Inference of combinatorial neuronal synchrony with Bayesian networks.

Various methods have been used to infer functional synchrony between neuronal channels using electrode signal recordings. Such methods vary from approaches to identify the groups of neuronal channels that show similar signal patterns to approaches to figure out connectivity between neuronal channels. The inference of detailed connectivity between neuronal channels from electrode signal recordings can be computationally more complex than identifying the groups of neuronal channels. For this reason, most of previous approaches to infer connectivity between neuronal channels were based on pairwise measures. In this work, we propose the degree of combinatorial synchrony (DoCS) based on Bayesian networks for more enhanced inference of neuronal synchrony. DoCS is evaluated as the likelihood of edge connections in Bayesian network structures that capture the combinatorial dependency between neuronal channels. From the comparison with a cross-correlation measure using artificial neuronal networks, we validate that the proposed DoCS shows more accurate inference of neuronal synchrony when target neuronal networks include combinatorial synchrony.