Hello, I am running fmriprep 1.5 with --use-aroma.
I was reviewing the html reports and the melodic_mixing.tsv files and noticed the number of components in the html file does not match the number of columns in the melodic_mixing.tsv file.
If I understand this correctly, those should contain the same number of components/columns correct?
I was hoping to do the following:
Inspect the html report and adjust the AROMAnoiseICs.csv file accordingly (to account for components that I believe to be noise that were not picked up by AROMA)
Then my plan was to to smooth (6mm FWHM) the desc-preproc_bold.nii.gz file and then use fsl_regfilt on the desc-preproc_bold.nii.gz file with the adjusted AROMAnoiseICs.csv file in addition to adding some other confound regresses for denoising.
If I understand this correctly I would have to create a custom confound_regressors.tsv file containing only the confounds I am wanting to regress out.
For clarity could someone literally type out the command line that I would be wanting to implement?
My best guess would be this:
fsl_regfilt -i sub-###_ses-###_task-###_space-###_desc-preproc_bold.nii.gz -d sub-###_ses-#_task-###_space-###_custom_desc-MELODIC_mixing.tsv -f sub-###_ses-###_task-###_space-###_custom_confound_regressors.tsv -o name_of_output_file
Where:
###= rest of file name in a chosen space" T1w, MNI152NLin6Asym, MNI152NLin2009cAsym (which would be best???)
custom_desc-MELODIC_mixing.tsv is the desc-MELODIC_mixing.tsv with the added confound columns added to the matrix)
custom_confound_regressors.tsv is a .tsv file that includes the column numbers that represents the AROMA noise components in the custom_desc-MELODIC_mixing.tsv and the other confounds chosen to be regressed from the data that were added to the custom_desc-MELODIC_mixing.tsv (WM, CSF, GS or others from the confound file produced by FMRIprep)
Is this correct? If so, how does the discrepancy between the html components and the mixing.tsv file impact this?
Thank you in advance