A PARETO TRADE-OFF BETWEEN TRACKING ACCURACY AND ACTIVE SENSING IN WEAKLY ELECTRIC FISH

Goal-directed behavior often requires a trade-off between stable movements that improve task performance and active movements that improve sensory sampling. This conflict is especially clear when sensing depends on self-motion, because perfect stability can reduce the sensory feedback needed for accurate control. Here we test how animals adjust this balance under different levels of sensory uncertainty using a longitudinal refuge-tracking task in weakly electric fish. Fish were required to match the motion of a moving refuge to remain inside it. We tested two species with different sensory specializations (Apteronotus albifrons and Eigenmannia virescens) across 54 conditions that independently varied illumination, water conductivity, and refuge geometry. To compare these manipulations within a single framework, we computed a Salience Index that summarizes the total sensory information available in each condition. Across both species and all conditions, behavior followed a consistent Pareto trade-off between tracking accuracy and active sensing. When salience was low, improved tracking precision was achieved only by increasing movement-based sensing. When salience was high, fish maintained near-perfect tracking while reducing movement. The combined dataset formed a common Pareto frontier, indicating that individuals with different sensory specializations converged on the same accuracy–effort boundary. We further modeled behavior as minimizing a cost function that penalizes tracking error and energetic effort. The observed strategies were consistent with optimal control under limited sensory information. Together, these results show that refuge tracking is constrained by a systematic accuracy–effort trade-off that can be captured by a single salience metric and an explicit control objective.