Morten Bakken

Assessment of positive emotions in animals to improve their welfare

It is now widely accepted that good welfare is not simply the absence of negative experiences, but rather is primarily the presence of positive experiences such as pleasure. However scientific investigation of positive emotions has long been neglected. This paper addresses two main issues: first, it reviews the current state of scientific knowledge that supports the existence of positive affective states in animals and, second, it suggests possible applications of this knowledge that may enhance quality of life under animal management conditions. In the first part of the paper, recent advances in psychology and neuroscience are reviewed to provide pragmatic frameworks based on cognitive processes (such as positive anticipation, contrast and controllability) for further investigations of positive emotions in animals. Thereafter, the neurobiological bases of positive emotions are highlighted in order to identify behavioral and physiological expressions of positive experiences in animals. Monitoring both the autonomic nervous system (via heart rate and its variability) and the immune system could offer relevant tools to better assess emotional states in animals, complementary to classical adrenocortical measures. In the second part of the paper, useful strategies for enhancing positive experiences (such as physical, social and cognitive enrichment or putative genetic selection) are outlined. Then this paper emphasizes practical applications for assessing and promoting positive emotions that may help in providing animals with a better quality of life. Play, affiliative behaviors and some vocalizations appear to be the most promising convenient indicators for assessing positive experiences in laboratory and farm animals under commercial conditions.

The effects of weight asymmetry and resource distribution on aggression in groups of unacquainted pigs

The relationship between weight asymmetry and aggression when mixing groups of unacquainted pigs in two different environments was investigated. Ten groups of 4 female (LandracexYorkshire), unacquainted pigs with a weight asymmetry of 3.1+/-0.2 kg (mean weight: 16.6+/-0.6 kg) between each pig, and another 10 groups with a weight asymmetry of 1.2+/-0.1 kg (mean weight: 14.1+/-0.1 kg), were mixed at the age of 7 weeks. Ten of the groups (five of each) were mixed in an experimental pen with a heterogeneous distribution of straw. The pen consisted of two main compartments with straw in one of them, and a passage area with concrete floor in between. The other 10 groups were mixed in the same pen, but with a homogenous distribution of straw (straw spread all over the pen).The results showed that fighting duration was significantly shorter in groups with large weight asymmetry than in groups with small weight asymmetry irrespective of the environment. The number of bites delivered during the fights in the heterogeneous environment was lower in groups with large weight asymmetry than in groups with small weight asymmetry. In the homogenous environment, however, there was no significant difference between groups with large and small weight asymmetries regarding the number of bites. The combination of a limited straw area and a small weight asymmetry resulted in the greatest number of bites. In groups with large weight asymmetry, the largest pig won around 50% of the fights, and 25% of the variation in the percentage of fights won was explained by the weight asymmetry.In groups with small weight asymmetry, less than 10% of the variation in the percentage of fights won could be explained by weight asymmetry. The two largest pigs delivered significantly more bites and spent significantly more time fighting than the smallest pig. The second largest pig received significantly more bites and body lesions than the smallest pig in the groups. Fights between the two largest pigs had a significantly longer duration than when other pigs were involved. The results were discussed in relation to sequential assessment theory and resource defence theory.

Melanin-based skin spots reflect stress responsiveness in salmonid fish

Within animal populations, genetic, epigenetic and environmental factors interact to shape individual neuroendocrine and behavioural profiles, conferring variable vulnerability to stress and disease. It remains debated how alternative behavioural syndromes and stress coping styles evolve and are maintained by natural selection. Here we show that individual variation in stress responsiveness is reflected in the visual appearance of two species of teleost fish; rainbow trout (Oncorhynchus mykiss) and Atlantic salmon (Salmo salar). Salmon and trout skin vary from nearly immaculate to densely spotted, with black spots formed by eumelanin-producing chromatophores. In rainbow trout, selection for divergent hypothalamus-pituitary-interrenal responsiveness has led to a change in dermal pigmentation patterns, with low cortisol-responsive fish being consistently more spotted. In an aquaculture population of Atlantic salmon individuals with more spots showed a reduced physiological and behavioural response to stress. Taken together, these data demonstrate a heritable behavioural-physiological and morphological trait correlation that may be specific to alternative coping styles. This observation may illuminate the evolution of contrasting coping styles and behavioural syndromes, as occurrence of phenotypes in different environments and their response to selective pressures can be precisely and easily recorded.

Aggression, fearfulness and coping styles in female pigs

Abstract The coping styles hypothesis states that animals have consistent behavioural and physiological coping responses that can be characterised on a continuum from reactive to proactive. According to this hypothesis, animals can be characterised by consistency in responses across social and non-social situations as well as over time. We evaluated the validity of this hypothesis for predicting the coping responses of pigs by testing for predicted correlations between responses to a variety of challenges. Animals were exposed to an immobility test at 3 weeks of age. At 8 and 24 weeks, animals were exposed to an unknown pig in the test animal’s home cage, a human in the home cage, and a novel object (bucket) in a novel arena. None of the predicted correlations were found between the duration of immobility and behaviour in the later tests. Many behavioural measures had some repeatability between the ages of 8 and 24 weeks, despite significant changes in the observed level of responses, but there were few correlations between responses to social and non-social situations. Responses to a human and novel object were positively correlated when measured at 8 weeks but not at 24 weeks. The results indicate that the coping styles hypothesis has very limited value in predicting the coping responses of pigs representing the population as a whole. The existing correlations could more parsimoniously be described as reflecting responses to specific or closely related stimuli. It is suggested that future research on pig personality should investigate the possibility that specific coping behaviours—for example, those mediated by fear and anxiety—may have consequences for different elements of coping ability related to welfare and productivity.

Relation between early fear- and anxiety-related behaviour and maternal ability in sows

Abstract Fear induced by exposure to humans, and anxiety induced by novel environmental changes may both function as stressors and have negative effects on reproduction in pigs. An earlier experiment indicated that reactions to humans and novelty had some repeatability between the ages of 8 and 24 weeks, thus reflecting personality traits. The present experiment therefore tested for predicted negative relationships between behaviour related to fear and anxiety measured at 8 weeks of age and later reproduction in 30 sows. Factor analysis of behaviour in a voluntary human approach test performed in the home cage and a novel object test performed in a novel arena indicated that behavioural measures related to investigation of these stimuli loaded on separate factors and may therefore reflect independent underlying dimensions. One factor related to investigation of the human was interpreted as reflecting the level of fear of humans, while another was correlated to investigation of the novel object and interpreted as indicating novelty-induced anxiety. Factor scores were generated for ‘fear’ and ‘anxiety’ and used in regression analysis as predictors of maternal behaviour and reproductive parameters. Higher levels of fear of humans were associated with longer durations of farrowing, larger variation in inter-birth intervals, and a higher number of piglets dying without milk in their stomachs. The duration of farrowing was positively correlated to the number of live born piglets dying within 3 weeks of age. Higher fear also tended to associate with a higher number of stillborn and a higher number of live born piglets dying within 3 weeks of age. High anxiety only tended to be associated with a higher number of stillborn. These results are interpreted as supporting the hypothesis that there are negative associations between fear of humans and maternal ability in sows.

Aggression and group size in domesticated pigs, Sus scrofa: 'when the winner takes it all and the loser is standing small'

Formation of artificial groups of domestic species occurs frequently in todays husbandry systems. These groups are kept in a closed space where the individuals are not able to withdraw from the group, and resources in the environment are most commonly limited and defensible. This creates a competitive environment where aggression and social stress are often seen. We developed a theoretical model that describes how aggression among unacquainted pigs is a function of group size. As the number of potential competitors increases, more individuals will benefit from not getting involved in costly fights. The reason for this is that the probability of being able to monopolize resources diminishes as group size increases. To verify predictions from this model, we compared aggression among unacquainted, weaned pigs in three group sizes (6, 12 and 24 pigs). There were significantly more fights per individual in groups of six and 12 pigs than in groups of 24 pigs, and the proportion of pigs not participating in fights was significantly higher in the largest group. For pigs that were willing to take the risk of getting involved in fights in the largest group size, the duration of each fight was longer than for fights in the smaller groups. These results support the predictions from our theoretical model, and we conclude that pigs are able to alter their behaviour according to how the actual competitive situation changes with group size.

Variation in behavioural responses of ewes towards predator-related stimuli

Thirty-two groups of six sheep, classified into three breed categories according to their weight class (L: light, one breed (n=7); M: medium light, two breeds (n=10); H: heavy, three breeds (n=15)) were tested for antipredatory behaviour towards seven stimulus regimes. Tests were performed on 2-5-years-old ewes with lambs inside standardised and fenced home pastures. Stimulus regimes were: man in rain poncho, trolley, ball on trolley (blind stimuli), stuffed wolverine on trolley, stuffed lynx on trolley, stuffed bear on trolley, and man in rain poncho with a dog on leash (carnivore stimuli). The L breed showed the longest recovery time, the longest flight distance and the tightest flocking behaviour. Significant regressions indicate that there were negative linear relationships between sheep weight and recovery time and between sheep weight and flight distance. Carnivore stimuli caused longer recovery times (P<0.001) and longer flight distances (P<0.001) than the blind stimuli. Our results confirm the hypothesis that lighter sheep breeds display stronger antipredatory reactions than heavier breeds.

Effects of environmental stressors on deep body temperature and activity levels in silver fox vixens (Vulpes vulpes)

Abstract The present study was performed to investigate the effects of 14 different environmental stimuli on stress-induced hyperthermia (SIH) and levels of locomotor activity in six (three infanticidal, three non-infanticidal) 2.5-year-old silver fox vixens. The effects of contact with humans (six experiments; handling for 5 min, handling of neighbouring animal for 5 min, presence of one person for 20 s, 5 and 90 min, presence of a group of humans for 5 min), exposure to unfamiliar foxes (four experiments; presence of an unfamiliar cagemate [female, male] and an unfamiliar neighbouring animal [female, male] for 90 min), and various recorded noise stimuli (four experiments; aircraft noise [duration 15 s, 100 dB], machine noise [duration 15 s, 90 dB], firing a shotgun [duration 1 s, 90 dB], human conversation [duration 15 s, 95 dB]) played back and repeated at 20 s intervals during 5 min were tested. Deep body temperature and activity levels were monitored with surgically implanted radio telemetry devices. All registrations were made during the 90-min period after stimulus presentation. The presence of humans and other silver foxes, but not exposure to loud recorded noise, resulted in a SIH. Comparison of the SIH between the normally reproducing vixens and the previously infanticidal vixens revealed significant differences. The SIH response was most pronounced in the previously infanticidal vixens, whereas the levels of physical activity were lowest in this group. The present study indicated that important means to improve animal welfare in silver foxes should include an improvement of the general human–animal relationship and emphasises the importance of a stable social environment.

Cortisol receptor expression differs in the brains of rainbow trout selected for divergent cortisol responses.

In rainbow trout (Oncorhynchus mykiss), selection for divergent post-stress plasma cortisol levels has yielded low (LR)- and high (HR) responsive lines, differing in behavioural and physiological aspects of stress coping. For instance, LR fish display prolonged retention of a fear response and of previously learnt routines, compared to HR fish. This study aims at investigating putative central nervous system mechanisms controlling behaviour and memory retention. The stress hormone cortisol is known to affect several aspects of cognition, including memory retention. Cortisol acts through glucocorticoid receptors 1 and 2 (GR1 and 2) and a mineralcorticoid receptor (MR), all of which are abundantly expressed in the salmonid brain. We hypothesized that different expressions of MR and GRs in LR and HR trout brains could be involved in the observed differences in cognition. We quantified the mRNA expression of GR1, GR2 and MR in different brain regions of stressed and non-stressed LR and HR trout. The expression of MR was higher in LR than in HR fish in all brain parts investigated. This could be associated with reduced anxiety and enhanced memory retention in LR fish. MR and GR1 expression was also subject to negative regulation by stress in a site-specific manner.

Peripheral temperature drop in response to anticipation and consumption of a signaled palatable reward in laying hens (Gallus domesticus).

The present study describes effects of anticipation and consumption of a palatable reward on comb surface temperature. The purpose was to investigate temperature responses as a potential physiological indicator of positive emotional states in laying hens. A rise in body temperature in response to stimuli predictive of or during exposure to unpleasant events has been interpreted as evidence of emotions in mammals and avians. However, this phenomenon has so far only been studied during anticipation of or exposure to negative events; i.e., emotions of a negative valence. Infrared thermography was used to record potential alterations in comb surface temperature to a conditioned cue signaling a reward (mealworms) and during reward delivery. On average, comb temperature dropped 1.5 °C (95% CI: +/-1.2 °C) after exposure to CS and consumption of reward (p~0.0014) when initial comb temperature was above 30 °C. Such temperature drop indicates a peripheral vasoconstriction and has clear resemblances to emotional fever as seen during negative emotional states. Thus, we propose that a drop in peripheral temperature reflects emotional arousal more than emotional valence. Substantial temperature responses due to diet-induced thermogenesis were found, further emphasizing a cautious interpretation of altered comb temperature in studies of animal welfare.