Every behavior of an animal is based on the interaction of many nerve cells in the brain, which form a close-meshed web called a neuronal network. However, what happens in the neuronal networks during particular behaviors is not easy for researchers to study—they have to describe the sequences of recurring postures and link them directly to the neuronal processes. Researchers at the University of Bonn have now presented a new method that is expected to improve this. DeepLabStream is the name of the system that uses artificial intelligence to estimate the posture and behavioral expressions of mice in real time and can respond to them immediately. Light beams are used to label and even influence specific neuronal populations in the brain in order to determine the causality between the neuronal networks and the observed behavior. The study has been published in the journal Communications Biology.