EVOLUTION-GUIDED PARCELLATION MAPS SOCIO-AFFECTIVE BRAIN ADAPTATION

Recent human brain evolution is linked to social environment changes like population density, language, and emotional complexity, but direct genetic studies are impossible without ancestral brain tissue. Ancient genomes and multimodal brain resources enable in silico analyses, yet integrated datasets linking molecular changes to cognition remain scarce.
Here, we introduce a novel workflow emphasizing socio-affective traits, mapping multigenic evolutionary signals to cognitive domains, functional networks, brain cell types, and molecular mechanisms. Unlike prior whole-brain approaches, we apply macro-scale brain parcellation guided by evolutionary rate rather than classical functional divisions, complementing micro-scale biopsy data.
We integrate expanded genetic data (including extinct hominins) with synthetic task-fMRI networks, first pre-clustering imaging data into concise psychological domains for interpretability. The evolution-guided parcellation is then applied, followed by a Genetic Algorithm for Generalized Biclustering (GABi) to mine multigenic evolution across brain space, identifying biclusters in socio-affective networks probed for cell-type signatures and Gene Ontology enrichment.
Peaks of adaptive selection emerge in networks for social interaction (language) and concepts (theory of mind), spanning hominid to modern human ancestry. Signatures involve glutamatergic/GABAergic neurons and non-neuronal cells; enrichments highlight cell signaling, synapses, and morphology—tuning circuits for social function.
This "computational archaeology" bridges genomics and neuroscience, with applications to neuropsychiatric traits or cross-species evolution.