The values of reproducibility and data sharing gain more and more importance in the research community. To make neuroscience studies reproducible and allow for easy data sharing, the Brain Imaging Data Structure (BIDS, https://bids.neuroimaging.io/) standard has been developed. It allows researchers to store their neuroimaging data in a standardized way, speeding up data sharing efforts. Since the development of this BIDS standard, the advantages of using the proposed data structure have been growing steadily: multiple bids-apps have been developed for common data processing pipelines used in the neuroscience community, allowing users to directly load or generate their BIDS-conform data set. Among those apps, is the bids-validator that easily checks the correctness and completeness of all files. BIDS extension proposals (BEPs) have followed for various other modalities used in neuroscience research, such as EEG or PET.
We recently proposed a new standard for storing the necessary data to reproduce a computational neuroscience modeling study. We exemplified our standard by putting a first data set for one of our computational modeling studies using the neuroinformatics platform The Virtual Brain (thevirtualbrain.org/) into the proposed data structure (Triebkorn et al., 2020). We have a first set of scripts allowing the conversion from the modeling and imaging data to the proposed standard but they are so far limited to the specifics of that particular study.
The goal of this project is to develop a user-friendly tool to store study details and data for full reproducibility of computational neuroscience modeling publications. We want to incorporate the existing tools for MRI data, like e.g. bidscoin, into our tool and possibly tools from other modalities.
As an end goal, it would be ideal to have a couple of diverse examples for using the developed tool in the form of computational modeling study publications prepared for sharing in this BIDS-conform standard. To be more user-friendly, the conversion tool should also offer a graphical user interface (Figure 1) where researchers can still add their user knowledge to the conversion tool. One could also aim for developing an extension of the bids-validator app for the newly developed standard to be able to easily check the validity and completeness of files.
For this project, we will need to stay in close communication with the research community. The GSOC contributor will be involved in the ongoing efforts of the neuroscience community to standardize the storage of data, code and simulation details. This will allow the GSOC contributor to build a large network involving different experts for the different modalities of neuroscience research. If successful, this project will represent the groundwork for future computational neuroscience studies offering an easy way to store the accompanying data and code in a standardized way enabling replication.
Lead mentor: Dr. Jil Meier @jilmeier (Charité Universitätsmedizin Berlin) - CET time zone
Co-mentors:
Dr. Michael Schirner @MichaelSchirner (Charité Universitätsmedizin Berlin, CET time zone),
Lia Domide @liadomide (Codemart, software development company based in Romania, EET time zone),
Prof. Daniele Marinazzo @Daniele_Marinazzo (Ghent University, CET time zone),
Prof. Petra Ritter @PetraRitter (Charité Universitätsmedizin Berlin, CET time zone)
Planned effort: 350 hours
Intended skill level: all levels welcome, project can be adapted to the contributor’s level
Pre-requisite skills: basic programming knowledge
Tech keywords: Python, docker, BIDS
Further references:
- Poldrack, R.A., … , Ritter, P. and Rogers, T.T., 2019. The importance of standards for sharing of computational models and data. Computational brain & behavior, 2(3), pp.229-232, doi.org/10.1007/s42113-019-00062-x
- bids-validator: https://github.com/bids-standard/bids-validator
- Pull request for computational modeling BEP: https://github.com/bids-standard/bids-specification/pull/850
- Triebkorn et al. (2020): https://www.biorxiv.org/content/10.1101/2020.03.26.009795v1.abstract
- Bidscoin: https://github.com/Donders-Institute/bidscoin