Swarming behavior is a remarkable phenomenon observed across various biological systems, from schools of fish to flocks of birds. When viewed as information processing, these collective systems can be regarded as a “liquid brain” in contrast to the “solid brain” of traditional neural networks (Sole et al., 2019), which have fixed structures. In a liquid ´ brain, the nodes (agents) and their connections are fluid and reconfigurable. We propose that, in such liquid brain systems, plasticity can be effectively modeled by embedding a social network within the swarm. This network underpins the interactions among agents, allowing the collective behavior to change dynamically without altering the intrinsic properties of individual agents. Swarming behavior, typically described by simple local interactions, usually relies on proximate interactions in models like Boids (Reynolds, 1987). However, by integrating a social network among agents, we can enhance its capability for collective memory and adaptability. While some studies have incorporated networks into Boids (Tang et al., 2018), none have addressed social network in Boids in terms of plasticity, and there is a lack of research on dynamic networking in Boids. This study presents a model where, despite fixed individual Boids parameters, the inclusion of a social network endows the swarm with plasticity, leading to observable changes in behavior.
