engleski

Behavioural elements as the early biomarkers of exposure

Toxicants are present in a wide range of concentrations in the environment, inducing lethal and sublethal effects in organisms. According to research to date, behaviour has a 10–1, 000 times higher sensitivity than the conventional LC50, and therefore its continuous monitoring is gaining more recognition. Analysis of behavioural parameters allows detecting behavioural elements characteristic for the exposed test organism, which can be used in preliminary screening and detection of sublethal concentrations of pollutants in the environment. The aim of the presented studies was to quantify the activity of aquatic (Daphnia magna) and terrestrial (Dendrobaena veneta) organisms exposed to sublethal concentrations of toxicants and to detect behavioural elements characteristic to the exposed organisms. Daphnia and earthworms were exposed to sublethal concentrations of ZnCl2 and H3BO3, respectively. Swimming activity of five Daphnia was monitored in transparent Petri dishes containing control and polluted media, with a smartphone, instantly upon exposure of the organisms to the toxicant (t0), as well as 1 h, 24 h and 48 h of exposure. The analysis of swimming activity was carried out in Python. Earthworm behavior was continuously monitored in a modified avoidance test setup, in glass terraria filled with control soil in one side and polluted soil in the other side. Automatic detection of burrowing earthworms was accomplished using a trained artificial neural network (ANN) model. Predictions of the ANN model were used to quantify the relative presence of earthworms in each side of the terrarium. Some of the most important behavioural elements of Daphnia exposed to pollutants observed are discontinuous swimming, spinning and movement along the wall of the dish. Exposed earthworms exhibited “exploring behaviour”, penetrating in the polluted half of the terrarium as they descended towards bottom layers of the soil.

continuous monitoring, swimming, burrowing, computer vision, U-net

nije evidentirano