Brain–body-environment interactions must move beyond isolated exposures toward the expotype: the dynamic configuration of physical, social, lifestyle, and internal factors that jointly shape brain–behaviour phenotypes. At @NatRevNeurosci (nature.com/articles/s41583-0…), we propose a future agenda to assess the exposome as a complex, time-varying system that requires nonlinear models to capture interactions, thresholds, synergistic effects, and cross-domain buffering mechanisms. Although predictive machine learning can support individual risk estimation, the next frontier is to move from prediction to mechanism, from association to dynamic synergetic inference. Multivariate learning, causal machine learning, aging clocks, longitudinal designs, and generative biophysical digital twins are beginning to provide this bridge. Brain–body–environment diversity can reveal how ecology becomes biologically embedded. We call for a future exposomic neuroscience that integrates nonlinear temporal modeling, generative mechanisms, and population diversity to understand, simulate, and modify trajectories of brain aging and disease. Congrats Sarah Genon, @MasoudTahmasian & @INM7_ISN