Core practices for credible neuroscience

Plan before collecting or analysing data 

Credible research starts with a clear research question, a justified design and an analysis plan that is appropriate to the data. Where possible, decide your primary outcome, inclusion and exclusion criteria, sample size, randomisation, blinding and statistical approach before collecting or analysing the relevant data. 

Planning does not prevent exploration. It simply helps distinguish between analyses that were specified in advance and analyses that emerged from the data. Both can be valuable, but they should be reported transparently so readers can understand how conclusions were reached. 

Preregister where appropriate 

Preregistration is a time-stamped record of a research plan, usually placed in an independent registry before data collection or before analysis of an existing dataset. A preregistration commonly describes the rationale, hypotheses, methods, sample size or number of replicates, outcome measures and intended statistical analyses. 

Preregistration can reduce bias, discourage undisclosed analytical flexibility and help researchers plan more carefully. It can also help protect priority, because it provides a time-stamped record of the study idea. Registries include the Open Science Framework, AsPredicted and ClinicalTrials.gov. In animal research, the Animal Study Registry may also be relevant. 

Consider Registered Reports 

Registered Reports are a publication format in which peer review happens before the results are known. The study rationale, methods and analysis plan are reviewed at Stage 1. If the design is accepted in principle, the study is then carried out and the completed paper is reviewed at Stage 2. The outcome of the results should not determine whether the study is publishable, provided the approved study has been conducted and reported appropriately. Journals that allow submission of Registered Reports can be found on the Center for Open Science website - https://www.cos.io/initiatives/registered-reports 

Manage data and code from the start 

Good data management makes research easier to understand, easier to check and easier to share. Decide how raw data, processed data, analysis scripts, metadata and notes will be organised before data collection begins. Use consistent file names, version control where appropriate, and a clear folder structure that will still make sense when a project member leaves. 

Electronic lab notebooks can help researchers keep track of experimental decisions, protocols, deviations, data collection and analysis. They are particularly useful when a lab wants consistent practice across projects or when staff turnover is likely. 

Share data, materials and code responsibly 

Sharing research outputs can increase transparency, support reuse and allow datasets, code and materials to be cited. It can also help researchers demonstrate good practice in grant applications, publications and promotion processes. Repositories such as OSF, Zenodo, figshare, Dryad and domain-specific repositories can help preserve materials and assign persistent identifiers. 

Sharing should be responsible. Human participant data, clinical data and other sensitive materials may require controlled access, consent language, data use agreements or institutional approval. Researchers should consider these issues during study design rather than after data collection. 

Report methods and deviations clearly 

Credible reporting includes enough information for others to understand how the work was done, how the data were analysed and how conclusions were drawn. If the study deviated from the preregistered plan or protocol, report what changed and why. Deviations do not invalidate a study, but hidden deviations make it harder to interpret. 

Use appropriate reporting guidelines where available. Examples include ARRIVE for animal research and COBIDAS for neuroimaging. Where no discipline-specific checklist exists, authors should still provide enough detail for a knowledgeable reader to evaluate the methods and analysis. 


References and links 
  • Center for Open Science preregistration: https://www.cos.io/initiatives/prereg 
  • Center for Open Science Registered Reports: https://www.cos.io/initiatives/registered-reports 
  • Open Science Framework: https://osf.io/ 
  • AsPredicted: https://aspredicted.org/ 
  • ClinicalTrials.gov: https://clinicaltrials.gov/ 
  • Animal Study Registry: https://www.animalstudyregistry.org/ 
  • Zenodo: https://zenodo.org/