ReproNIM Principle 1: Study planning

Study planning represents a foundational step in ensuring reproducible neuroimaging research. Many problems that lead to reproducibility issues can be avoided by proper planning

  1. Implement good science basics, e.g, power analysis, statistical consults
    If your study is underpowered or your study design is flawed, you will have trouble publishing your work and, more to the point, your work will likely not be reproducible (Scuzs and Ioannidis, 2020).

    • ReproNim resources**:**

    • Other resources:

      • NINDS-funded Community for Rigor (C4R): tools and resources for improving rigor and reproducibility, from basics to advanced

      • Statistical consult: Most universities have statisticians available for consultation. The most important time to consult with a statistician is at the conceptualization phase, not after the data are collected.

  2. Use pre-existing data for planning and/or analysis
    Leveraging pre-existing data, whether your own or data available from public repositories, can help validate methodological approaches before collecting new data and potentially increase your sample size

    • ReproNim resources**:**
      ReproNim has tools and approaches to make it easier to find and reuse your own data or publicly available data

    • Other resources:

      • OpenNeuro (and other repositories)

  3. Create a data management and sharing plan (DMSP)
    A key component of study planning is carefully considering data management needs from the start. By anticipating the volume and types of data that will be generated, researchers can allocate appropriate resources for storage, processing, and sharing. The NIH’s requirement for a Data Management and Sharing Plan (DMSP) in all proposals helps formalize this planning process. The DMSP requires researchers to specify crucial details like which data standards will be implemented and where data will ultimately be shared. By addressing these requirements during the planning phase, researchers can integrate data sharing preparations into their workflow from the beginning, rather than treating it as an burdensome afterthought at the study’s conclusion.

  4. Adopt open consent to allow broad sharing of data
    To ensure that data can be broadly shared, it is important that research participants provide consent for broad public sharing of the data, rather than for narrow purposes.

    • Resources
      • The Open Brain Consent project provides guidance, tools and templates for incorporating open consent into neuroimaging studies while still providing protections for human subjects through de-identification of data.

  5. Pre-register your study to increase transparency and build trust
    Pre-registering a study involves documenting and publishing the research plan, including details about study hypotheses, design and analysis, before the study begins. Preregistration makes research more transparent and reproducible, helps reduce duplication of research, can improve study design and separates hypothesis-generating (exploratory) from hypothesis-testing (confirmatory) research. Some journals support Registered Reports, a peer reviewed study pre-registration which guarantees the publication of the subsequent results if the study adheres to the protocol, regardless of whether or not the hypothesis was supported.