Switching fmriprep version mid-study

1

Hi everyone,

I’m curious about the implications of changing fMRIPrep versions in the middle of a study. Specifically, is it safe to switch between different minor releases, or even between major releases? My intuition says it’s probably not ideal, but I’d love to understand the reasons behind this.

Thanks for any insights!
Shane

2

Hi @swwalsh1,

You should not switch releases mid-study unless you reprocess everyone. Even minor versions can produce changes, e.g., Continuous Evaluation of Denoising Strategies in Resting-State fMRI Connectivity Using fMRIPrep and Nilearn - PubMed. You should address the changelog to see if a minor update would affect your outcomes (they may not depending on your workflow), and always assume major changes are incompatible with each other.

Best,
Steven

3

I was going through the FAQ recently and I remeber they have an entry for that:
https://fmriprep.org/en/stable/faq.html#a-new-version-of-fmriprep-has-been-published-when-should-i-upgrade