Using consummatory behaviour to measure the affective state and welfare of laboratory mice

Abstract

Mice are the most widely used model in biomedical research, making it important to know how life in the laboratory impacts on their welfare. Whilst it is difficult to evaluate how a mouse might ‘feel’ because of their inability to self-report, behavioural and physiological measures can give insights into their current welfare state. One approach is to measure their ‘affective state’ through how they respond to reward. Humans with depression often report a lack of enjoyment from reward (known as anhedonia), which can also be measured in laboratory rodents by measuring sucrose consumption: the less they consume, the more anhedonic they are. However, consumption is confounded by other factors, particularly motivational state. Therefore, an alternative method assesses the microstructure of the animals’ licking patterns, which may better reflect an animal’s hedonic response towards reward, i.e. how much it ‘likes’ it. The aim of my thesis was to determine how stress influences the hedonic responses of laboratory mice, and determine whether assessing changes in consumption or licking microstructure could be used to infer a mouse’s affective state, in order to make evidence-based improvements to their welfare. Experiments using standard depressogenic methods (i.e. chronic mild stress and chronic corticosterone administration) were ineffective at altering affective state, and sucrose consumption and licking microstructure were unchanged. However, I found that current methods used to handle laboratory mice were sufficient in inducing changes in the animal’s affective state. I found that the standard practice of handling mice by their tails causes alterations in reward perception, revealing a depressivelike state. These experiments provide more support for refinements to be made with regards to the existing handling practices of laboratory mice. I discuss my findings in relation to implications for animal welfare and scientific data collection across a number of in-vivo models