Abnormal FWHM for multi-echo data

AFNI Message Board
,
1

AFNI version info (afni -ver): 23.1.10

3dFWHMx -input ./MNI152NLin2009cAsym_desc-preproc_bold.nii.gz  -mask ./MNI152NLin2009cAsym_desc-brain_mask.nii.gz -acf

3dFWHMx.1D
3dFWHMx.1D1024×787 17.3 KB

Hello again,

I cannot thank the community enough for the help I have been getting with my analysis. I am on my way to finalizing the GLM script for my analysis with my multi-echo data (pre-processed through fMRIprep and tedana).

I have used 3dFWHMx in order to smooth the data for 3dDeconvolove but it seems like my FWHM is unusually large for my functional data. Would anyone happen to know why this might be the case? The current functional data that I am using has been denoised and optimally combined the 3 echos that I have. Would you recommend that I do not smooth the data for the GLM?

None of the preprocessing steps should have any smoothing preformed, and I think the issue might be coming from the fact that the functional data has been combined for all 3 echos that I have.

Any insight would be very much helpful, thank you!

2

Note that afni_proc.py can use tedana if you tell it to, though that was not the question.

But try adding -detrend, as afni_proc.py would do. This is usually done at a slightly higher level than what would be used in a linear regression.

-rick

3

I don't think optimal combination should change the smoothness of the data. You can run 3dFWHMx on each echo separately to see if this increased smoothness is already present in the separate echoes. --Dan