Controlled generation of functional human neural circuits

Biological neural networks show an inherent structure and connections that are reproduced across individuals. It is still largely unknown how the structure of such networks is correlated with their observed functional dynamics and computational output. Detailed investigation of network structure and its functional implications is challenging. For instance, in vivo studies do not allow for isolated analysis of network motifs. As basic elements of neuronal systems, these motifs need careful consideration. In contrast, the main obstacle in vitro is to create defined network structures down to the single cell level. This lack of control is hindering reproducible experimental research on small-scale circuits. Thus, precise study of functional features inherent to structural motifs is hardly possible. Here, we present a method to generate predefined 2D neural networks by placing neurons in a scaffold atop a multielectrode array (MEA). We have built a setup for printing single human induced pluripotent stem cell (hiPSC)-derived neurons with µm precision. Microchannels help to guide neurites growth direction and generate predefined neuronal connections. It is possible to incorporate multiple neuronal cell types in one circuit. Our approach can be extended to incorporate precise stimulation of neurons by MEA electrodes or optogenetic actuators. Using our method, we engineer first and second order network motifs which are abundant in biological neural systems like the retina. Advanced control over neuronal network formation facilitates to compare computational predictions to experimental results and might therefore help refining existing models. Insights into structure-function relationships could potentially aid a more detailed understanding of large neuronal systems.