Development of objective methods to quantify reward learning impairments, a 'cognitive biomarker' of affective state in different species

Achieving emotional wellbeing in animals is a core objective of animal welfare. We know from studies in humans that emotional wellbeing can suffer without overt signs of distress or obvious changes in behaviour, and it is only through specially designed questionnaires that poor mental health is identified. The methods used for humans are largely based on subjective self-report, something which cannot be used in non-human animals.

However, to deliver positive animal welfare requires the ability to quantify the animals affective or 'emotional' experience. Although overt changes in the animal's behaviour can provide some information, these do not offer sufficient sensitivity to understand the emotional impacts. Indirect measures such as ill health, survival rates and productivity provide only a crude measure and are not sensitive to positive welfare change.

Our research has pioneered the development of objective methods to measure animals' affective state. By translating behavioural methods from human psychology, we have developed and validated novel assays which can accurately measure animals positive or negative affective experience of different interventions. We have recently discovered that reward learning is sensitive to different affective states in rats and mice and the specific nature of the effects we observe are replicated in patients with major depressive disorder.

We now have a substantial amount of evidence to support the idea that changes in reward learning are directly related to the core affective state of the subject. As affective state is strongly linked to emotional wellbeing in animals, being able to quantify this objectively has the potential to transform how we deliver positive welfare interventions and enable an unbiased, animal perspective which is not influenced by human perception and interpretation.

In this project we have planned two different phases of research that each address an objective related to our overarching goal of developing and validating a method to automatically quantify the emotional wellbeing of a non-human animal. The first phase of the project will investigate different types of reward learning in three key species, mouse, chicken and dog tested in their home environments.

These will be the first studies to explore different species in parallel studies using equivalent behavioural tests accommodated in the real-world living environment. We will learn how animals interact with our reward delivery devices and how this relates to reward learning in tests of different levels of difficulty and reinforcement schedules. These studies will provide new insights into the way animals learn to perform specific behaviours to obtain reward when this is available via a device within their normal home environment.

Previous studies have removed animals to test in specialist equipment and it is important to understand the impacts of continuous access in the home environment to identify the best task to use for subsequent experiments. We will then use this knowledge to design experiments to then validate our approach using representative populations for each species and associated with relatively more positive or negative affective states.

By the end of the planned research programme we will understand reward learning and its relationship with affective state in three representative species. Our methodological advances and associated species-relevant reward learning devices will be made available as open source resources to enable other researchers to use them within their own programmes.

We are also optimistic that the simplicity of the methods and their automation will, in the future, enable more general use beyond animal welfare science.