Porcupine: A visual pipeline tool for neuroimaging analysis

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Scientific Software Development

Background:

• Neuroimaging research increasingly demands reproducible data acquisition and analysis.
• Complex analysis pipelines often involve multiple tools and programming languages, hindering code sharing and reproducibility.
• Current methods make it difficult for researchers to share or recreate complex analysis scripts.

Purpose of the Study:

• To introduce Porcupine, a novel tool for visually constructing neuroimaging analysis pipelines.
• To automate the generation of analysis code and shareable execution environments (Docker images).
• To improve the understanding, reproducibility, and shareability of neuroimaging analyses.

Main Methods:

• Porcupine enables visual construction of analysis workflows.
• The tool automatically generates analysis code, currently integrating with Nipype functionalities.
• It creates Docker images to ensure a consistent and shareable execution environment.

Main Results:

• Porcupine facilitates the creation of reproducible and shareable neuroimaging analysis pipelines.
• The visual interface improves understanding and allows for manual code modification.
• It bridges the gap between conceptual analysis design and practical implementation.

Conclusions:

• Porcupine simplifies the development and communication of complex neuroimaging analyses.
• It lowers the barrier for users to generate reusable and reproducible analysis pipelines.
• The tool promotes reproducible science by making analysis construction, visualization, and sharing more accessible.