HALFpipe: Advancing reproducible fMRI analysis based on fMRIPrep
Short description and the goals for the OHBM BrainHack
Using standardized pipelines leads to consistent results, right? We ran fMRIPrep 20.2.7 one hundred times and then calculated functional connectivity matrices using HALFpipe.
We did this in a multiverse approach where we varied which denoising strategy was applied before connectivity matrix calculation. An interesting next step here is to figure out how to quantify the variability in these matrices depending on the strategy. Are there any strategies that better than others?
Any interested contributors may download the data and attempt to answer this question with us!
We would like to repeat this procedure with newer versions of fMRIPrep, and compare to the baseline that we already ran. How does the variability between versions compare to the variability within a single version?
This will require some programming to load in derivatives files from different versions of fMRIPrep and then apply the Nipype workflows from HALFpipe to them. We are welcoming contributors willing to expand their familiarity with Nipype workflows in Python or BIDS derivatives.
The Enhancing Neuro Imaging Genetics through Meta Analysis (ENIGMA) Consortium conducts the largest brain imaging studies in the world, involving over 500 institutions in 45 countries worldwide. At this time, the ENIGMA Consortium has processed more than fifty thousand fMRI scans using fMRIPrep, mostly using HALFpipe as a user interface to fMRIPrep for ease of use. HALFpipe also calculates downstream results such as functional connectivity maps and matrices.
While fMRIPrep supports resampling fMRI data to the cortical surface, the current HALFpipe workflows currently can not use those outputs. We would like to change this, and have a roadmap of the Nipype workflows in HALFpipe that need to be adapted.
We would love to discuss the best approaches and potentially even try implementing some of the changes.
Title
HALFpipe: Advancing reproducible fMRI analysis based on fMRIPrep
Short description and the goals for the OHBM BrainHack
Using standardized pipelines leads to consistent results, right? We ran fMRIPrep 20.2.7 one hundred times and then calculated functional connectivity matrices using HALFpipe.
Link to the Project
https://github.com/HALFpipe/HALFpipe
Image/Logo for the OHBM brainhack website
Project lead
Main Hub
Seoul
Link to the Project pitch
No response
Other hubs covered by the leaders
Skills
Brain Imaging Data Structure (BIDS)
Recommended tutorials for new contributors
No response
Good first issues
No response
Twitter summary
No response
Short name for the Discord chat channel (~15 chars)
halfpipe
Please read and follow the OHBM Code of Conduct