Graph theoretic characterization of in vitro neuronal network development

Multielectrode arrays (MEA) are routinely used to observe the development of neuronal networks in vitro, making simultaneous recordings of small proximal subpopulation activity within a cell culture. We propose the use of graph theoretic measures to deduce behavioral properties of neuronal networks as they develop in culture. We utilize 878 recordings from embryonic rat cortex cell cultures collected from 60-electrode, grid-type (200 μm) MEA’s [3]. We modeled each recording as a directed weighted graph with weights describing electrode connectivity. For each recording, we calculated each electrode’s clustering coefficient [1]. We averaged them over a sliding 5-day window, interpolating as needed and generating ~1700 features that describe each culture’s