Measuring cognition and emotion of animals to understand their welfare

The major knowledge gap in assessing animal welfare is our inability to measure how animals actually feel in the different situations in which we farm them. As animals are unable to speak, we must find ways of gaining insight into their minds. Our animal welfare research at CSIRO includes a focus on understanding the cognitive abilities of animals in order to gain insight into how they feel and to thereby enable us to improve their welfare. Animals perform better in tests of learning that require skills related to their fitness and survival. For example, grazing animals, such as cattle and sheep perform well in maze tests, indicating they have good spatial memory, which allows them to graze with optimum efficiency. Species that wander across a territory searching for random food sources, such as chickens, pigs and dogs do not perform as well in tests of spatial learning. By understanding the cognitive abilities of different species we can use cognitive principles, based on human psychological theories, to develop scientific approaches for measuring emotive state in animals. An example of this is the assessment of cognitive bias. From the response of an animal to an ambiguous situation we can infer its emotional state. Thus, when an animal is content, it is more likely to assess the ambiguous scenario as potentially delivering a positive outcome. In contrast, when an animal is in an adverse mental state, it is more likely to assess the same scenario as delivering a negative outcome, and respond accordingly. The paper will describe the application of a range of cognitive tests to sheep. If we understand how animals think and feel, we may be able to do a better job at optimising their welfare. The Challenge of Measuring Animal Welfare Society is increasingly concerned with the welfare of animals used for food production and this concern transfers into the product choices consumers make. The science of objective measurement of animal welfare is relatively new. There exist some methods of assessing animal welfare, such as measuring blood variables indicative of changes in physiology or immunology. Studies of animal behaviour have also been used to indicate obvious states such as pain or discomfort, or preferences for different environments, however, the information gained may be relatively limited. The major knowledge gap in assessing animal welfare is our inability to measure how animals actually feel in the different situations in which we farm them. As animals are unable to speak, we must find ways of gaining insight into their minds. Our animal welfare research at CSIRO includes a focus on understanding the cognitive abilities of animals in order to gain insight into how they feel and to thereby enable us to improve their welfare. The challenge is to achieve this insight into animal mental states in a scientifically rigorous way. In order to measure animal welfare, we need to get animals to demonstrate to us how they feel, and to do this, we need to better measure and understand their cognitive abilities. Understanding Animal Cognitive Abilities Cognition is the mental process of knowing, thinking, learning and judging. Learning involves a modification in animal behaviour that occurs as a result of experience. Animals have evolved their cognitive abilities to increase their chance of survival. To highlight some of the concepts we wish to convey in this paper, we will use the sheep as an example. Sheep traditionally and unfairly have had a reputation as being “dumb”. This perception of sheep may have developed from the difficulties of handling individual sheep when conducting routine management practices due to their strong flocking instinct. Sheep have evolved this instinct to increase their chance of survival and help protect them from predators, basically, through safety in numbers. The assessment of cognitive processes, such as spatial learning and memory has been commonly performed in rats using maze tests, such as the Morris water maze and the radial-arm maze (Hodges, 1996). To further our understanding of the cognitive and learning abilities of sheep, we developed a maze test for sheep. The maze test utilised conspecifics and the strong flocking instinct of sheep as motivation to move through the maze, thereby negating the need for prior training. An 18.4 x 8.2 m outdoor maze was constructed with 1.2 m high opaque external walls and a number of internal barriers (Figure 1). The inner walls were made of open-barred portable fence panels to enable animals to view conspecifics at the opposite end of the maze, thus providing motivation to traverse the maze and join their flock mates. The first time an animal navigates the maze indicates cognitive ability, whereas the improvement in completion time over successive testing indicates learning. We validated that the maze was measuring spatial memory in sheep by administering scopolamine, a muscarinic receptor antagonist that has been shown to impair memory function (Ferreira et al., 2003). Sheep receiving scopolamine were unable to improve their time to navigate the maze over 3 consecutive days of testing, whereas control animals significantly improved over the 3 days (Lee et al., 2006). Figure 1. Maze layout We studied the learning profile of Merino sheep (Figure 2) and found that sheep significantly reduced the total times taken to complete the maze over 3 consecutive days. When sheep were tested 6 weeks later (Days 42, 43 and 44), they improved further and were faster than on Day 3. With experience, the sheep committed fewer errors (the time spent in blind compartments of the maze), improving from Day 1 to Day 3. There was no further improvement in error times when sheep were tested 6 weeks later, with no differences shown between these days or Day 3. In a separate study, when sheep were tested one year later, they were significantly faster to complete the maze, indicating that they did not only retain the memory of the maze but had improved in their performance. 0 20 40 60 80 100 120 To ta l t im e (s ec )