What do we really know about the interplay between brain, behavior, and cognition from childhood to early adulthood? An international group effort to generate and share simulated datasets.

Abstract

Background: Neuroimaging has contributed considerably to our understanding of brain development and its relationship to cognition and behavior. However, despite advancements in neuroimaging, replicability in research remains a key issue and there are no gold standard models that quantify neuroanatomical correlates of cognition, behavior and their interplay. Methods: Research groups across the globe have each independently created simulated datasets containing the interplay between brain development, behavior, and cognition. Each group has worked independently and unaware of the approaches and assumptions made by the other groups. Each group was provided the same number of variables and were instructed to create three datasets with each embedding how they envision the interplay between brain development, behavior, and cognition emerges throughout development. Results: We are releasing these simulated datasets to challenge/invite the research community to determine the underlying patterns and assumptions within the simulated datasets. Each dataset contains 10,000 participants over 7 longitudinal waves, ranging from age 7 to 20. The data can be found at https://socoden.github.io/Simulation/. The code and descriptions of the models that were used to create these datasets will be released mid 2025. The research community is invited to explore underlying models and data patterns in simulated datasets and submit their findings. Discussion: Findings from those in the community analyzing data will be evaluated by the groups that generated the simulated datasets. Evaluations will be qualitative and will be used to identify common themes across patterns in simulated datasets and models applied by the community, to learn which proposed neurodevelopment patterns were picked up, which ones were missed and what were common assumptions made. Lessons learned will be disseminated to the research community and may also be beneficial to test with real data as large developmental datasets, such as the ABCD Study and the China Child Brain Development (CCBD) Study, come to fruition.