brainprep¶
BrainPrep is an integrated tool designed to preprocess MRI data with a strong emphasis on transparency, traceability, usability, and portability.
Core Principles¶
Transparency
BrainPrep is built to be understandable. Every step in the preprocessing workflow is explicitly defined, documented, and inspectable. Users can trace how data is transformed, which parameters are applied, and which tools are invoked without needing to dig through opaque scripts or hidden configurations.
Traceability
Reproducibility is central to scientific integrity. BrainPrep ensures that every preprocessing run is traceable, with automatic logging of inputs, outputs, tool versions, and execution metadata. This makes it easy to audit results, share workflows, and reproduce findings across labs and platforms.
Usability
BrainPrep is designed for researchers, not just developers. Its interfaces are intuitive, its wrappers are consistent, and its workflows are modular. Whether you’re a novice or an expert, BrainPrep aims to reduce the cognitive load of preprocessing so you can focus on your science.
Portability
To ensure consistent execution across different systems, BrainPrep is distributed and executed using Docker or Apptainer. This containerized approach guarantees that all dependencies, tools, and configurations are bundled together, eliminating environment-specific issues and simplifying deployment.
Design Trade-offs¶
BrainPrep consciously deprioritizes:
Performance: While efficiency matters, BrainPrep favors clarity over speed. It avoids aggressive parallelization or optimization that might obscure the logic of the workflow.
BIDS Compliance: BrainPrep supports BIDS principles but does not enforce strict adherence. Instead, it provides flexible input/output handling that can accommodate diverse data structures and legacy formats.
Wrapper Architecture¶
All preprocessing steps in BrainPrep are implemented using decorators, which wrap either:
Command-line tools: These wrappers return a command or list of commands to be executed, along with a tuple of generated output paths.
Python functions: These wrappers return only a tuple of generated output paths, encapsulating logic written directly in Python.
This architecture promotes modularity, reusability, and clarity, making it easy to build, inspect, and extend workflows.
Get started with brainprep
Discover functionalities by reading examples
Learn about brainprep
