Hunger and knowledge: foraging decisions in an uncertain and social world

即使在看似同质的栖息地中，也存在大量的时间和空间上的小规模异质性。因此，当动物探索它们的环境时，个体之间在遇到食物时存在差异，导致种群层面的饥饿和知识个体差异。个体之间内在一致的行为差异（通常被称为动物“个性”）往往会进一步放大这种差异，这种差异可能是由遗传差异和成年后稳定的发育差异驱动的。个体如何在信息有限的情况下平衡进食和避免被捕食的风险是理解动物行为的一个长期存在的问题。当我们考虑到许多动物，特别是那些被其他动物捕食的动物，生活在群体中时，这个故事就变得更加复杂了。生活在一起对躲避捕食者和寻找资源有好处，部分原因是动物可以分享信息并相互学习；另一方面，协调何时何地执行特定行为成为一个新问题。信息何时以及如何获得和社会传播，以及饥饿如何影响决策和社会行为，已经进行了广泛的研究，但彼此是分开的。在这个项目中，我们将首次把这些因素结合在一起，以了解当动物单独和集体做出觅食决定时，饥饿和信息是如何相互作用的。我们的研究将利用最近对跨分类群动物群体决策的兴趣激增，以及最先进的动物跟踪计算机视觉软件的进步。我们将讨论一系列关于饥饿和信息在决策中的悬而未决的问题：在获取信息方面，饥饿和冒险倾向的相对贡献和相互作用是什么？群体内个体之间的相互作用如何影响这些过程？知识和食物如何在个体之间分配的后果是什么？在一个群体中，将消息灵通和naïve，饥饿和饱足的个体混合在一起会产生什么影响？例如，naïve个体是否会跟随并更好地从饱足的个体那里学习，而不是从可能会把他们甩在后面的饥饿的信息灵通个体那里学习？领导力能否从由社会互动驱动的饥饿差异中产生？群体中更有效的觅食行为是通过分享信息来解释的，还是仅仅是因为有一个伙伴的额外动机？这个项目有可能大大增加我们对决策的理解，以及对社会动物个体差异的重要因素的理解。实验工作将以小型淡水鱼为模型系统。与研究鸟类或灵长类动物等更复杂的动物相比，研究鱼类有很多优势。信息和饥饿可以很容易地控制，并且要使用的物种（三棘棘鱼和孔雀鱼）具有“裂变融合”的社会系统，其中个体经常改变群体成员，使我们能够轻松地操纵群体组成。由于决策和信息交换在大多数鱼类的速度和方向变化中是明显的，最先进的计算机视觉软件可以跟踪每个个体的位置，并将其用于客观地量化个体的决策。此外，鱼类表现出许多“更聪明”的脊椎动物的认知能力，这表明我们的研究结果可能适用于其他动物，甚至是人类。在行为实验的同时，我们将创建新的动物群体模型，考虑当个体能够了解其环境时，决策是如何发生的。在实验工作的同时开发模型将最大限度地提高我们充分理解真实动物中预期的复杂相互作用的能力。

Even within seemingly homogeneous habitats there is a great deal of small-scale heterogeneity over time and space. Thus, as animals explore their environment, there is variation between individuals in encountering food, resulting in population-level individual differences in hunger and knowledge. This variation is often further magnified by intrinsic consistent differences in behaviour between individuals (frequently referred to as animal 'personality'), which can be driven by differences in genetics and development that are stable in adulthood. How individuals manage the balance between feeding and avoiding predation risk when they have limited information is a long-running question in understanding animal behaviour. The story becomes much more complicated when we consider that many animals, especially those that are prey to others, live in groups. Living together has benefits in avoiding predators and locating resources, in part because animals can share information and socially learn from one another; on the other hand, coordinating when and where to do particular behaviours becomes a new problem. When and how information is gained and socially transmitted, and how hunger affects decisions and social behaviour, have been studied extensively but separately from one another. In this project, for the first time we will bring these elements together to understand how hunger and information interact when animals make foraging decisions both individually and collectively.Our study will utilise the recent explosion of interest in decision making by animal groups across taxa and advances in state-of-the-art animal tracking computer vision software. We will address a range of unanswered questions regarding hunger and information in decision making: What are the relative contributions of, and interactions between, hunger and risk taking tendency in the acquisition of information? How do interactions between individuals within groups influence these processes and what is the consequence for how knowledge and food are distributed between individuals? What is the effect of mixing informed and naïve, and hungry and sated, individuals in a group, e.g. do naïve individuals follow and learn better from satiated rather than hungry informed individuals that are likely to leave them behind? Can leadership emerge from differences in hunger that are driven by social interactions? Can more effective foraging in groups be best explained by sharing of information or simply by the extra motivation of having a partner? The project has the potential to greatly increase our understanding of decision making and the factors important in individual variation in social animals. The experimental work will use small freshwater fish as a model system. Compared to studying more complex animals such as birds or primates, there are a number of advantages to studying fish. Information and hunger can be easily controlled, and the species to be used (three-spined sticklebacks and guppies) have 'fission-fusion' social systems where individuals change group membership often, allowing us to easily manipulate group composition. As decisions and information exchange are evident in changes in speed and direction in most fish species, state-of-the-art computer vision software that tracks each individual's position over time will be used to quantify individual decisions objectively. Moreover, fish show many of the cognitive abilities found in 'smarter' vertebrates, suggesting that our results may be applicable to other animals, even humans. In parallel with the behavioural experiments, we will create new models of animal groups that consider how decision-making occurs when individuals are able to learn about their environment. Developing the models alongside the experimental work will maximise our ability to fully understand the complex interactions expected in the real animals.