Symmetries and synchronization from whole-neural activity in the <i>Caenorhabditis elegans</i> connectome: Integration of functional and structural networks

Understanding the dynamical behavior of complex systems from their underlying network architectures is a long-standing question in complexity theory. Therefore, many metrics have been devised to extract network features like motifs, centrality, and modularity measures. It has previously been proposed that network symmetries are of particular importance since they are expected to underlie the synchronization of a system's units, which is ubiquitously observed in nervous system activity patterns. However, perfectly symmetrical structures are difficult to assess in noisy measurements of biological systems, like neuronal connectomes. Here, we devise a principled method to infer network symmetries from combined connectome and neuronal activity data. Using nervous system-wide population activity recordings of the <i>Caenorhabditis elegans</i> backward locomotor system, we infer structures in the connectome called fibration symmetries, which can explain which group of neurons synchronize their activity. Our analysis suggests functional building blocks in the animal's motor periphery, providing testable hypotheses on how descending interneuron circuits communicate with the motor periphery to control behavior. Our approach opens a door to exploring the structure-function relations in other complex systems, like the nervous systems of larger animals.