Harnessing population diversity: in search of tools of the trade

Danilo Bzdok,Guy Wolf,Jakub Kopal
DOI: https://doi.org/10.1093/gigascience/giae068
IF: 7.658
2024-09-28
GigaScience
Abstract:Big neuroscience datasets are not big small datasets when it comes to quantitative data analysis. Neuroscience has now witnessed the advent of many population cohort studies that deep-profile participants, yielding hundreds of measures, capturing dimensions of each individual's position in the broader society. Indeed, there is a rebalancing from small, strictly selected, and thus homogenized cohorts toward always larger, more representative, and thus diverse cohorts. This shift in cohort composition is prompting the revision of incumbent modeling practices. Major sources of population stratification increasingly overshadow the subtle effects that neuroscientists are typically studying. In our opinion, as we sample individuals from always wider diversity backgrounds, we will require a new stack of quantitative tools to realize diversity-aware modeling. We here take inventory of candidate analytical frameworks. Better incorporating driving factors behind population structure will allow refining our understanding of how brain–behavior relationships depend on human subgroups.
multidisciplinary sciences
What problem does this paper attempt to address?