Functional Complexity of Engineered Neural Networks Self-Organized on Structured 3D Interfaces

Abstract

Engineered neural networks are indispensable tools for studying neural function and dysfunction in controlled microenvironments. In vitro, neurons self-organize into complex assemblies with structural and functional features reminiscent to those observed for in vivo circuits. Traditionally, these models are established on planar interfaces, but studies suggest that the lack of a 3D growth space affects neuronal organization and function. While methods supporting 3D growth exist, reproducible 3D neuroengineering techniques compatible with electrophysiological recording methods are still needed. In this study, a reproducible biocompatible interface made of the polymer SU-8 to support 3D network development is developed. Using electron microscopy and immunocytochemistry, it is shown that neurons utilize these 3D interfaces to self-assemble into complex, multi-layered 3D networks. Furthermore, interfacing the 3D structures with custom microelectrode arrays enables characterizing of electrophysiological activity. Both planar control networks and 3D networks display complex interactions with integrated and segregated functional dynamics. However, control networks show stronger functional interconnections, higher entropy, and increased firing rates. In summary, the interfaces provide a versatile approach for supporting neural networks with a 3D growth environment, compatible with assorted electrophysiology and imaging techniques. This system can offer new insights into the impact of 3D topologies on neural network organization and function.