Aligning multirun data with blip-down fieldmap

,
1

Dear Users,

I am trying to preprocess an fMRI dataset collected from a single subject with 60+ runs. These runs were collected across a period of 5 days and each time the subject was in the scanner, 4-5 runs were collected. Every time the participant was put inside the scanner, a new run of blip-down sequence (for fieldmap correction) was collected.

My question is if I apply TOPUP on each dataset (before motion correction), would it be valid to align all the runs collected on different days, with different fieldmap (i.e. shimming) together, and coregister to a single T1?

Another way is to run the preprocessing separately for each group of runs where the blip-down sequence (i.e. shimming) is the same for that set of acquisitions. What would be the rationale/benefit for analysing in this way?

Relatedly, on AFNI, I do not think it is possible to indicate multiple “blip_reverse_dataset” for a list of runs (please correct me if wrong).

Thank you so much in advance.

2

Regarding AFNI, you are correct. But in this case I would process the different days separately anyway.
The question would be how to align the EPI across days. One could use the same volreg base across sessions, or have each session align to the anat (using afni_proc.py -volreg_align_e2a). A third option is to use an affine alignment across all runs, initially aligning each run to its own MIN_OUTLIER, say. It is not clear which of these will work the best. I would be inclined to either rely on anatomical alignment across sessions (so process one day at a time), or to use cross run allineate (-volreg_post_vr_allin).