What animals can tell us about how they perceive the world

Recognition

One of the most fundamental questions that needs to be addressed for understanding the perceptual abilities of any species is which senses they use for finding their way around their environment, to recognise palatable from unpalatable foods and recognise both friends and enemies. Scientists at the Babraham Institute for example have shown that, contrary to expectation, sheep do not rely on their sense of smell to discriminate between different palatable foods and other sheep and even humans. In fact, like us, they mainly use vision. Indeed, they have a visual acuity that is considerably better than companion animals such as dogs and cats. The one important exception to this is mother ewes recognising their lambs in the period immediately after giving birth where smell is the most important sense. However, even in this case, as the lambs get older their mothers can also recognise them by sight and by the sound of their voices.

One of the key areas where understanding the recognition abilities of animals is of importance for an appreciation of their welfare needs is the extent to which they are able, and need to, distinguish between other individuals. If it can be shown that any animal can recognise many different individuals in its social environment this immediately underlines the importance of its social environment for its welfare, otherwise why would it have developed such abilities.

Scientists at the Roslin Institute have demonstrated for example that domestic chicks and quail can discriminate between cage mates and strangers. Scientists at the Babraham Institute have shown that sheep can actually recognise at least 50 different sheep and 10 humans visually from their faces. Indeed, their face-recognition skills and the way their brains are organised to perform this difficult task show remarkable similarities to humans and monkeys. They can use cues from the internal features of the face (eyes, nose and mouth) to recognise highly familiar individuals and, like humans, make more use of the features from the side of the face appearing in the left visual field (i.e. the right half of a perceived face) for recognition. This is because both sheep and humans mainly use the right side of the brain for recognising faces and this is preferentially accessed by visual cues from the left visual field.

Research by the Babraham scientists has also shown that for sheep faces are not just used for recognition but are also a source of attraction and comfort. Thus, females will not only show a differential preference for male and female faces depending on their state of sexual receptivity but when they are sexually interested in males they will also show preferences for the faces of specific individual males. Also, both behavioural and physiological responses to the stress of being isolated are markedly reduced in animals just exposed to the sight of face-pictures of familiar sheep. The importance that the sheep have not just for recognising individuals from their faces but also for assessing their relative attraction can be seen by the way their brains actually encode faces. Cells in the temporal cortex that selectively encode face stimuli are actually sub-divided into groups responding to particular categories of face that have a similar emotional or behavioural significance:

• faces which have horns and how big the horns are (an important indicator of dominance and gender)
• faces of sheep of the same breed and particularly familiar individuals (sheep prefer to socialise with members of their own breed and need to be able to recognise familiar animals)
• faces of humans and dogs (the main species that represent a potential threat to the sheep).

Recognition of individuals from their faces clearly requires close proximity and for detecting potentially threatening individuals the animals need to use other cues that can be utilised at a distance. Thus, humans can be recognised by body shape and posture and another group of cells in this same region respond selectively to the human body shape. They respond in the same way even if the head and shoulders of the shape are covered, and regardless of whether it is viewed from behind or from the front. However, most of the brain cells respond only when the human shape moves towards the sheep and not if it is stationary or moving away. If the human shape moves towards the sheep by crawling "on all fours" rather than walking, there is less response. This suggests that the sheep's brain is processing information in a way that recognises not just a human shape but also the significance of its movements.

However, humans are clearly not always perceived in a negative way by sheep, and sheep will actually seek proximity with a highly familiar human carer. In this case it has been shown that the face of such a familiar human can become encoded by cells in the brain that normally only detect the faces of familiar sheep in their flock.

Like humans, chicks use the two sides (hemispheres) of their brains for different tasks: the left hand side being more involved in decision-making. Research at the University of Sussex has shown that chicks that are carrying out a plan of action such as approaching to lift a lid use their right eye to fixate the object that they will grasp. This strategy of using one or other eye to bring into action the appropriate side of the brain is probably general, since chicks also respond differently to objects according to whether they lie on their right or left sides. Partners are more likely to be attacked for example if they appear on the left. There is also evidence that the system fed by the left eye is more important in determining whether a chick is frightened by a sight.

Other research at the University is investigating how chicks perceive colour and form in a task that resembles foraging. This could provide new insights into the importance of colour for chickens, and could suggest how colour might be used to enrich the environment of housed birds to improve their mental well-being. The design of housing should take into account hens' abilities to navigate and make choices. Furthermore, research at Roslin found that televised images sustained interest and reduced fear in chickens. Welfare of birds may be enhanced by making it easier for them to look at reassuring and interesting images.

Out of sight

An animal's mental well-being may be determined, at least in part, by whether it can hold "mental pictures" independent of its surroundings i.e. if it understands that objects continue to exist even when out of its sight. The following examples raise the question of whether such cognitive ability increases an animal's potential to suffer stress.

• If nesting materials are withdrawn from a hen, will she still perceive their existence and search for them? Or not?
• If one sheep is removed from a flock, will the others remember it, perceive its continued existence elsewhere and search for it? Or not?

The Babraham research has shown that these visual cells in the temporal cortex of the sheep's brain that respond to the actual sight of an individual's face can also respond when that individual is expected to appear or is temporarily hidden from view or can only be heard through a recording of their vocalisations. In humans, this same specialised part of the brain is activated in a similar way when we see or imagine faces and perhaps therefore sheep also have some ability to imagine the faces of absent flock members or even humans. This possibility receives further support from recent studies at Babraham showing that sheep's brain cells may still be "tuned" to the images of faces of individual sheep and humans even after one or two years of absence. Formal behavioural experiments also confirmed that sheep retained the ability to discriminate between 50 different faces even when they had not seen them for over two years.

There are other implications of an animal being able to hold "mental pictures", including its ability to navigate around its surroundings. If a mental picture exists, say of the site of a feeding container, then the animal will be able to navigate to it from any location nearby; otherwise it will be limited to finding the container by trial and error or by following a set route. This is important in determining how easily individual animals can find their way around their environment, and has welfare implications for the design of new housing or accommodation for animals.

Behavioural studies at the Roslin Institute, funded by the former Ministry of Agriculture, Fisheries and Food, showed that hens can form mental pictures of the location of objects which researchers had hidden behind screens, and can use these pictures to search for the objects. However, the results indicated that although the animals have this ability they do not always use it. Motivation is important. The Roslin study showed that hens can store a mental picture of the quality of the food hidden in this way and that this picture can influence how motivated they are to search for the food.

Researchers at Oxford ran a similar experiment on mink, as they were interested in seeing whether the sight, smell and sound of environmental enrichments affects the motivation animals have to interact with them. This is important because if such stimuli have powerful 'persuasive' effects, then investigating how hard animals will work for enrichments that are right under their noses (see page 23 on preference tests) would not give a very good picture of how much an animal in a barren cage would like them. The researchers measured how hard mink would work for toys, food, a water-bath and the chance to see another mink, both when they could see these resources from their home cage, and when they could not. The animals were more motivated to reach the toys when they could see, hear and smell them, but for all the other resources, in contrast, out of sight was not out of mind: the animals worked just as hard for them even when their stimuli were screened off. So just as with hens, the mink could learn and remember where resources are, even when they can't see them, and they also showed high motivations to reach them. But for some resources (the toys in this instance), stimuli have a powerful effect on motivation. This means that we have to be careful when we design and run preference tests: preference tests with screened resources could give us quite different results from tests with resources the animals can see and smell first.

In a new research project at the University of Cambridge, scientists are exploring the ability of Western Scrub Jays, (a US native species of bird similar to the British Jay) to use memories of past experiences to plan for the future. These researchers have previously shown that birds with experience of stealing other birds' hidden caches of food seem to use this knowledge when hiding their own supplies.

Social behaviour and understanding

If animals can both recognise each other, and retain mental images of each other, then the questions arise "Can they think about each other" "Can they learn from one another?" and even "Can they understand what other animals are experiencing or know?" Answers to these questions will be important in understanding the complexity of animals' minds, their potential for suffering and the extent to which it might be possible to provide conditions that promote desirable behaviours and inhibit those, such as stereotypic behaviours (see page 30) that are undesirable.

Scientists at the University of Bristol have studied whether hens are able to learn from each other, and whether such learning is a relatively simple innate response, or whether it indicates a cognitive understanding of what other birds are doing and the consequences of their behaviour.

In these experiments, some hens are trained to carry out specific behaviours, for example, to peck at a green key, rather than a red one, in order to get a food reward. These are known as "tutor" hens. Scientists then watch to see whether "observed" hens, that have no previous experience of the green key and its relationship to food, can learn to peck at the green key by watching the tutor bird. The research showed that hens that watched a tutor did indeed learn much more quickly to peck at the green key than did hens that had not seen the tutor.

It clearly makes sense for animals to be able to learn from one another, not least because in the wild this could save them valuable time and effort. This raises a question, however, about which animals others would naturally choose to observe and copy. The researchers at Bristol found that, in their experiments, observer hens learnt faster from tutors that were dominant hens in the flock (perhaps because the dominant birds pecked with more confidence) and that they also learnt faster from watching hens rather than cockerels.

These findings do not necessarily mean that observer hens understand what the tutor bird is doing, it could be that they simply copy its behaviour, regardless of outcome. Indeed, subsequent experiments showed that pecking is copied, even if the observers do not see the tutor getting its food reward. This suggests that hens are not taking consequences into account.

The Bristol team investigated whether hens can recognise and respond to chicks that are behaving in a way which may seem inappropriate to the mother hen. Hens perform a variety of innate behaviours to show chicks what food to eat, for example, they repeatedly pick up and drop the food in front of the chicks. Researchers trained hens to select food that was coloured red and reject food that was coloured green. They then trained half their chicks in the same way, and the other half with the reverse colour preference i.e. to select green and reject red. When the hens watched both sets of chicks foraging they performed their innate behaviours faster and in a more exaggerated way with the chicks that were selecting the green food. This indicates that the hens could recognise that the chicks were behaving "wrongly". It does not necessarily mean that the hens found this stressful.

Other researchers at the University of Bristol are investigating potentially much more sophisticated social understanding in pigs. Their objectives are both to increase basic understanding of the mental abilities of pigs and to use information about these abilities to design husbandry methods that enhance pig welfare on the farm.

One of the biggest practical welfare issues affecting pigs that are kept in groups on farms is their propensity to fight and injure each other. Pigs differ greatly in temperament: some are naturally more aggressive than others. Individuals can even be "scored" for their tendency to fight: so-called "rapid responders" will fight much more quickly than "slow responders".

In studies at Edinburgh and Bristol, fewer fights occurred in a newly mixed group of unfamiliar pigs that contained both fast and slow responders, than in groups containing either all fast responders or all slow responders. This suggests that when pigs of different aggressiveness are mixed they make some form of assessment of each other's fighting potential and use it to sort out their social status, in preference to fighting. In other words, slow responders choose not to pick fights that they might lose.

The Bristol scientists have developed an experimental system to investigate whether pigs are able to assess the aggressive motivations and fighting ability of others without having to come into direct contact with them. Pigs of different temperament are given a choice to enter or avoid a pen that is adjacent to one containing a rapid or slow responder. Initial results suggest that pigs behave differently to, and hence may discriminate between, rapid and slow responders.

Studies on scent signalling by mice and other rodents, by researchers at the University of Liverpool, are beginning to reveal how a combination of marking behaviour and chemical information enables the animals to communicate about, for example, their sex, individual identity, kinship, and their social and reproductive status. Scent marks of male mice provide a continuous record of competitive challenges between individuals. By covering their defended territory with fresh scent marks and preventing other males from depositing competing marks, territory owners indicate that they are defending their territory successfully. If any other males deposit competing scents, the territory owner will counter-mark immediately, ensuring that his marks are the freshest. The researchers at Liverpool have found that mice assess the competitive ability of a territory owner from the presence of fresh counter-marks from competing males. Female mice prefer males that are able to counter-mark any intruder scents; while males are more likely to challenge owners of territories that contain fresh scents of other males. Crucially, this depends on the acute ability of rodents and other animals to recognise different individuals from their genetically-determined scents.

Odour cues, from saliva or urine, might be one mechanism by which pigs sense social status and behavioural characteristics in others. Researchers at Silsoe and Roslin found that pigs could differentiate between familiar and unfamiliar animals just on the basis of airborne olfactory cues. Recent research at Bristol has shown that young pigs are able to discriminate between urine from different unfamiliar individuals. The next questions to be addressed are whether rapid and slow responders can be discriminated according to differences in their saliva or urine. A better understanding of these processes may help farmers to provide conditions that minimise fighting, for example by facilitating the use of individual recognition and assessment behaviour. By combining behavioural information with genetic data, it might also be possible to breed pigs that are naturally less aggressive.

Bristol scientists and researchers at the University of St Andrews are investigating whether pigs can learn how to profit from the knowledge of others, whether individual pigs can change their behaviour to "deceive" others, and even whether they can develop "counterdeceptive" tactics to out-compete others. These are complex social behaviours that have previously been studied mainly in primates.

To do this, the researchers are using specially designed foraging experiments in which some pigs are trained so that they can repeatedly find food in the test area. These act as "informed" pigs: they know where the food is hidden (they are rather like the tutor hens in the experiments described above). An informed pig is then mixed with a "noninformed" pig that does not know where the food is, and researchers observe and record how the animals interact. Results indicate that non-informed pigs can use their informed partners to lead them to food, and that informed pigs can develop behaviours which act to decrease their chances of being exploited, such as visiting the food source more often when the non-informed pig is out of sight.

Related experiments are investigating whether pigs are able to discriminate between individuals that do or do not have reliable information about the location of food.

This research requires rigorous design and interpretation in order for true social understanding of the knowledge of others to be distinguished from situations in which animals respond to the behaviour, rather than the knowledge or mental states, of others. Some of the behavioural studies used with livestock species have been derived from those first developed for primates. Information about how animals understand the knowledge of others will have practical applications - for instance in identifying how animals might be stressed by observing particular treatment of others, and therefore identifying what practices should be kept hidden from other individuals. Recently, a large-scale study has been started by scientists at the University of St Andrews to examine how tool use and foraging skills are transmitted by observational learning between primates. This study will be based on observations of young wild-born chimpanzees living in African sanctuary enclosures.

Remembering and forgetting

If farm animals' memory for information relating to husbandry procedures, how to locate and use resources, and the identity of conspecifics, is disrupted in any way, this may lead to stress, poor welfare and the need for relearning or re-training. Several areas of research are pursuing ways of minimising interventions which could have such disruptive effects.

Research at Bristol has shown that everyday husbandry events such as moving to a new pen, or meeting an unfamiliar animal can disrupt short-term spatial memory performance in pigs. Whether this is also true for social memory now needs to be investigated. For example, taking away an individual sow from an established and settled group to give birth on her own, and returning her some weeks later, results in an initial period of fighting when the sow is reintroduced. It seems that she and the others have "forgotten" their relationships and need to re-establish the group's social interactions. Understanding how long such social memory lasts, how it is disrupted, and whether it can be prolonged, are important research questions.

Conditions which enhance animal memory may thus enhance welfare and minimise stress. A better understanding of how animals establish and retain memories could aid the design of husbandry practices that minimise stress and the time required for management procedures.