Abstract
Understanding the intricate structure–function relationships of neuronal circuits is crucial for unraveling how the brain achieves efficient information transfer. In specific brain regions, like the hippocampus, neurons are organized in layers and form unidirectional connectivity, which is thought to help ensure controlled signal flow and information processing. In recent years, researchers have tried emulating these structural principles by providing cultured neurons with asymmetric environmental cues, namely microfluidics' microchannels, which promote directed axonal growth. Even though a few reports have claimed to achieve unidirectional connectivity of in vitro neuronal circuits, given the lack of functional characterization, it remains unknown if this structural connectivity correlates with functional connectivity. We have replicated and tested the performance of asymmetric microchannel designs previously reported in the literature to be successful in promoting directed axonal growth, as well as other custom variations. A new variation of 'Arrowhead', termed 'Rams', was the best-performing motif with a ∼76% probability per microchannel of allowing strictly unidirectional connections at 14 d in vitro. Importantly, we assessed the functional implications of these different asymmetric microchannel designs. For this purpose, we combined custom microfluidics with microelectrode array technology to record the electrophysiological activity of two segregated populations of hippocampal neurons ('Source' and 'Target'). This functional characterization revealed that up to ∼94% of the spiking activity recorded along microchannels with the 'Rams' motif propagates towards the 'Target' population. Moreover, our results indicate that these engineered circuits also tended to exhibit network-level synchronizations with defined directionality. Overall, this functional characterization of the structure–function relationships promoted by asymmetric microchannels has the potential to provide insights into how neuronal circuits use specific network architectures for effective computations. Moreover, the here-developed devices and approaches may be used in a wide range of applications, such as disease modeling or preclinical drug screening.