AROMA and mean FD regression

[arwang]

Hi,

I am analyzing a resting-state dataset and have used ICA-AROMA as well as wm, csf, and global signal regression and a bandpass filter (xcpEngine).

I read a paper published in 2013 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3896129/) that recommends adding mean FD for each individual as a nuisance regressor in group-level analyses even after making corrections for motion in each individuals’ scans.

I know this is not specifically an ICA-AROMA forum or motion-artifact removal forum but was just wondering if this is still recommended when utilizing ICA-AROMA. Any potential for re-introducing noise if, for example, I include mean FD + age in a partial correlation between functional connectivity and a behavioral measure?

Thanks

[Mazz]

Hi,

I’ve just started doing fMRI analyses this year, so do take what I say with a grain of salt.

The link below suggests you would still add mean FD as a nuisance regressor at the group-level (fmriprep also uses ICA-AROMA).

Confounds from fmriprep: which one would you use for GLM?

Opinions on this topic are divided and this is why FMRIPREP provides those regressors instead of cleaning up the data for you. I personally would include 6 motion parameters, FD, and aCompCor on run level and mean FD on group level (for both task and rest).

Could you clarify what you mean by re-introducing noise? Are mean FD + age your covariates in your partial correlation?

Any potential for re-introducing noise if, for example, I include mean FD + age in a partial correlation between functional connectivity and a behavioral measure?