Longitudinal multi-centre brain imaging studies: guidelines and practical tips for accurate and reproducible imaging endpoints and data sharing

BackgroundResearch involving brain imaging is important for understanding common brain diseases. Study endpoints can include features and measures derived from imaging modalities, providing a benchmark against which other phenotypical data can be assessed. In trials, imaging data provide objective evidence of beneficial and adverse outcomes. Multi-centre studies increase generalisability and statistical power. However, there is a lack of practical guidelines for the set-up and conduct of large neuroimaging studies.MethodsWe address this deficit by describing aspects of study design and other essential practical considerations that will help researchers avoid common pitfalls and data loss.ResultsThe recommendations are grouped into seven categories: (1) planning, (2) defining the imaging endpoints, developing an imaging manual and managing the workflow, (3) performing a dummy run and testing the analysis methods, (4) acquiring the scans, (5) anonymising and transferring the data, (6) monitoring quality, and (7) using structured data and sharing data.ConclusionsImplementing these steps will lead to valuable and usable data and help to avoid imaging data wastage.