Marianne Farish

Genetic validation of postmixing skin injuries in pigs as an indicator of aggressiveness and the relationship with injuries under more stable social conditions

The objective of the study was to estimate genetic correlations between skin lesions and aggressive behavior postmixing and under more stable social conditions as a potential means of selecting against pig aggressiveness. Postmixing aggression in commercial pig production is common, compromises welfare and profitability, and cannot be significantly reduced by low-cost changes to the environment. A genetic component to individual aggressiveness has been described in pigs and other species. Selective breeding against aggressiveness ought to be possible if an easily measured indicator trait can be shown to be genetically associated with aggressive behavior. Aggressive behavior was recorded continuously for 24 h after mixing, and a count of skin lesions (lesion count, LC) was recorded at 24 h and 3 wk postmixing on 1,663 pigs. Two behavioral traits were found to have a moderate to high heritability similar to that of growth traits; duration of involvement in reciprocal fighting (0.43 +/- 0.04) and delivery of nonreciprocal aggression (NRA; 0.31 +/- 0.04), whereas receipt of NRA had a lower heritability (0.08 +/- 0.03). Genetic correlations (r(g)) suggested that lesions to the anterior region of the body 24 h after mixing were associated with reciprocal fighting (r(g) = 0.67 +/- 0.04), receipt of NRA (r(g) = 0.70 +/- 0.11), and to a lesser extent, delivery of NRA (r(g) = 0.31 +/- 0.06). Lesions to the center and rear were primarily genetically associated with receipt of NRA (r(g) = 0.80 +/- 0.05, 0.79 +/- 0.05). Genetic correlations indicated that pigs that engaged in reciprocal fighting delivered NRA to other animals (r(g) = 0.84 +/- 0.04) but were less likely to receive NRA themselves (r(g) = -0.41 +/- 0.14). A genetic merit index using lesions to the anterior region as one trait and those to the center or rear or both as a second trait should allow selection against animals involved in reciprocal fighting and the delivery of NRA. Positive correlations between LC 24 h and 3 wk after mixing were found, especially for lesions to the center and rear of the body, indicating that postmixing lesions are predictive of those received under more stable group conditions. As well as reducing immediate aggression at mixing, selection on postmixing LC is expected to have a long-term impact on injuries from aggression, even after dominance relationships are established.

Aggressiveness as a component of fighting ability in pigs using a game-theoretical framework

Understanding animal contests has benefited greatly from employing the concept of fighting ability, termed resource-holding potential (RHP), with body size/weight typically used as a proxy. However, victory does not always go to the larger/heavier contestant and the existing RHP approach thereby fails to accurately predict contest outcome. Aggressiveness, typically studied as a personality trait, might explain part of this discrepancy. We investigated whether aggressiveness forms a component of RHP, examining effects on contest outcome, duration and phases, plus physiological measures of costs (lactate and glucose). Furthermore, using the correct theoretical framework, we provide the first study to investigate whether individuals gather and use information on aggressiveness as part of an assessment strategy. Pigs, Sus scrofa, were assessed for aggressiveness in resident–intruder tests whereby attack latency reflects aggressiveness. Contests were then staged between size-matched animals diverging in aggressiveness. Individuals with a short attack latency in the resident–intruder test almost always initiated the first bite and fight in the subsequent contest. However, aggressiveness had no direct effect on contest outcome, whereas bite initiation did lead to winning in contests without an escalated fight. This indirect effect suggests that aggressiveness is not a component of RHP, but rather reflects a signal of intent. Winner and loser aggressiveness did not affect contest duration or its separate phases, suggesting aggressiveness is not part of an assessment strategy. A greater asymmetry in aggressiveness prolonged contest duration and the duration of displaying, which is in a direction contrary to assessment models based on morphological traits. Blood lactate and glucose increased with contest duration and peaked during escalated fights, highlighting the utility of physiological measures as proxies for fight cost. Integrating personality traits into the study of contest behaviour, as illustrated here, will enhance our understanding of the subtleties of agonistic interactions.

Complex contests and the influence of aggressiveness in pigs

Animal contests vary greatly in behavioural tactics used and intensity reached, with some encounters resolved without physical contact while others escalate to damaging fighting. However, the reasons for such variation remains to be fully explained. Aggressiveness, in terms of a personality trait, offers a potentially important source of variation that has typically been overlooked. Therefore, we studied how aggressiveness as a personality trait influenced escalation between contestants matched for resource-holding potential (RHP), using detailed observations of the contest behaviour, contest dynamics and escalation levels. We predicted that winner and loser behaviour would differ depending on personality. This was tested by examining 52 dyadic contests between pigs, Sus scrofa. Aggressiveness was assayed in resident–intruder tests prior to the contest. Contests were then staged between pigs matched for RHP in terms of body weight but differing in their aggressiveness. In 27% of the contests a winner emerged without escalated physical fighting, demonstrating that a fight is not a prerequisite between RHP-matched contestants. However, the duration of contests with or without fighting was the same. In contests without a fight, opponents spent more time on mutual investigation and noncontact displays such as parallel walking, which suggests that ritualized display may facilitate assessment and decision making. Winners low in aggressiveness invested more time in opponent investigation and display and showed substantially less aggression towards the loser after its retreat compared to aggressive winners. Aggressiveness influenced contest dynamics but did not predict the level of escalation. Prominent behavioural differences were found for the interaction between personality and outcome and we therefore recommend including this interaction in models where personality is considered. Analyses based on contest duration only would miss many of the subtleties shown here and we therefore encourage more detailed analyses of animal contests, irrespective of the level of contest escalation.

Towards on-farm pig face recognition using convolutional neural networks

Abstract Identification of individual livestock such as pigs and cows has become a pressing issue in recent years as intensification practices continue to be adopted and precise objective measurements are required (e.g. weight). Current best practice involves the use of RFID tags which are time-consuming for the farmer and distressing for the animal to fit. To overcome this, non-invasive biometrics are proposed by using the face of the animal. We test this in a farm environment, on 10 individual pigs using three techniques adopted from the human face recognition literature: Fisherfaces, the VGG-Face pre-trained face convolutional neural network (CNN) model and our own CNN model that we train using an artificially augmented data set. Our results show that accurate individual pig recognition is possible with accuracy rates of 96.7% on 1553 images. Class Activated Mapping using Grad-CAM is used to show the regions that our network uses to discriminate between pigs.

The influence of experience on contest assessment strategies

Animal contest behaviour has been widely studied, yet major knowledge gaps remain concerning the information-gathering and decision-making processes used during encounters. The mutual assessment strategy, where the individual assesses its own fighting ability (Resource Holding Potential, RHP) and compares it to that of its opponent, is least understood. We hypothesise that individuals need experience of agonistic encounters to become proficient at mutual assessment. Pigs (Sus scrofa, n = 316) were contested twice. In between contests, animals did or did not (control) receive intense fighting experience. A substantial proportion of the contests reached an outcome with a clear winner without fighting. Non-escalation was highest in RHP asymmetric dyads of the second contest, irrespective of experience. In contest 1 (no experience) and in contest 2 for the experienced animals, costs increased with loser RHP and where unaffected by winner RHP, suggesting a self-assessment strategy. In contest 2 control dyads, which only had experience of one prior contest, a negative relation between winner RHP and costs suggested mutual assessment during the pre-escalation phase but not during escalated aggression. This reveals that a brief and relatively mild experience can be beneficial in the development of mutual assessment whereas profound experience may result in adoption of a self-assessment strategy.

Intensity of aggression in pigs depends on their age and experience at testing

Faces are an essential source of information for social species. Sheep are one of the most studied farm species in terms of their ability to process information from faces, but little is known about their face-based emotion recognition abilities. We investigated whether sheep could use images of faces displaying different emotional states as cues in a simultaneous discrimination task. To that end, we took photos of faces of four sheep in three social situations: two with a negative valence (social isolation or aggressive interaction) and a neutral situation (ruminating in the home pen). In a two-armed maze, sheep (n=16) were then presented with pairs of images of the same familiar individual taken in the neutral situation and one of the negative situations. Sheep had to learn to associate one type of image from a pair with a food reward. Once they had reached the learning criterion, sheep then had to generalise the task to new pairs of images of different conspecifics displaying the same emotions as in the previous phase. For every run in the maze, the latency to choose an arm and the outcome of the choice (success or error) were recorded, as well as the total number of runs needed to learn the task (learning speed). Influence of the type of image rewarded and of the side of presentation of the rewarded image were analysed by linear mixed models and learning speed was analysed by Mood’s median test. All sheep learned the task with images of faces. Sheep that had to associate a negative image with a reward learned faster that sheep that had to learn the neutral-reward combination (medians: 40 vs 75 runs, χ2=4.00, df=1, P=0.046). With the exception of sheep from the Aggression-rewarded group, sheep could generalise the discrimination task to images of new faces (F3,26.6=3.37, P=0.033). Sheep chose an arm correctly more often (F1,288.1=5.02, P=0.026) and more quickly (F1,289.3=23.92, P<0.001; right: 8.4±0.6 sec, left: 9.6±0.8 sec) when the rewarded image was displayed on the right side, suggesting the influence of a right hemisphere/left visual field bias in face-based perception of emotions. Our results strongly suggest that sheep can perceive the emotional valence displayed on faces and that this valence affects learning processes.

Lateralization influences contest behaviour in domestic pigs

Cerebral lateralization, i.e. hemispheric asymmetries in structure and function, relates in many species to a preference to attack from their left. Lateralization increases cognitive capacity, enabling the simultaneous processing of multiple sources of information. Therefore, lateralization may constitute a component of fighting ability (Resource Holding Potential), and/or influence the efficiency of information-gathering during a contest. We hypothesized that lateralization will affect contest outcome and duration, with an advantage for more strongly lateralized individuals. In 52 dyadic contests between weight-matched pigs (Sus scrofa; n = 104; 10 wk age), the direction of orientation towards the opponent was scan sampled every 10 s. Laterality indexes (LI) were calculated for the direction and strength of lateralization. Up to 12.5% of the individuals showed significant lateralization towards either the right or left but lateralization was absent at the population level. In line with our hypothesis, animals showing strong lateralization (irrespective of direction) had a shorter contest duration than animals showing weak lateralization. Winners did not differ from losers in their strength or direction of lateralization. Overall the results suggest that cerebral lateralization may aid in conflict resolution, but does not directly contribute to fighting ability, and will be of value in the study of animal contests.

Socialisation, play behaviour, and the development of aggression in domestic pigs

Faces are an essential source of information for social species. Sheep are one of the most studied farm species in terms of their ability to process information from faces, but little is known about their face-based emotion recognition abilities. We investigated whether sheep could use images of faces displaying different emotional states as cues in a simultaneous discrimination task. To that end, we took photos of faces of four sheep in three social situations: two with a negative valence (social isolation or aggressive interaction) and a neutral situation (ruminating in the home pen). In a two-armed maze, sheep (n=16) were then presented with pairs of images of the same familiar individual taken in the neutral situation and one of the negative situations. Sheep had to learn to associate one type of image from a pair with a food reward. Once they had reached the learning criterion, sheep then had to generalise the task to new pairs of images of different conspecifics displaying the same emotions as in the previous phase. For every run in the maze, the latency to choose an arm and the outcome of the choice (success or error) were recorded, as well as the total number of runs needed to learn the task (learning speed). Influence of the type of image rewarded and of the side of presentation of the rewarded image were analysed by linear mixed models and learning speed was analysed by Mood’s median test. All sheep learned the task with images of faces. Sheep that had to associate a negative image with a reward learned faster that sheep that had to learn the neutral-reward combination (medians: 40 vs 75 runs, χ2=4.00, df=1, P=0.046). With the exception of sheep from the Aggression-rewarded group, sheep could generalise the discrimination task to images of new faces (F3,26.6=3.37, P=0.033). Sheep chose an arm correctly more often (F1,288.1=5.02, P=0.026) and more quickly (F1,289.3=23.92, P<0.001; right: 8.4±0.6 sec, left: 9.6±0.8 sec) when the rewarded image was displayed on the right side, suggesting the influence of a right hemisphere/left visual field bias in face-based perception of emotions. Our results strongly suggest that sheep can perceive the emotional valence displayed on faces and that this valence affects learning processes.