Deep Learning Unveils Brain Evolution Secrets Human Speech Gene Identified

A study in *Science* reveals insights into genetic switches defining brain cell types across species using deep learning. Researchers analyzed regulatory codes in human, mouse, and chicken brains, finding conserved and divergent codes between birds and mammals over 320 million years. These models may aid cancer and brain disorder studies. Separately, *Nature Communications* reports the identification of a human-specific NOVA1 protein variant potentially linked to the origins of spoken language. Experiments replacing the mouse NOVA1 protein with the human variant altered vocalizations. This research may lead to early detection of speech and language needs.