[derivatives][fmriprep] unbalanced number of anatomical sessions

courtois-neuromod / anat

Anatomical sub-dataset of Courtois-Neuromod project.

2 stars 0 forks source link

[derivatives][fmriprep] unbalanced number of anatomical sessions #2

Closed bpinsard closed 4 years ago

bpinsard commented 4 years ago

For the beta release, not all participants have multiple anatomical sessions. To run fmriprep, we planned to average the T1w/T2w from the different sessions. Should we try to have the same number of session for all participants (1) or the max for each participants (1,2 or 3). I would go with the second choice as I think it shouldn't bias the subsequent functional analysis.

@pbellec @julieaboyle1 @arnaudbore what is your take on that?.

bpinsard commented 4 years ago

No answer, the choice was average all available session for that release. Results in unbalanced number of session across particpants. Should not matter much as models are individual. T1 quality is high so even with 1 session, segmentation/registration should be pretty stable.