Multivariate Functional Data Analysis Uncovers Behavioral Fingerprints in Invertebrate Locomotor Response to Micropollutants

The need for effective biomonitoring in wastewater has become clear due to the impracticality of continuously tracking all chemicals and emerging contaminants in the aquatic exposome. Effect-based biomonitoring provides a cost-effective solution. The ToxMate device, which uses videotracking of locomotor behavior in aquatic invertebrates, has proven efficient for real-time detection of micropollutant surges in effluents. To extend the approach, this proof-of-concept study evaluates the potential to formalize behavioral fingerprints from real-time videotracking data to characterize qualitative variations in effluent contamination. We present the first application of a functional data analysis (FDA) framework in ecotoxicology. Data were obtained by simultaneously tracking three sentinel organisms from distinct taxa (a crustacean, an annelid, and a gastropod) during pulse exposures to four chemicals in the laboratory (two metals, one pharmaceutical, and one insecticide). Individual and multispecies responses were analyzed to determine whether combining species enhances the resolution of contamination fingerprints through multidimensional FDA. Applying the same data-driven approach to field data from a wastewater treatment plant (WWTP) revealed four recurring types of micropollution events. This proof of concept demonstrates the potential of behavioral fingerprints to improve wastewater monitoring and reduce pollutant transfer to the environment.

由于无法持续追踪水生暴露组中的所有化学品与新兴污染物，对污水开展有效生物监测的必要性已愈发明确。基于效应的生物监测（effect-based biomonitoring）则提供了一种经济高效的解决路径。ToxMate设备通过对水生无脊椎动物的运动行为进行视频追踪，已被证实可高效实现对污水出水中微污染物突增的实时检测。为拓展该方法的应用场景，本项概念验证研究探索了通过实时视频追踪数据构建标准化行为指纹的可行性，以表征污水污染物的定性变化特征。本研究首次将功能数据分析（Functional Data Analysis, FDA）框架应用于生态毒理学领域。研究通过实验室条件下的脉冲暴露实验获取数据：对来自3个不同分类群的指示生物（1种甲壳类、1种环节动物与1种腹足类动物）同时开展追踪，并设置4种污染物暴露（2种金属、1种药物污染物与1种杀虫剂）。本研究通过分析单一物种与多物种的响应特征，验证了通过多维功能数据分析（FDA）整合多物种数据，可提升污染物指纹的识别分辨率。将该数据驱动方法应用于某污水处理厂（WWTP）的现场监测数据后，研究识别出4种反复出现的微污染事件类型。本项概念验证研究证实，行为指纹可有效优化污水监测流程，并减少污染物向环境中的迁移扩散。