Monitoring animal behaviour in infectious disease studies with accelerometers and video

Automated monitoring of behaviour is a promising method to fa-cilitate reduction and refinement in animal experiments. We ex-plored the use of accelerometers and video to monitor behaviourin infectious disease trials with group-housed animals under bio-safety level 2 and 3 conditions. In a recent study, three rams wereequipped with two 25 Hz accelerometers, one on the ear and oneon the back, and an Aruco-marker for video analysis. Data werecollected starting 5 days prior to infection with Toxoplasma gon-dii through to 21 days post infection. An activity index (AI) wascalculated from accelerometer data and time spent drinking wasquantified with an automated approach using pixel changes fromvideo. The ear-based AI was higher than the back-based AI, butboth showed similar patterns over time. Three days post infection,the AI halved during the daytime/active period, while it slightly in-creased during the night. These data supported a partial disappear-ance of circadian rhythm and a general restlessness. Time spentdrinking also decreased for each ram. These trends correspond-ed well to increased body temperatures measured from days 3-9.At 9-11 days post infection, the AI and drinking time returned topre-infection patterns. As the animal caretakers scored the ramsas “inactive” only during days 5-7 post infection, our monitoringprovided additional insights into behavioural patterns and diseasesymptoms. We believe that we can work towards a real-time con-tinuous and automated monitoring system that can facilitate refine-ment and reduction in experiments with a variety of group-housedspecies in high-containment conditions.