In a data communication network, the nodes can be organized or arranged in various ways. For small networks, there are some typical arrangements of nodes. In this paper we discuss how the standard butterfly network, a network of six nodes, might be implemented in the nervous system by means of current-mediated coupling between axons. The implementation is in the context of the well-known network coding gains. Specifically, we show how two current-coupled axons can exchange information in a `contact-less' manner. Our strategy provides throughput-enhancement, that is, a network coding gain, resulting from the synchronization of action potentials. We therefore move incrementally towards a better understanding of networks of neurons and how information is transmitted by them and processed in them.