Abstract:Brain function is substantially linked to the highly organized modular structure of neuronal networks. However, the structure of in vitro assembled neuronal circuits often exhibits variability, complicating the consistent recording of network functional output and its correlation to network structure. Therefore, engineering neuronal structures with predefined geometry and reproducible functional features is essential to precisely model in vivo neuronal circuits. Here, we engineered microchannel devices to assemble 2D and 3D modular networks. The microchannel devices were coupled with a multi-electrode array (MEA) electrophysiology system to enable recordings from circuits. Each network consisted of 64 modules connected to their adjacent modules by micron-sized channels. Modular circuits within microchannel devices showed enhanced activity and functional connectivity traits. This includes metrics such as connection weights, clustering coefficient, global efficiency, and the number of hub neurons with higher betweenness centrality. In addition, modular networks demonstrated an increased functional modularity score compared to the randomly formed circuits. Neurons within individual modules displayed uniform network characteristics and predominantly participated in their respective functional communities within the same or neighboring physical modules. These observations highlight that the modular network structure promotes the development of segregated functional connectivity traits while simultaneously enhancing the efficiency of overall network connectivity. Our findings emphasize the significant impact of physical constraints on the activity patterns and functional organization within engineered modular networks. These circuits, characterized by stable modular architecture and intricate functional dynamics-key features of the brain networks-offer a robust in vitro model for advancing neuroscience research.

Engineered modular neuronal networks-on-chip represent structure-function relationship