Comparison of Message-Switched and Packet-Switched Communication Simulated on the Human Connectome

Makoto Fukushima, Kenji Leibnitz

Abstract: There is a growing interest in modeling the communication dynamics in the network of structural brain connections (i.e., the connectome). Recently, communication simulated on the connectome has been compared between two model assumptions about whether abstract signal units, or messages, are divided into smaller packets (packet switching) or not (message switching) when transmitted between their source and destination nodes. However, previous analyses were limited to simulations on the nonhuman connectome and did not compare any edgewise communication metrics computed under the two modeling assumptions. Here, we simulate communication on the human connectome and confirm the previous finding that packetization in the connectome improves communication speed for physiologically plausible propagation strategies. We also find characteristic differences between message switching and packet switching in an edgewise metric, the association weights of node pairs in the connectome. This study using the human connectome data corroborates previous findings with the nonhuman data and provides new comparison results for message-switched and packet-switched communication on the connectome.