Individual differences in affective processing and implications for animal welfare: a reaction norm approach

Individuals differ in how they view the world. Some are more optimistic than others about the future or the outcome of ambiguous situations. Some are more sensitive to stimuli that are pleasant or rewarding. And some are more strongly affected by unpleasant things. These differences impact how individuals respond to opportunities or challenges in their lives with knock-on effects for their general wellbeing and vulnerability to affective (emotional) disorders such as depression. This is not just the case for people. In animals there is growing evidence that individuals vary in 'optimism' / 'pessimism' when deciding whether an ambiguous signal heralds a positive or negative outcome, and some evidence that these individual, or personality, characteristics predict how well they cope with challenge.

The aim of this project is to comprehensively investigate these individual differences and their implications for animal welfare. We will use a cognitive task developed in our lab and designed to measure 'optimistic' or 'pessimistic' responding to ambiguous cues as a marker of, respectively, positive or negative affective states. This task has been used in more than 160 studies across a wide range of mammalian and bird species, and there is evidence of personality differences in 'optimism' that can influence an animal's vulnerability to stress. Furthermore, computational modelling of data from this task also allows us to derive measures of an individual's sensitivity to pleasant and unpleasant events.

In this project, we will first quantify individual differences in the average expression of these characteristics (personality differences) in a widely used lab animal, the rat. We will also quantify how individuals differ in their stability of expression of the characteristics across repeated tests (variability or predictability), and in how flexible their responses are when tested across different contexts (plasticity). This will provide us with a more detailed picture than before of individual differences in 'optimism' and sensitivity to pleasant and unpleasant events.

We will then evaluate how different types of individual respond to changes in the presence of opportunities or challenges in their home cage environments, for example by enriching their cages or introducing some unpredictable events. We will measure markers of their welfare and use the same analysis approach as above to establish each individual's average welfare across contexts and, in particular, the extent to which its welfare improves when in an opportunity-filled environment or becomes poorer in more challenging environments. We will thus be able to test, for example: whether more 'optimistic' individuals, especially those showing predictable rather than variable levels of 'optimism', cope better with challenge; whether individuals who are more sensitive to pleasant events benefit more from an increase in opportunities; whether individuals who show greater sensitivity to unpleasant events are more strongly affected by challenge. We will thus map links between individual characteristics and robustness or vulnerability in welfare-enhancing or welfare-challenging environments.

We will then ask whether we can alter individuals' characteristics and improve robustness by exposing them to experience of play in complex environments with other animals. There is growing interest in the use of play as a practical way of improving welfare in lab animals, and we will systematically evaluate this possibility which, if supported, could inform animal management practices. Moreover, if play experience results in greater stability of individual responses in standard behavioural tests (e.g. of anxiety-like states, cognitive ability) despite changes to home environments, this will not only indicate that it acts to maintain welfare in the face of challenge, but also that it can enhance cross-context reproducibility of testing, an important goal in modern science.

Differences in how people process affective information (optimism-biases; reward/punishment sensitivity) influence their affective states, wellbeing and vulnerability to affective disorders. In animals, evidence from the judgement bias task (JBT) which measures 'optimistic' or 'pessimistic' decision-making under ambiguity as a marker of positive or negative affect, also indicates individual differences in 'optimism' that can influence vulnerability to stress. JBT data also yield reward/punishment sensitivity parameter values via computational modelling.

Here we will study individual differences in these characteristics and their implications for animal welfare in rats. We will analyse data from repeated JBTs using double-hierarchical (DH)GLMs to compute individual differences in mean response (personality; intercepts in mean model); change across time (plasticity; slope in mean model); variability across repeat-testing (predictability; intercept in dispersion model). Individuals will then experience changes in home cage opportunities (enrichment) and challenges (unpredictable events), and measures of welfare will be collected across contexts. Each individual's mean welfare (intercept) and how it changes across home cage contexts (reaction norm; slope) will be quantified using DHGLMs. GLMs will compute links between individual affective processing and welfare reaction norms to test hypotheses, e.g. 'optimistic' animals cope better with challenge; reward- and punishment-sensitive animals respond more strongly to opportunities and challenges respectively.

Affective processing and welfare reaction norms will be quantified in rats with or without playpen experience to evaluate the effectiveness of this practicable method for improving welfare. Play-induced flattening of reaction norms will indicate maintenance of welfare in the face of challenge, and better cross-context reproducibility of behavioural tests; findings of both practical and scientific importance.