A joint research team from the Institute of Automation, Chinese Academy of Sciences, and Peking University has proposed a novel neural network framework called CATS Net, designed to simulate the formation, understanding, and communication processes of human concepts. This framework comprises two core modules: concept abstraction and task-solving. It can compress high-dimensional visual information into low-dimensional concept vectors and dynamically adjust network activities to accomplish specific tasks. The system can autonomously generate new concepts and form a concept space, enabling knowledge transfer between networks through concept vectors. Research shows that the concept space generated by CATS Net aligns with human cognitive semantic models and the activity patterns of relevant brain regions, laying the foundation for developing next-generation intelligent systems capable of forming and applying human-like concepts. The related findings have been published in Nature Computational Science.