Validation of tri-axial accelerometers and the impact of environmental enrichment on behaviour and welfare of domestic dogs (Canis familiaris) : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Animal Science at Massey University, Palmerston North, New Zealand
Loading...
Files
Date
2024
DOI
Open Access Location
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Massey University
Rights
The author
Abstract
Monitoring and quantifying behaviour in dogs can provide valuable insight into their overall health and welfare. Behavioural observation in dogs can often be labour-intensive and time consuming. However, recent advances in remote sensing technologies, such as accelerometery, offer an automated method for continuously measuring behaviour without an observer present. The first aim of this thesis was to investigate the use of remote sensing technology, particularly tri-axial ActiGraph® WGT3X-BT accelerometers, along with machine learning (ML) algorithms to automatically classify behaviour in six colony-housed domestic dogs (Chapter 2). A total of 132,295 seconds (~36.7 hours; ~6.1 hours per dog) of video footage of behaviour were recorded. Five modelling rounds were created using ML techniques, with model 4 achieving the highest overall accuracy and kappa coefficient while still capturing a wide range of behaviours. The study found that the ActiGraph® WGT3X-BT accelerometer can accurately classify behaviour in domestic dogs. However, it also revealed challenges in differentiating behaviours with similar acceleration profiles, particularly in classifying the "standing" behaviour. As a result, behaviours were grouped during the model-building process to improve overall accuracy. The refined models significantly improved over time, indicating a promising method for detailed and remote assessment of domestic dog behaviour. The secondary aim of the thesis was to evaluate the potential use of tri-axial accelerometery and a validated random forests model for determining the efficacy of environmental enrichment treatments and to assess the effect of food, olfactory, and tactile enrichment treatments on the behaviour and activity of six colony-housed domestic dogs (Chapter 3). Significant differences were observed among enrichment treatments regarding active/inactive behaviour, ODBA levels, individual behaviours, and interaction durations. The use of ActiGraph® devices were demonstrated to be an accurate and objective method for measuring the success of enrichment activities, particularly when used alongside other observational methods such as interaction duration. The study emphasised the impact of environmental factors, individual differences among dogs, and seasonal variations on the effectiveness of enrichment. It highlighted the need for personalised enrichment programs to enhance the overall effectiveness of treatments.