In recent years, behavior-related endpoints have been proposed as rapid and reliable ecotoxicological tools for risk assessment. In particular, the use of detritivores to test the toxicity of pollutants through feeding is currently becoming a well-known method. Experiments combining feeding with other behavioral endpoints can provide relevant information about direct and indirect toxicological effects of chemicals. We carried out a feeding experiment with the shredder Gammarus pulex in order to detect indirect (through leaf conditioning) and direct effects (through water exposure) of two pollutants at environmentally relevant concentrations: the fungicide prochloraz (6 mg/L) and the antidepressant fluoxetine (100 ng/L). Prochloraz inhibited fungal growth on leaves, but it did not affect either the microbial breakdown rates or the C:N ratio of the leaves. Individuals of G. pulex that were fed with treated leaves presented lower consumption rates, not only those fed with prochloraz-treated leaves, but also those fed with fluoxetine-treated leaves, and those fed with the mixture-treated leaves. Mixed-effects models revealed that the swimming velocity of the amphipods after the experiment was modulated by the exposure to fluoxetine, and also by the exposure to prochloraz. We demonstrate that both the antidepressant and the fungicide may cause significant sublethal effects at low concentrations. The combination of behavioral endpoints together with the application of mixed models provided a useful tool for early detection of the effects of toxicity mixtures in freshwater ecosystems.