search

NBCLab / power-replication

A replication and extension of Power et al. (2018)
https://www.overleaf.com/read/swgjxcjqytxg
Apache License 2.0
2 stars 0 forks source link

fMRIPrep settings #8

Closed tsalo closed 3 years ago

tsalo commented 3 years ago

I'm working on the fMRIPrep processing job and I just wanted to check about the settings. Here are some initial thoughts:

  1. --fs-no-reconall: I'm running fMRIPrep on ~800 participants, so processing with recon-all will add a lot of time. The analyses don't use surface space, so I don't think it's that important. EDIT: I actually included masks that are only available from fMRIPrep in my analysis plan, so I have to use Freesurfer :(
  2. --output-spaces MNI152NLin6Asym:res-native anat:res-native func:res-native: I'll be analyzing the echo-wise files grabbed from the working directory, but I think these are the target spaces. Does anyone have a concern about analyzing everything in native resolution?
  3. Any other settings worth adding/amending? A number of things (CompCor, FD outliers) will be calculated as part of the analyses, rather than taken from fMRIPrep, but there could be other options worth using. Here are the current version's arguments.

EDIT: Tagging @62442katieb @mriedel56 @emdupre @handwerkerd

tsalo commented 3 years ago

I've started submitting the jobs, so I'm closing now.