A new method for evaluating the complexity of animal behavioral patterns based on the notion of Kolmogorov complexity, with ants' hunting behavior as an example

摘要

We suggest a method for evaluating the complexity of animal behavioral patterns based on the notion of Kolmogorov complexity, with ants' hunting behavior as an example. We compared complete and incomplete hunting stereotypes in members of a natural ant colony and in naive laboratory-reared ants. We represent behavioral sequences as "texts", and compress them using a data compressor. Behavioral units (10 in total), singled out from video records and denoted by letters, served as an alphabet. Successful hunting stereotypes appeared to be characterized by smaller complexity than incomplete ones. A few naive "born hunters" which enjoy "at once and entirely" complete hunting stereotypes are characterized by a lower level of complexity of hunting behavior. We conclude that innate complete stereotypes have less redundancy and are more predictable, and thus less complex. We suggest that this method for evaluating the complexity of behavioral "texts" can serve different purposes, from estimating behavioral variability within populations of animals to comparative analysis of neuronal assemblies within the brain. The method can also be applied to distinguishing between initial and transformed behavioral patterns in many fields of biology and medicine, including studying and diagnosis of neurological diseases reflected in the behavior.