Network Analysis Of High-Density Microelectrode Recordings

Abstract

A high-density microelectrode arrays (HDMEA) with 3,150 electrodes per square millimetre was used to capture neuronal activity across various scales, including axons, dendrites, and networks. We present a new method for high-throughput segmentation of axons based on the spatial smoothness of signal delays. Comparison with both ground truth and receiver operator characteristics shows that the new segmentation method outperforms previous methods based on the signal-amplitude-to-noise ratio. Structural and functional neuronal network connectivity were reconstructed using a common extension of \"Peters rule\" and a inter-spike histogram method, respectively. Approximately one third of these connections are putative chemical synapses. We evaluated the spike patterns but did not find evidence for \"polychronisation\" (non-synchronous but precisely timed spike sequences). The developed framework can be used to investigate the relationship between the topology of neuronal connections and emerging temporal spike patterns observed in dissociated neuronal cultures.