Containerizing BIDSme : A Reproducible Tool for BIDS Conversion

2026-06-08

AI Summary

This work addresses the complexity of converting neuroimaging data to the Brain Imaging Data Structure (BIDS) standard and the limited portability and reproducibility of existing conversion tools. To overcome these challenges, the authors propose a containerized solution based on Docker and Docker Compose that encapsulates the BIDSme semi-automated conversion tool. This approach enables lightweight, flexible, and user-friendly deployment while significantly enhancing cross-platform compatibility, computational reproducibility, and seamless integration with neuroimaging platforms such as Neurodesk. Empirical validation demonstrates that the proposed method efficiently supports standardized conversion of multimodal neuroimaging datasets, offering a robust and scalable workflow for the neuroimaging community.

Abstract

The "Brain Imaging Data Structure" (BIDS) has become a widely adopted standard for organizing and sharing neuroimaging datasets of various modalities. However, converting raw brain imaging data into BIDS framework remains a complex and time-consuming task. BIDSme is a semi-automated tool developed to streamline this conversion process, but until recently, it lacked the portability and accessibility needed for widespread adoption. This paper presents the containerization of BIDSme using Docker and Docker Compose, improving usability, reproducibility, and integration into existing platforms like Neurodesk. It also details the design choices, iterative refinements, and validation process that led to a flexible, lightweight, and user-friendly containerized application.

Problem

• BIDS
• neuroimaging data conversion
• reproducibility
• portability
• data standardization

Innovation

• containerization
• BIDS
• Docker
• reproducibility
• neuroimaging