We aimed to investigate the principle of emerging interactions between swarms using the functional differentiation theory of the brain. We propose a heterogeneous swarms model, where two swarms having different parameters evolve to maximize transfer entropy between them. In our simulation, we found the emergence of heterogeneous behavior among the swarms, and the appearance of several interaction patterns depending on the degree of the transfer entropy. Our results imply that the same principle of functional differentiation may underlie both the brain and swarms, leading to a novel design of brain-inspired swarm intelligence.