Connecting the Connectome to Neural Activity

Understanding how neural networks learn complex functions involves studying the dynamic interplay of the continuous change of their connections. Cultured neurons, maintained outside a living organism, produce complex dynamics similar to the activity seen in in vivo brain tissues. Recent advancements in connectome imaging allow large-scale synaptic reconstructions, adaptable to cultured tissue, providing access to the ground truth synaptic network. We aim to map functional dynamics onto structural connectivity by comparing inferred connectivity from dynamical modeling to the actual network structure. Our approach combines various models at different levels of detail to assess scalability and accuracy. Additionally, we investigate the relationship between experimentally accessible collective behavior and the underlying network properties, aiming to gain insight into intrinsic neural dynamics observable in experimental settings. This work bridges functional and structural perspectives, advancing our understanding of the dynamics and statistical properties of neural systems.