Linking the exposome to the brain-behaviour phenotype

A range of environmental, lifestyle and biological exposures across the lifespan - varying in timing, duration and intensity - interact with genetic factors to shape an individual's neurocognitive phenotype. By referring to the totality of exposures that an individual has experienced in their life so far, the exposome offers a valuable concept to better understand interindividual variability in not only brain-behaviour phenotype but also vulnerability and resilience to brain diseases. Numerous large-scale neuroimaging projects are enriched with extended exposomic data including sociodemographic, biomedical, lifestyle and environmental measurements. Yet, deciphering how this complex web of influences collectively and dynamically shapes individual brain-behaviour phenotypes comes with substantial challenges. In this Perspective, we first emphasize how the exposome concept refers to a set of interrelated and interacting factors and outline how multivariate pattern learning enables us to account for this complexity. We also highlight temporality as a key challenge, as the timing, duration and sequencing of exposures importantly shape their associations with brain-behaviour phenotypes, and we discuss existing approaches to address these dynamics and their limitations. We further underscore the challenge of causal inferences in population datasets, especially given bidirectional exposome-brain relationships. We then refer to traditional statistical design, as well as generative models and causal machine learning, as promising perspectives to approach data entanglement. Finally, we conclude that, to truly benefit global health, the field also needs to address the lack of global diversity in brain-behaviour exposomic research.