24 October 2023

Weijia WANG – PhD defense

"Analysis of collective phases in fish school models with burst-and-coast swimming"

Defense in english

  Zoom link

Team : Collective Animal Behaviour (CAB), CRCA-CBI

Supervisor : Guy THERAULAZ (CRCA-CBI) & Zhangang HAN, Beijing Normal University

Committee members :

  • Pr. Wang Dahui, Beijing Normal University, Chine (Examinateur)
  • Pr. Zhixin Liu, Academy of Mathematics and Systems Science, Chine (Rapporteure)
  • Pr. Fernando Peruani, CY Cergy Paris Université, France (Rapporteur)
  • Pr. Nicolas Destainville, Université Toulouse Paul Sabatier, France (Examinateur)
  • Pr. Zhangang Han, Beijing Normal University, Chine (Co-directeur de thèse)
  • Dr. Guy Theraulaz, CNRS, Université Toulouse Paul Sabatier, France (Co-directeur de thèse)


Collective motion in groups of organisms is a ubiquitous phenomenon observed at all biological scales. From the collective migration of cells during a immune response to the collective response of fish schools to predators attacks, researchers have investigated the mechanisms that enables the coordination of movements and the formation of ordered structures in time and space in these living systems composed of a large number of constituents. It has been well established that the coordination of movements among individuals in a moving group results from a self-organization process, in which individuals repeatedly interact with their local neighbors. On the other hand, many existing computational models, due to their excessive simplification and lack of biological relevance, still face challenges in revealing the interaction rules from which collective motion emerges in various biological systems.. In this thesis, focusing on the challenging issues mentioned above, we have investigated the interacting rules involved in the emergence of collective states in data-driven fish school models. The models describe the interactions involved in burst-and-coast swimming in groups of Hemigrammus rhodostomus. We have first investigated the impact perceptual and cognitive factors on collective states. We then investigated the long-term collective behavior of schools in burst-and-coast swimming fish.

1. In natural conditions, social interactions can be modulated by the perceptual and cognitive abilities of animals as a result of adaptive fitness to the environment. The resulting collective states of the moving group will therefore change. To quantitatively investigate the impact of perceptual and cognitive factors on collective swimming mediated by social interactions, we comprehensively analyze the phase plane of a data-driven model that characterize social interactions involved in burst-and-coast swimming in schools of Hemigrammus rhodostomus. We find that coordinated swimming patterns of schooling and milling can emerge even if fish only interact with their most or two most influential neighbors, providing the existence of a minimal level of attraction between fish to maintain cohesion. We also find that the range of social interactions, which characterizes the perceptual ability of fish, has similar effects on collective states as the strength of interactions.

2. The burst-and-coast swimming mode is widely adopted by various species of fish. Fish are sensitive to the movements of other conspecifics in the burst phase, but neglect this information during the coasting phase. Therefore, this swimming mode potentially affects the coordination between fish. However, few studies have focused on the long-term collective behaviors of fish with a burst-and-coast swimming mode. We address this question by analyzing the phase plane and the long-term behavior with a model that quantitatively describes the general pattern of burst-and-coast swimming and social interactions between fish. We find that fish can effectively coordinate their behaviors in a broad range of interaction strengths when information about only one or two most influential neighbors is perceived. In long-term, the stability of schooling and milling states depends on whether group cohesion can be maintained.

Keywords : Collective Behavior, Self-organization, data-driven modeling, fish school.

24 October 2023, 09h3012h00
Beijing Normal University, China