Exploring the consequences of social responsiveness on leadership using an interactive biomimetic robotic fish

个体对社会伙伴行为的反应可能始终不同。这些差异很可能会影响群体中个体的互动，例如领导力。领导力包括两个主要的功能方面，这两个方面都可能受到社会反应的个体差异的影响：（i）追随者可能会根据领导者的行为来决定跟随哪个领导者，而（ii）领导者的领导能力可能取决于其对追随者行为的反应。然而，在领导背景下的社会反应的个体差异的经验数据是罕见的。很难通过实验来测量和操纵一个人的社会反应。此外，对整个动物群体进行系统的操纵往往是不可行的。克服这些困难的一种方法是控制群体中的个人行为。因此，我将使用一个新的交互式仿生机器鱼系统（Robofish）。机器鱼被孔雀鱼（Poecilia reticulata）接受为同种，并且能够将自己融入活孔雀鱼群中。这是通过实时跟踪活动物和机器鱼来实现的，机器鱼然后能够根据活孔雀鱼的动作调整自己的行为。通过这个新工具，我试图回答以下基本问题。首先，我问孔雀鱼是否在社会反应方面一直存在差异。为了回答这个问题，我将分析活的孔雀鱼对机器鱼的反应，机器鱼在预定义的路径上游泳通过水族馆。第二，我问一个人的社会反应能力是否会影响他在不同领导者中的选择能力。我预测，一个人对社会的反应越积极，就越能在不同的领导者之间做出适应性的选择。为了测试这个预测，孔雀鱼可以在两个同时出现的Robofish领导者之间进行选择。这两个机器鱼领导者中的一个将事先被认为是通向食物来源的，而另一个则不是。越是响应个别孔雀鱼，它应该遵循的领导人，是与食物之前。第三，我问领导者的社会反应如何影响他们有效领导团队的能力。强有力的领导者往往表达自己的偏好，因此，他们自己对社会的反应较少。理论模型预测，与较弱但对社会反应更敏感的领导者相比，这些强有力的领导者更有能力领导更大的群体。然而，当领导者的响应低于特定阈值时，分裂事件应该增加。为了验证这一预测，我将操纵交互式机器鱼领导者的社会反应，并量化不同大小的孔雀鱼群体的以下行为。我的方法是新的，所提出的实验可以提供经验数据，以评估最近的几个理论模型的集体运动和领导的动物群体。

Individuals can differ consistently in how they respond to the behavior of their social partners. These differences may well affect interactions of individuals in a group, for example leadership. Leadership comprises two main functional aspects that both could be influenced by individual differences in social responsiveness: (i) a follower may base its decision on which leader to follow on the leaders´s behavior while (ii) a leader´s ability to lead may depend on its responsiveness to the behavior of the follower(s). However, empirical data on individual differences in social responsiveness in a leadership context is rare. It is difficult to measure and manipulate an individual's social responsiveness experimentally. In addition, a systematic manipulation of entire animal groups is often not feasible. One approach to overcome these difficulties is to gain control over individuals´ behavior within a group. For that reason, I will use a newly available interactive biomimetic robotic fish system (Robofish). Robofish is accepted as a conspecific by guppies (Poecilia reticulata) and is able to integrate itself into shoals of live guppies. This is achieved through real-time tracking of both live animals and the Robofish that is then able to adjust its behavior to the live guppies´ actions. With this new tool, I try to answer the following fundamental questions. First, I ask whether guppies differ consistently in their social responsiveness. To answer this question, I will analyze the response of live guppies towards a Robofish that swims on a predefined path through an aquarium. Second, I ask whether an individual´s social responsiveness affects its ability to select among different leaders. I predict that the more socially responsive an individual is, the better it should make adaptive choices between different leaders. To test this prediction, guppies can choose between two simultaneously presented Robofish leaders. One of these two Robofish leaders will have been perceived beforehand as leading to a food source while the other one was not. The more responsive the individual guppy, the more it should follow the leader that was associated with food before. Third, I ask how the social responsiveness of leaders influence their ability to lead groups effectively. Strong leaders express most often their own preferences and, thus, are less socially responsive themselves. Theoretical models predict that those strong leaders are better able to lead larger groups compared to weak but more socially responsive ones. However, splitting events should increase when responsiveness of the leader is below certain thresholds. To test this prediction, I will manipulate interactive Robofish leaders´ social responsiveness and quantify the following behavior of live guppy groups of different size. My approach is new and the proposed experiments could provide empirical data to evaluate several recent theoretical models on collective motion and leadership in animal groups.