Similarity and Cluster Analysis of the C. ELEGANS Neural Network

(with K. Oshio)

We classify theoretically the 39 sensory neuron classes of the C. elegans by the analysis of the structure of C. elegans neural network.
In the analysis, we first use a random walker model and the connectivity database of the real C. elegans in order to investigate the accessibility from each sensory neuron to other neurons. Next, we classify the sensory neurons by applying cluster analysis to the accessibility data with regard to the sensory neurons.
In the above analysis, we consider the data only about the connectivity, not even about the inhibitory / excitatory nature of the synapses; nevertheless, the result of the analysis matches quite well with the experimental results regarding the functions of the sensory neurons. This implies that the topological nature of the structure of the network itself plays an important role in the information processing of the C. elegans neural network.