Pipelines
The large amount of complex processing steps in neuroimaging analysis makes the use of pipelines and automatisation necessary. Due to the existence of a variety of options these processes have been shown to contribute to the reproducibility crisis [1,2].
In our subgroup we aim to streamline neuroimaging pipelines to promote reproducibility and sustainable practices by:
- Encouraging sharing of data, meta data and the complete pipelines used
- Promoting transparency of code and dependencies inherent to the pipelines
- Facilitating reproducibility of specific pipelines by promoting dockers and neuroimaging data structures like BIDS [3]
- Developing pipelines and promoting practices that focus on reducing the environmental costs
References:
- Lindquist, Martin. “Pipeline choices alter neuroimaging findings.” Nature 582.7810 (2020): 36-37.
- Altuna Akalin, 2018. Scientific Data Analysis Pipelines and Reproducibility. Published online
- Gorgolewski, Krzysztof J., et al. “The brain imaging data structure, a format for organizing and describing outputs of neuroimaging experiments.” Scientific data 3.1 (2016): 1-9.