Optimization of an electro-optical representation of the C. elegans connectome through neural network cluster analysis

Using C. elegans as a model organism, we present on an optimization strategy for reducing the spatial needs and power consumption in an optical connectome implementation. By means of a cluster analysis algorithm1, the interconnectivity of 279 neurons can be subdivided into 3 groups. This clustering reveals 2 independent neural populations, whose members interconnect only within their cluster-community and through a relay group of inter-cluster connections. Using this strategy, the expected spatial needs could be cut down by one fourth, thereby reducing the required light intensities by the same amount. A follow-up sub-partitioning of the individual clusters led to an additional power saving of up to 7%.