Neuroimaging Toolkit

This toolkit gives neuroimaging researchers practical steps for strengthening credibility from study planning to data sharing. It focuses on preregistration, sample size, privacy, data organisation, reporting, FAIR principles and sharing data and code responsibly. 
Planning your study:

Preregister the study 

Preregistration is particularly useful in neuroimaging because many aspects of acquisition, preprocessing and analysis can introduce flexibility. A preregistration should describe the research question, hypotheses, acquisition plan, sample size, inclusion and exclusion criteria, preprocessing steps, primary contrasts or outcomes, statistical approach and planned exploratory analyses. 

Templates can help researchers specify the many parameters involved in neuroimaging. Field-specific templates exist for some fMRI and EEG/ERP studies. Link this section to the wider Preregistration and Registered Reports page rather than repeating general definitions here. 

Plan sample size realistically 

Neuroimaging research can be expensive and time-consuming, which can limit recruitment. However, small samples can reduce reliability and generalisability. Sample size should be justified using previous literature, simulations or power calculation tools where appropriate. The design matters: task-based studies may require different sample sizes from resting-state or brain-wide association studies. 

Where manual decisions are unavoidable, apply them blind to the experimental conditions where possible. Quality assessment should be orthogonal to the contrast of interest. Exploratory analyses can be included, but they should be distinguished from analyses planned in advance. 

Consider data protection early 

Researchers should decide before data collection whether and how data might be shared. Neuroimaging data can be personal data, and true anonymisation may be difficult where a key is retained or where the data could reveal health information. Consent, institutional approval and data use agreements should be planned early if the data may be reused by other researchers.

During your study:

Use a clear data structure 

A robust data structure makes neuroimaging projects easier to analyse, audit and share. The Brain Imaging Data Structure, BIDS, provides a standard for organising neuroimaging data across modalities. Using BIDS can make datasets easier to validate, combine and analyse with community tools. 

Tools such as dcm2bids and dcm2niix may help convert folders into BIDS format, although the status of converter tools should be checked because some are no longer under active development or may not support recent BIDS extensions. Researchers should also maintain multiple secure copies of their data. 

These videos explain these processes in more detail and are taken from BNA credibility webinars.

Neuroimaging 1.png

 

Neuroimaging 2.png

 

After data collection:

Check that sharing is compliant 

Before sharing neuroimaging data, confirm whether consent, institutional approval and any data use agreement permit the intended form of sharing. If the data are personal or potentially identifiable, controlled access may be more appropriate than fully open release. 

Follow FAIR principles 

Data intended for sharing should be findable, accessible, interoperable and reusable. This means depositing data in a suitable repository, providing documentation and metadata, using persistent identifiers, and describing access conditions clearly. NeuroVault can be useful for sharing statistical maps, and other repositories may be appropriate depending on data type and consent. 

Report results clearly 

Transparent reporting is essential for neuroimaging research. Authors should explain acquisition parameters, preprocessing, quality control, exclusions, software versions, analysis choices and statistical thresholds. Where relevant, activation maps can be uploaded to NeuroVault so others can browse or reuse them. 

Share data and code 

Sharing data can make datasets secure, citable and reusable. Sharing code can strengthen reproducibility by showing how analyses were conducted and allowing others to identify errors or adapt methods. Platforms such as GitHub and Bitbucket can host code, but code should also be archived in a stable repository where a DOI can be assigned when possible.

References and links 
  • BIDS: https://bids.neuroimaging.io/ 
  • NeuroVault: https://neurovault.org/ 
  • Open Brain Consent: https://open-brain-consent.readthedocs.io/ 
  • Andy’s Brain Book: https://andysbrainbook.readthedocs.io/en/latest/ 
  • FSL documentation: https://fsl.fmrib.ox.ac.uk/fsl/docs/#/ 
  • SPM documentation: https://www.fil.ion.ucl.ac.uk/spm/docs/ 
  • MRC CBU list of openly available datasets: https://imaging.mrc-cbu.cam.ac.uk/methods/OpenDatasets 
  • COBIDAS guidance: Nichols TE et al. Best Practices in Data Analysis and Sharing in Neuroimaging using MRI. 2016. bioRxiv doi: 10.1101/054262 
  • Poldrack R, Baker C, Durnez J et al. Scanning the horizon: towards transparent and reproducible neuroimaging research. Nature Reviews Neuroscience, 2017. https://doi.org/10.1038/nrn.2016.167 
  • Marek S et al. Reproducible brain-wide association studies require thousands of individuals. Nature, 2022. 
  • Niso G et al. Open and reproducible neuroimaging: From study inception to publication. NeuroImage, 2022. doi: 10.1016/j.neuroimage.2022.119623 

 This toolkit was produced in 2024 by Grace Cooper with guidance from Olivia Kowalczyk, Roni Tibon, Josefina Weinerova and Esther Walton.