In the field of odour detection, dogs are generally considered superior to both human and technological alternatives. However, the complexities of odour detection in a real-world environment poses a significant challenge to the development of reliable training methods, as in reality the target odour is not found in isolation. A vast number of irrelevant odours, along with significant variations in the properties of the substance, can affect the dog’s ability to detect the target odour. The ability of the dog to generalise the target odour from a training scenario to account for these variations is key to their success, and research into how this is affected by the training method is currently growing. A compound training method is widely reported to improve generalisation compared to the more traditional sequential approach, however recent work with rats provided interesting results for a third method described as “intermixed”. This method requires the dog to learn several target odours concurrently. The current study uses a basic odour detection line-up to assess the three groups of pet dogs who were trained to detect two target odours, A and B. The intermixed group trained on A and B concurrently, the compound group on a mixture of AB, and the sequential group on either A or B, prior to learning the opposite odour. Each dog was tested on all combinations of the test stimuli (A, B and AB), as well as combinations containing a novel interferent (AC, BC, ABC). Results demonstrated the intermixed training group correctly identified 85% of test stimuli, performing significantly better than the compound group which identified 62.8% (p = 0.025), and sequential group which identified 60% (p = 0.013), providing evidence in favour of the use of an intermixed training method to promote generalisation in detection dog training.
PLEASE NOTE: You must be a member of the University of Lincoln to be able to view this dissertation. Please log in here.