01:06, SUBSCRIBE FSL Anonymous <[log in to unmask]> wrote:

Hi,

Can I use a 4D brain parcellation (e.g. a brain image parcellated into 200 ROIs that is subsequently converted into a 4D .nii file with one ROI per volume; 200 volumes total) in place of a melodic_IC file as an input to dual_regression (DR)? Ultimately I would like to use only the stage 1 output of dual_regression as input for FSLNets.

I could create 200 individual masks from the 200 ROIs and generate 200 averaged timecourses (per subject), however this seems painstaking and I'm concerned about getting everything into the format preferred by FSLNets. Since I know that DR stage 1 output will give me the right kind of FSLNets input (and will do everything in one step) I was hoping to do things with DR. Is this a methodology that makes sense?

Thanks very much!

########################################################################

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1