Remotely monitoring heart-rate and feeding behaviour of fish by using electronic sensor-tags

Sammanfattning: Modern electronic sensor tags (BioLoggers) can collect high resolution data on movement and physiology of fish. These tools make it possible to remotely monitor physiological stress responses as well as complex behaviours, such as feeding and hunting, in free swimming wild fish, data that often is difficult to collect by other means. Quantification of stress and feeding behaviour is valuable in behavioural ecology and fish conservation, and are also important in fish welfare (e.g. in hatchery environments and in fisheries). This study had two main goals: 1) Evaluating the use and function of a newly developed cordless heart rate bio-logger to measure stress in hatchery brown trout (Salmo trutta), and 2) Evaluating the use of a high resolution 3D accelerometer bio-logger to remotely capture different foraging behaviours in Eurasian perch (Perca fluviatilis). Bio-loggers are often inserted surgically in the fish, and in study 1 the first aim was to evaluate if the placement of the bio-logger inside the fish affects the quality of the data produced. Results showed that placement close to the heart produced significantly higher quality of data compared to when the logger was placed in the belly of the fish. The direction of the logger with regards to the electrodes did not impact the quality of the data. Based on this result, a protocol for the surgery and placement of cord-less HRT-sensors in trout was developed. To evaluate the ability of the bio-logger to capture change in heart rate as a function of stress, hatchery brown trout was tagged with heart rate sensors and exposed with a GABAergic anxiolytic drug known to reduce stress in fish. The fish was then exposed to handling stress, and heart rate in exposed fish was compared to control fish that was not exposed to the anxiolytic drug. The result showed that the control fish had significantly higher heart rate during handling periods compared to non-handling periods, but that such difference could not be seen in the drug exposed fish, hence indicating that the new cord-less heart rate logger is capable of quantifying induced stress and stress relief in fish. In study 2, two different feeding behaviours in perch, feeding on small insects and feeding on small fish, were carefully studied in aquaria. The two behaviours were then manually simulated using a euthanized perch tagged with 3D accelerometer bio-logger. The data collected by the accelerometer was then analyzed using a machine learning approach, with the aim to be able to distinguish between the two feeding behaviours using only the acceleration profiles. The result showed that it is possible to distinguish between feeding behaviours in perch with high accuracy (99 %) using acceleration bio-loggers. The results are discussed based on the potential usability of heart rate and accelerometer bio-loggers in animal welfare and ecological research.