A human-specific genetic modifier reconfigures large-scale cortical network dynamics underlying behavioral performance

Cortical expansion in the human lineage was accompanied by alterations in cortical circuit architecture, including an elaboration of cortico-cortical connectivity. Yet how these changes reshape network computation to support behavior remains poorly understood. Here we leverage a mouse model expressing SRGAP2C, a human-specific gene duplication that modifies cortical circuit development and increases cortico-cortical connectivity, to ask how this remodeled architecture shapes network dynamics during learning. As mice acquire expertise on a sensory discrimination task, SRGAP2C drives broader interhemispheric correlation, distributed task-relevant encoding, and enhanced directed influence from frontal and associative regions. Texture representations in primary sensory cortex become more separable during motor preparation, supporting improved discrimination under demanding conditions. These findings demonstrate that human-specific changes in cortical circuit architecture do not simply scale up but rather reconfigure the cortical dynamics that organize sensory-to-motor transformation in a manner that tracks behavioral performance under demanding conditions, linking genomic innovations in the Homo lineage to large-scale cortical network function.