Developmental cognitive neuroscience initiatives for advancements in methodological approaches: Registered Reports and Next-Generation Tools

Why we wrote it

This is an editorial, not a study. In 2020 Developmental Cognitive Neuroscience began accepting two new kinds of articles, Registered Reports and Next-Generation Tools, and we wrote it to explain what they are and why the journal wanted them. Developmental researchers increasingly work with big, open, longitudinal datasets, and those datasets create particular problems for how we plan and report analyses.

Registered Reports

In a Registered Report, authors submit the introduction and methods before they collect or analyze data. Reviewers comment on the plan, and if it passes, the journal commits to publishing the paper whatever the results turn out to be. That takes the pressure off getting a striking result, which is a common source of p-hacking and publication bias. Exploratory analyses are still welcome, and they are labeled as exploratory.

We put most of our attention on Registered Reports of secondary data analysis, because we expect those to be common in this field. When the data already exist, the plan should say what the authors already know about the dataset, including their own earlier work, unpublished analyses, and what others have published. The plan should also guard against overfitting, for example with cross-validation. We use the Adolescent Brain and Cognitive Development (ABCD) Study as an example. Because so much about that sample is already public, almost any new analysis of it counts as secondary, and we explain why we will not issue a call for Registered Reports at every data release.

Next-Generation Tools

The second article type is a home for methods work: best practices, corrections to old procedures, and new software, ideally with code others can use. Motion correction in resting-state MRI is our example of a field revising its methods, and earlier findings had to be revisited as a result. We also see this as a way for early-career researchers to get credit for tools they spend a long time building.

What it means

We describe both formats as an evolving process rather than a finished recipe. In our view, developmental cognitive neuroscience needs new publication formats and new analytic tools if it is to make good use of big data. We hope other journals and research groups find the same approach useful.