Optimal decision-making in a school of fish uses a flexible system based on role sharing, researchers find

Theoretical biology states that in a flock of birds or school of fish exhibiting collaborative behavior, being in a state of criticality (i.e., a “flexible system” that can changeably respond to the environment) is crucial for flexible and quick decision-making. However, the reason why such criticality is relatively easy to achieve in animal groups has not been elucidated.

Using integrated information theory, researchers at University of Tsukuba investigated heterogeneous criticality of a 10-fish school of Plecoglossus altivelis (Ayu). They discovered that the role sharing between supercriticality and subcriticality subgroups leads to a state of heterogeneous criticality for the entire group more easily than theoretical predictions.

For this study, published in Scientific Reports, researchers quantified the degree of criticality and assigned that value to all subsets. They discovered that the distribution of the degree of criticality was nonuniform within the group. However, averaging over this distribution confirmed the traditional criticality status of the group as a whole.

These findings indicate that the “flexible system” that leads to decision-making includes role sharing between individuals who overreact to surrounding stimuli (supercriticality subgroups) and those who are less responsive (subcriticality subgroups). Such role sharing leads to group-wide critical states much more quickly than theoretical predictions. This information structure was not observed in several theoretically constructed models.

Researchers also observed that the subcriticality subgroups function as hubs in communicating information for the entire group. Thus, the choices of the entire group are reduced, and the decision-making of the supercriticality subgroups is smoothly integrated.

These findings indicate that critical states in animal groups consist of heterogeneous structures, and role sharing among individuals is critical for overall optimized decision-making.