Neuroimaging studies are complex, sensitive, costly to generate, and valuable for reuse, making them a stress test for reproducibility, privacy, governance, and long-term preservation. This talk will argue that the reproducibility crisis is best understood as part of a broader transparency crisis: science becomes trustworthy only when others can evaluate the evidence, methods, data, assumptions, and infrastructures on which claims depend.
Using examples from neuroimaging, Cyril will discuss transparency across the research lifecycle: consent and participant autonomy, pre-registration, BIDS-based data organisation, scripted workflows, multiverse analyses, privacy-risk quantification, governed data access, and sustainable open-science infrastructure. He will highlight tools and frameworks such as Open Brain Consent, Multiverse Analytics, MetaPrivBIDS, and PublicnEUro to demonstrate how transparency can be engineered into research practice. The central message is that open science is not simply about making data public; it is about designing research systems that make evidence inspectable, reusable, ethically governed, and trustworthy over time.
