Nonlinear Dose-Response of Cadmium on Drosophila Larval Locomotion: Machine Learning Decodes Behavioral Complexity.

The prevalent environmental contaminant cadmium threatens ecosystems, yet the lack of high-resolution behavioral kinematics hinders assessment of cadmium neurotoxicity in ecologically critical insect larvae. This study integrated machine learning-based trajectory tracking methodologies, to meticulously quantify dose-dependent effects of cadmium on the locomotion velocity, angular velocity, directional preference, and trajectory alterations, using Drosophila larvae as a model organism. Results demonstrated that cadmium exposure not only increased the larval movement speed and the proportion of active duration but also substantially diminished the angular velocity and the duration of high angular velocity. Notably, the average speed curve among the cadmium treatment groups exhibited a "U"-shaped distribution. At a lower concentration (1 mg/L), an increase in speed and the duration of straight movement were prominent. The medium concentrations (4 and 8 mg/L) were characterized by the highest trajectory complexity and the largest individual disparities, respectively. Despite the enhanced activity at the high concentration (16 mg/L), there was a concurrent increase in movement complexity. These behavioral changes are likely related to factors such as neurotransmitter regulation, visual damage, and antioxidant mechanisms. This study not only reveals the complex effects of cadmium on the movement behavior of insects, but also provides a reference basis for the research on the neurotoxic mechanism of cadmium on organisms and the assessment of the ecological risks of heavy metal pollution.