The knowledge of the intuitive link between muscle activity and the finger movement is vital for the design of commercial prosthetic hands that do not need user pre-training. However, this link has yet to be understood due to the complexity of human hand. In this study, a relation spectrum using a controllable and human-readable polynomial neural network (CR-PNN) was developed for the first time and applied to analyze the muscle-finger system. We established CR-PNN models for six degrees of freedom ( DOFs) in 8 subjects. Multiple fingers may be controlled by a single muscle, or multiple muscles may control a single finger. Thus, the research is based on two aspects: muscle synergy and muscle coupling for hand. The research gave the relation spectrum of the muscle-finger system and the knowledge of muscle coupling. The article is very short but significant. The contributions of this paper can be divided into two parts: (1) The findings of hand can contribute to design prosthetic hands. (2) The relation spectrum using CR-PNN in the method can provide a reference for analyzing complex systems in multiple areas. (We're strong believers in Open Source, and provide CR-PNN code for others. GitHub: https://github.com/liugang1234567/CR-PNN#cr-pnn. And you can find the relation spectrum in" Read C-R PNN for System identification_LG.pdf" )