According to the hypernetwork theory of the brain, repeatedly processed signals from the external world converge in the deep layers of the neural network, forming cognitive clusters. These "atoms" of intelligence emerge not as a superstructure on top of the existing neural network, but by transforming it from the ground up, until individual neurons develop cognitive properties. How does this happen? What tools help experimentally study deep cognitive processes in the living brain? How can we make the brain optically transparent and reveal its internal cognitive architecture?