fMRI multiband/SMS preprocessing

Hi, I am pretty new to fMRI analysis. I am using ADNI2 data. I have done some preprocessing using fmriprep docker for single-band scans. All the single band scans have TR value of about 3000msec.
I am using this command:
cmd = “fmriprep-docker " + bids_path + " " + output_path + " --participant_label 01 --fs-license-file /home/yadnesh/Shreya/BIDS/freesurfer.txt --fs-no-reconall --output-spaces MNI152NLin6Asym:res-2”

Now, I came across some scans with multiband with TR value 607msec. I just want to know what changes I need to make in the above command to make it compatible with multiband scan preprocessing.

Hi @Harshit_Samani,

You don’t necessarily need to adjust your fmriprep command above for mutliband data; fmriprep can detect it from the metadata. Two fmriprep flag options that could potentially be affected by multiband data would be --ignore slicetiming and --use-aroma.

For high multiband acceleration factors (e.g. 8), some argue there’s no need for slice timing correction, hence the --ignore slicetiming selection. Others will have stronger opinions on this, but if you plan on performing slice-time correction, I’d suggest reading this recent Poldrack & Mumford blog post to better understand fmriprep’s implementation of slicetiming correction.

fmriprep has a ICA-AROMA denoising option (i.e. --use-aroma), which as far as I know was only developed on singleband data and therefore the denoising may not perform as well on multiband data, though you would have to see for yourself.

Hope this helps a bit.


Another similar discussion here with additional information: