Self-organization and evolution of structure and function in cultured neuronal networks

Abstract

The study of cultured neuronal networks (CNNs) has recently gained significant relevance as an alternative to in vivo models. CNNs provide a simplified version of the central nervous system while still exhibiting complex selforganization. This in vitro approach has proven invaluable in understanding the intricate relationship between structure and dynamics in neuronal networks and how it evolves during the developmental process. However, our understanding is often hindered by the challenges associated with resolving the intricate structure of neural networks. To overcome these limitations, our experimental approach allows for the simultaneous study of the detailed structure and dynamics of the cultured networks. This enables us to compare the topological and functional networks throughout the formation process, providing valuable insights into the interplay between structure and function in neuronal networks. We cultured neuronal networks from initially isolated neurons extracted from the locust species Schistocerca gregaria for a duration of two to three weeks. Throughout this period, we closely monitored the growth and development of these networks, allowing us to characterize the time evolution of their structural connectivity. In Chapter 4, we focused on CNNs developing in Petri dishes, which exhibited a self-organized complex circuitry characterized by a small-world topology. This topology was characterized by abundant neuronal loops (high clustering) and short distances between nodes. As a first step to further investigate the relationship between structure and dynamics, we employed a biophysically plausible dynamical model of a neuron known as the Morris-Lecar model. This model was simulated on top of the experimental network’s nodes, allowing us to emulate the neuronal dynamics embedded in a network. By analyzing the statistical complexity of each individual node’s dynamics, we discovered an intriguing anti-correlation with their degree of centrality in weak coupling regimes, where nodes with higher degrees exhibited lower complexity levels in their dynamics. These findings have significant implications, suggesting that it is possible to infer the network’s degree distribution solely from individual dynamical measurements. This provides a new insight into the relationship between network structure and dynamics, highlighting the potential for understanding the functional properties of CNNs based on their individual node dynamics. In Chapter 5, we are in a condition to perform a fully experimental analysis of the relationship between the structural and functional network. To conduct this study, we cultured the neuronal networks on top of microelectrode arrays (MEA), enabling us to simultaneously record the electrophysiological signals and capture the detailed anatomical neuronal circuitry through microphotography at the single-link level. The structural networks exhibit features consistent with CNNs in Petri dishes. Meanwhile, in the dynamics we observed a peak in synchronization events among electrodes, coinciding with the spatial percolation of the neuronal network. At this stage, we found that the correlation between the structural and corresponding functional networks was the strongest, with approximately 15% of the physical links connecting electrodes firing synchronously. As the culture matured, we observed a gradual decline in this correlation, although it became more statistically significant. Interestingly, we discovered that the electrodes more involved in coherent dynamics, the functional hubs, were not hubs in the physical circuitry but nodes with an average degree. Finally, in Chapter 6, we present a state-of-the-art methodology to investigate the growth of cultured neuronal networks and monitor their structural properties using a neuron-on-a-chip device. This innovative device comprises a chamber fabricated from polydimethylsiloxane (PDMS), vinyl, and glass. The chamber is connected to a microfluidic platform that enables the continuous perfusion of the culture medium. We compared the growth of CNNs in the chip to controls cultured in Petri dishes, showing similar structural features, including the number of connected neurons, the clustering coefficient, and the shortest path between any pair of neurons throughout the culture’s lifespan. However, our results demonstrated that the neuronal circuits on the chip have a more stable network structure and longer lifespan than those developing in conventional settings. This suggests that the neuron-on-a-chip setup provides an advantageous alternative to current culture methods for studying neuronal networks. Furthermore, this technology holds promising applications, such as batch drug testing of in vitro cell culture models. From an engineering perspective, one of the notable advantages of this device is the flexibility to develop custom designs more efficiently than other microfluidic systems. This opens up new possibilities for tailoring the chip to specific experimental requirements and further advancing our understanding of neuronal network dynamics. Our studies offer new insights into the relationship between the structural and functional networks of self-organized neuronal systems, as well as the strengths and limitations of functional measures as proxies for the underlying structural networks. Our findings open up new paths for future investigations to elucidate the intricate interplay between structure and function in other scenarios using diverse methodologies and tools to study CNNs.