How do dogs respond to olfactory changes associated with human health and stress?

Abstract

Dogs' ability to detect the odour of human disease has led to an increasing number of medical detection roles, and as companions, dogs are commonly believed to sense our emotions. However, we know little about the odours dogs detect or how they affect their emotions. Using a combination of volatile organic compound (VOC) analysis, behavioural analysis and cognitive bias testing, this
thesis explores how dogs detect and respond to human health and stress odours.

Medical Alert Assistance Dogs are trained for a range of chronic conditions. We investigated differences in VOCs from individuals with type 1 diabetes and Postural Tachycardia Syndrome (PoTS) and found significant differences between samples collected during medical events compared to control samples. Objective behavioural analysis of a PoTS-alert dog confirmed they reliably warn their owner before the onset of symptoms, and interviews with five PoTS-alert dog owners revealed these dogs improved their health, safety and quality of life. To assess dogs' responses to remote disease samples, we compared the behaviours of eleven bio-detection dogs during detection tasks to handlers' behaviour ratings. We found that most behaviours exhibited when dogs encountered a target were correctly identified as alerts by handlers. Inter-rater reliability tests showed moderate to substantial agreement on decisions but identified opportunities to improve reliability when recording dog performance. Finally, we assessed the response of eighteen companion dogs to the odour of an unfamiliar person when stressed and relaxed and found that stress odours led to more pessimistic choices in a cognitive bias test.

Together, these findings demonstrate dogs’ ability to detect and respond to target odours of health using clear behavioural responses and show that detecting human stress odours may negatively affect their emotions and behaviour. Future research should focus on better understanding the implications of these findings on dogs' working ability, performance and welfare.

Date of Award	10 Dec 2024
Original language	English
Awarding Institution	University of Bristol
Sponsors	Biotechnology and Biological Sciences Research Council & Medical Detection Dogs
Supervisor	Nicola J Rooney (Supervisor), Carsten Muller (Supervisor), Claire Guest (Supervisor) & Andrew Dowsey (Supervisor)