fMRIprep MNI normalization issue

[arnaudleroy59]

Hi,

I use fMRIprep which is working well with most of my MRI data. But I have an issue with a few patients when using fMRIprep. Coregistration is ok and no issue reported when looking at the html file. However, when I'm looking at the functional coverage, here is the output and there is nearly a flat time course. Native data seem ok. Have you got any explanation and tips to fix this issue?

Best

Arnaud Leroy

[oesteban]

What you mean by functional coverage?

Can you share the reports with us?

[jdkent]

Answering @oesteban's questions will provide necessary context, but as a hail mary, if @arnaudleroy59 is talking about an ICA-AROMA denoised brain, has ICA-AROMA specified, and is using fmriprep version < v1.2.0, then this reminds of some of the aroma masks before #1332.

I'm likely wrong, but I wanted to throw this out there while I was thinking of it.

[arnaudleroy59]

The green area represent the functional data (using brain voyager or SPM).

This is an ICA-AROMA denoised brain, but it is the same without this correction.

file:///home/cure/Documents/MATLAB/Sebastien/FMRIPREP_DATA/Fmriprep_sub-3064SEAHS1bis/BIDS_OUTPUT/fmriprep/sub-3064SEAHS1bis.html

Thank in advance for your help

Best

[arnaudleroy59]

On addition we are using fMRIprep v1.4. Do you need more information?

[arnaudleroy59]

Hi,

What do you think about these outputs?

Do you need more to answer? If yes, how can I send you?

Thank you in advance for your help

Best regards

Arnaud Leroy

[oesteban]

Besides the full html report (which I requested via private email), could you write here the exact name of the file represented in the figure of your first post?

[oesteban]

Hi @arnaudleroy59, I've checked on your HTML report and it indeed looks good. Can you provide me the exact name of the file you used for the plot on top?

[oesteban]

Okay, I just realized that you added it to the shared file. So, if I understand correctly, the file represented above is sub-3064SEAHS1_task-rest_space-MNI152NLin2009cAsym_desc-smoothAROMAnonaggr_bold.nii.gz, correct?

Could you please post the same plot for sub-3064SEAHS1_task-rest_space-MNI152NLin6Asym_desc-smoothAROMAnonaggr_bold.nii.gz? If this one doesn't look bad, then I think you just confirmed the problem surfaced in #1809 and https://neurostars.org/t/ica-aroma-fmriprep-using-child-template/5139/2, which will be addressed in #1816.

[oesteban]

I'll wait for @arnaudleroy59 to check on the space-MNI152NLin6Asym aroma derivative. If that one looks good then we can close this and follow up on #1816.

[arnaudleroy59]

Dear Doctor Esteban,

The issue is the same using MNI152Lin6Asym aroma output.

Best regards

Dr Arnaud Leroy

Envoyé de mon iPhone

Le 11 oct. 2019 à 21:17, Oscar Esteban notifications@github.com<mailto:notifications@github.com> a écrit :

I'll wait for @arnaudleroy59https://github.com/arnaudleroy59 to check on the space-MNI152NLin6Asym aroma derivative. If that one looks good then we can close this and follow up on #1816https://github.com/poldracklab/fmriprep/issues/1816.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/poldracklab/fmriprep/issues/1766?email_source=notifications&email_token=ANEXK5I4FKJZD2NN636O7DLQODGKBA5CNFSM4IVTEJ72YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEBA6UMA#issuecomment-541190704, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ANEXK5PVDEE2UOVKOXKJZY3QODGKBANCNFSM4IVTEJ7Q.

[oesteban]

Hi @arnaudleroy59, we are still looking into this. Our team member @rciric will soon ask you for further details to follow up. Thanks for your patience.

[rciric]

Hello — to get a better sense of the issue, can you confirm whether the same issue occurs in the preprocessed brain prior to denoising with ICA-AROMA (sub-3064SEAHS1_task-rest_space-MNI152NLin6Asym_desc-preproc*_bold.nii.gz)? You previously mentioned that “[the issue was] the same without [ICA-AROMA]”, and I’d like to make sure that this problem doesn’t relate to fsl_regfilt.

[arnaudleroy59]

Dear Dr Rastko Ćirić,

I confirm that the same issue occurs in the preprocessed brain prior to denoising with ICA-AROMA.

I red previous posts about other issues with ICA-AROMA . Does it mean that only MNI152NLin6Asym is available for all patients when using ICA-AROMA (I didn’t understand if MNI2009 was also available after denoising with ICA-AROMA)?

Thank you in advance for your help

Best regards

Dr Arnaud Leroy

De : Rastko Ćirić [mailto:notifications@github.com] Envoyé : vendredi 1 novembre 2019 22:31 À : poldracklab/fmriprep fmriprep@noreply.github.com Cc : LEROY Arnaud1 Arnaud.LEROY@CHRU-LILLE.FR; Mention mention@noreply.github.com Objet : Re: [poldracklab/fmriprep] fMRIprep MNI normalization issue (#1766)

Hello — to get a better sense of the issue, can you confirm whether the same issue occurs in the preprocessed brain prior to denoising with ICA-AROMA (sub-3064SEAHS1_task-rest_space-MNI152NLin6Asym_desc-preproc*_bold.nii.gz)? You previously mentioned that “[the issue was] the same without [ICA-AROMA]”, and I’d like to make sure that this problem doesn’t relate to fsl_regfilt.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/poldracklab/fmriprep/issues/1766?email_source=notifications&email_token=ANEXK5MVJABVU6RG63RWT7LQRSNY3A5CNFSM4IVTEJ72YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEC4G4SQ#issuecomment-548957770, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ANEXK5LJNJRRBS2VFYMWATTQRSNY3ANCNFSM4IVTEJ7Q.