Complex network analysis of zebrafish locomotion based on time series and visibility graphs

Zebrafish are increasingly being used as a model organism to study various biological processes, including both normal and pathological conditions. Understanding zebrafish locomotor behavior is essential for gaining insights into human movement disorders. Despite an abundance of research on zebrafish locomotion, studies utilizing time series data remain limited. In this study, we employ the visibility graph method to examine how container size influences zebrafish locomotor characteristics under normal conditions. We further characterize specific behavioral indicators under normal, panic, and intoxication conditions. Our findings highlight the effectiveness of this method in identifying the behavioral states of individual zebrafish irrespective of container size. Notably, under normal conditions, the step series of individuals in containers of varying sizes consistently exhibit a non-trivial, strongly correlated pattern. These patterns are characterized by hub nodes that display long-range correlations in their positions within the step series. For other time series, including direction-changing series under normal conditions and both step and direction-changing series under panic and intoxication conditions, the strong patterns are trivial. In these cases, hub nodes do not form motifs, and the positions of motifs within the series exhibit randomness.