Hi Pablo,

What is your objective besides 'nuisance regression' in the RS-fcon analysis? Formerly, a nuisance variable is a random variable that is fundamental to the probabilistic model, but that is of no particular interest in itself or is no longer of interest (i.e., head motion). Thus, nuisance variables are treated as covariates of no interest in most of neuroimaging studies, so actually, they are regressed out from data.

I recommend you to visit the forum of SPM for related issues:

http://www.fil.ion.ucl.ac.uk/spm/support/

PD. Depending on your aim, you would be interested in parametric regression approaches: i.e., predict the RS data associated to parameters of interest.

Best,

Hi Martín,

thanks for your answer, very much appreciated! The nuisance regression is part of my resting state preprocessing pipeline. The overall objective is to submit the cleaned data to a Graph theoretical analysis, in which I compare basically how FC changes in relation to psychopharmacological treatment. So, I had the idea to submit the data to a nuisance regression and write out a new 4D Volume which I then go on to process further.

Indeed I could also do the Nuisance Regression in SPM12 as covariates, but the ROI to ROI approach is, as far as my understanding goes, a lot easier to conduct with other, specialized toolboxes like DPABI or GRETNA.

If you have further advice, I'm very grateful for help!

Best regards,

Pablo

Hi, Pablo.

I guess that you want to regress out the signals from the white matter and CSF as a part of denoising step. If this is the case, you are supposed to normalize c2 and c3 images into the MNI space assuming that your functional images are also in the MNI space. Note that the voxel size should be matched to the functional image (e.g., 2*2*2). Of course, you can do that with SPM12 (normalization).

best regards,

Haemy