Group behaviours in fish are widespread, and are increasingly used to develop new optimisation algorithms for solving complex tasks. However, how external factors regulate individual spatial location in a fish school represents an overlooked aspect that is fundamental to understand collective dynamics. Herein, the neon tetra Paracheirodon innesi, a highly social fish species, has been used to analyse the effect of different ecological contexts, including food availability and predation risk, on the place preference of individual agents in the group. A robotic school of fish has been developed to establish a biohybrid interaction enabling accurate isolation of the individual choice of living fish without disturbance from other conspecifics. Results show that the individual optimal position in the group is context dependent, as individual fish continuously modify their location in the school according to different external scenarios. Individual fish stayed as close to food as possible to better compete with conspecifics. In the presence of a robotic predator, fish preferred to locate in the middle of the aggregation to favour dilution effects. When no stimuli were provided, fish preferred to stay at the back of the school, probably to exploit the increased likelihood of predator detection by conspecifics swimming ahead. This study provides new basic knowledge on how swarm intelligence and self-organisation processes of fish are affected by the individual agent. Furthermore, these findings can inspire innovative task optimisation approaches potentially applicable in engineering contexts.

鱼中的群体行为很普遍，并且越来越多地用于开发新的优化算法来解决复杂的任务。但是，外部因素如何调节鱼群中的单个空间位置代表了一个被忽略的方面，这对于理解集体动态至关重要。在此，霓虹灯四Paracheirodon innesi，一种高度社会化的鱼类，已被用来分析不同生态环境（包括食物供应和捕食风险）对个体中个体所处位置的影响。已经开发了一种鱼类机器人学校，以建立生物杂交相互作用，从而能够准确隔离活鱼的个体选择，而不会受到其他物种的干扰。结果表明，个体鱼群中的最佳位置取决于环境，因为个体鱼群会根据不同的外部场景不断修改其在学校中的位置。个别鱼类应尽可能靠近食物，以更好地与种竞争。在存在机器人食肉动物的情况下，鱼优选位于聚集体的中间，以利于稀释作用。如果没有刺激，鱼类更喜欢留在学校的后面，这可能是为了利用特定物种在前面游动而捕食者的可能性增加。这项研究提供了有关个体群体如何影响鱼类群智能和自组织过程的新基础知识。此外，这些发现还可以激发创新的任务优化方法，这些方法可能适用于工程环境。