Biodegradable Plastic Particles Slow Larval Movement in Fruit Flies, AI Tracking Reveals

Biodegradable plastics have long been marketed as the environmentally friendly answer to the world’s plastic crisis, but a new study suggests that even these greener polymers may carry hidden costs for living organisms. Researchers Feiyue Feng and Jie Shen of Hangzhou Dianzi University in China have found that microscopic particles of polyhydroxyalkanoates, or PHAs, a family of microbially synthesized polyesters touted as sustainable alternatives to conventional plastics, can measurably disrupt the movement of fruit fly larvae. The work, published in the journal Ecotoxicology, is among the first to apply deep-learning pose estimation to quantify how biodegradable microplastics alter the fine structure of behavior in a terrestrial insect, and it raises uncomfortable questions about whether bioplastics are as benign as their reputation implies.

The research team focused on third-instar Drosophila melanogaster larvae, the maggot-like stage of one of biology’s most intensively studied model organisms. Larvae were exposed for 96 hours through their diet to 5-micrometer PHA particles at three concentrations: 0.25, 2.5, and 7.5 milligrams per liter. These doses were deliberately sublethal, meaning the larvae survived the exposure. The question was not whether the particles killed the animals, but whether they changed how the animals behaved, a subtler class of toxicity that ecologists increasingly recognize as critically important for survival in the wild.

To capture those behavioral changes with precision, the researchers turned to SLEAP, a deep-learning system for multi-animal pose tracking originally developed by computational biologists and published in Nature Methods. Rather than relying on human observers to eyeball crawling larvae, the system uses neural networks to track body positions frame by frame, converting hours of video into quantitative trajectories. From these trajectories, the team extracted speed and angular-speed interval distributions along with nine spatial features, including total distance travelled, displacement, path straightness, convex-hull perimeter and area, and radius of gyration, a measure of how far an animal roams from the center of its activity.

The results were striking at the highest dose. Compared with unexposed control larvae, animals exposed to 7.5 milligrams per liter of PHA particles crawled 37.8 percent more slowly on average and covered 40.7 percent less distance. The distribution of speeds shifted in a telling way: larvae spent more time in the low-speed category and less time in the medium- and high-speed categories. In other words, exposed larvae were not simply moving at a uniformly reduced pace; they were spending a larger fraction of their time nearly stationary, a pattern consistent with impaired or disorganized locomotor function.

The spatial features told a complementary story. Exposed larvae showed reduced displacement, meaning they ended up farther from where their normal trajectories would have carried them, and reduced path straightness, indicating more meandering, less directed movement. Their convex-hull perimeter and area, which describe the spatial footprint of their wandering, shrank, as did their radius of gyration. Together these measures paint a picture of larvae confined to smaller, less purposeful territories, exploring less of their environment than their unexposed counterparts.

Intriguingly, angular speed behaved differently from linear speed. The occupancy of angular-speed categories shifted from the low category toward the medium and high categories, suggesting exposed larvae made more frequent or sharper turns, yet the mean angular speed itself did not differ significantly among the treatment groups. This dissociation between linear and rotational movement is exactly the kind of nuance that only high-resolution, machine-learning-based tracking can reveal, and it hints that PHA particles may affect the neural or muscular control of crawling in ways that a simple speed measurement would miss entirely.

Perhaps most concerning from an environmental standpoint, the study detected effects even at the lowest concentration tested. Changes in the proportion of time spent at low speeds and low angular speeds were evident at 0.25 milligrams per liter, the mildest exposure in the experiment. While the broadest and most severe response occurred at 7.5 milligrams per liter, the fact that subtle alterations appeared at the low end of the dose range suggests that the behavioral effects of biodegradable microplastics may begin at concentrations lower than those typically flagged as hazardous in conventional toxicity testing.

Why does larval crawling matter? Locomotion is not a luxury for a fruit fly larva; it is the engine of survival. Larvae rely on crawling to find food, occupy suitable habitats, disperse across their environment, and escape predators and other threats. A larva that moves nearly 40 percent less distance and wanders less effectively is a larva that feeds less efficiently, disperses less successfully, and may be more vulnerable to predation. The authors argue that these coordinated changes constitute an ecologically relevant effect, one that could translate into real consequences for insect populations even when no individual animal dies. Behavioral ecotoxicology has increasingly emphasized that sublethal effects on movement, foraging, and escape responses can ripple through food webs in ways that lethality assays never capture.

The study arrives at a moment of rapid growth for the bioplastics industry. European Bioplastics market updates report continued expansion in biodegradable polymer production, and PHAs in particular have been celebrated as the greenest plastics so far, because bacteria synthesize them naturally and soil microbes can break them down. But the new findings join a growing body of evidence that biodegradable does not mean harmless. Previous work has documented that incomplete degradation can leave PHA-derived particles in soil and water, that secondary nanoplastics released from biodegradable microplastics can severely impact freshwater environments, and that poly-3-hydroxybutyrate, a common PHA, can exert acute, chronic, and multigenerational effects on the water flea Daphnia magna. Related studies have shown hazardous effects of polylactic acid nanoplastics in Drosophila, another widely used bioplastic.

The broader lesson is that the environmental fate of biodegradable plastics deserves the same scrutiny applied to conventional polymers. Degradation is a process, not an instant, and during the window between release and full breakdown, bioplastic particles behave much like any other microplastic: small, persistent, and available for ingestion by organisms at every trophic level. Feng and Shen’s findings show that biodegradable-polymer particles can elicit sublethal effects in terrestrial insects, a group that has received far less attention than aquatic organisms in microplastics research, and the authors argue that such particles warrant inclusion in ecological hazard assessment. As bioplastics flood into packaging, agriculture, and consumer products, the study serves as a timely reminder that solving the plastic problem requires not just swapping materials, but rigorously testing what those new materials do once they fragment into the environment, one crawling larva at a time.

Subject of Research: Sublethal behavioral toxicity of biodegradable polyhydroxyalkanoate microplastics in Drosophila melanogaster larvae

Article Title: Behavioral toxicity of polyhydroxyalkanoate microplastics in Drosophila melanogaster larvae revealed by deep-learning pose estimation

Article References: Feng, F., & Shen, J. (2026). Behavioral toxicity of polyhydroxyalkanoate microplastics in Drosophila melanogaster larvae revealed by deep-learning pose estimation. Ecotoxicology, 35(8), Article 179. https://doi.org/10.1007/s10646-026-03162-0

Image Credits: AI Generated

DOI: 10.1007/s10646-026-03162-0

Keywords: polyhydroxyalkanoates, biodegradable plastics, microplastics, Drosophila melanogaster, locomotor behavior, SLEAP, deep learning, pose estimation, ecotoxicology, sublethal effects, behavioral toxicity, terrestrial insects