Explore the vast range of titles available.

Physical exercise can act as a physiological and a mental stressor. Monitoring exercise-induced stress is therefore essential to understand racehorses ‘responses to effort and to ensure their welfare. Stress perceived by the horse during physical activity can be measured using various indicators, including stress-related hormones such as cortisol and adrenaline, and other neuromodulators such as serotonin, all involved in the stress response and its regulation. Another approach to assess physiological and emotional responses to stimuli such as exercise is through behaviours and facial movements. In this study, we aimed to 1) evaluate the changes in these three hormones following a trotting exercise, 2) determine the changes in behaviour and facial movements in response to the same exercise and 3) investigate potential relationships between hormonal variations and specific behavioural patterns that could serve as indicators of exercise-induced stress in horses. Fourteen French Standardbred horses from two stables were monitored over one day. In the morning, they performed an interval training trotting exercise. Behaviours and facial movements were recorded via video for 2 min 30 both before and just after exercise. Saliva and blood samples were collected at four time points: before exercise, just after exercise, 1 h post-exercise and 24 h post-exercise to assess salivary cortisol, and serum concentrations concentration of adrenaline and serotonin. Results showed significant post-exercise increases in all three hormones with peak concentrations observed immediately after exercise, and elevated cortisol and adrenaline levels persisting one hour later. These variations are consistent with normal physiological responses to physical effort, reflecting activation of regulatory systems rather than necessarily indicating negative stress. However, inter-individual variability in the magnitude of these responses suggests that horses did not all experience the exercise in the same way, highlighting potential differences exercise-induced stress. In terms of behaviour, horses exhibited higher frequencies of facial movements, particularly mouth movements, after exercise than before. Moreover, increases in serotonin and adrenaline concentrations were positively associated with agitation-related behaviours (pawing and head turning) and mouth movements. Overall, our findings suggest that a behavioural profile characterised by increased agitation and mouth movements may reflect a post-exercise arousal response in French Standardbreds. These behaviours, in association with hormonal changes, could provide a useful non-invasive tool to assess to assess horses’ response to exercise, and potentially exercise-induced stress. However, further studies are needed to confirm this interpretation, particularly by investigating the potential effects of post-exercise management practices such as cross-tying, which may induce frustration.

This study investigated whether sheep ( Ovis aries ) discriminate the body odour of unfamiliar humans subjected to a contrastive emotion-inducing procedure. Axillary secretions were collected from 38 humans after they watched fearful or joyful films. The behavioural responses of adult ewes ( n  = 35) and 6-day-old lambs ( n  = 22) to these odours were assessed using a habituation-dishabituation paradigm validated in this species. Both ewes and lambs exhibited habituation (reduced tactile-olfactory investigation), but no significant dishabituation, limiting strong conclusions about odour discrimination. However, ewes displayed ear postures suggestive of differential emotional reactivity to assumed valence, but these effects were subtle and potentially influenced by test conditions. The lambs exposed to fear odour (FO), relative to those exposed to joy odour (JO), showed increased escape attempts and oral grasping of the odour pad, a conflicting pattern of responses that complicates interpreting FO as aversive. In sum, these results suggest that lambs may react to FO as a distinct stimulus, but they may not clearly associate it with negative valence. Ewes showed behavioural differences in response to odours, yet the absence of clear dishabituation makes it difficult to confirm discrimination ability. Further research including physiological measures is needed to clarify ovine perception of human emotional odours.

Behavioural and physiological markers of discrete positive emotions remain little investigated in animals. To characterise new markers in horses, we used tactile stimulations to induce emotional situation of contrasting valence. In the Gentle grooming group (G, N = 13) horses were gently groomed during 11 sessions on the body areas they appreciated the most. Horses in the Standard grooming group (S, N = 14) were groomed using a fixed procedure, reported to induce avoidance reactions in some horses. At session 11, G horses expressed significantly more contact-seeking behaviours than S horses, which expressed significantly more avoidance behaviours. This result suggests positive emotions in G horses and negative emotions in S horses. Blood cortisol, oxytocin, heart rate and heart rate variability never differed between before and after the grooming session. However, after the 11 sessions, basal oxytocin levels were lower in the G than in the S group. This difference was unexpected, but supports studies showing that a low level of basal oxytocin could be a marker of better well-being. Analyses of facial expressions during grooming revealed significant differences between groups. These expressions appear to be more sensitive than behavioural indicators because they alone enabled differentiating emotions according to the group when horses were re-exposed to neutral grooming one year after the treatment.

In this study, we explore machine learning models for predicting emotional states in ridden horses. We manually label the images to train the models in a supervised manner. We perform data exploration and use different cropping methods, mainly based on Yolo and Faster R-CNN, to create two new datasets: 1) the cropped body, and 2) the cropped head dataset. We train various convolutional neural network (CNN) models on both cropped and uncropped datasets and compare their performance in emotion prediction of ridden horses. Despite the cropped head dataset lacking important regions like the tail (commonly annotated by experts), it yields the best results with an accuracy of 87%, precision of 79%, and recall of 97%. Furthermore, we update our models using various techniques, such as transfer learning and fine-tuning, to further improve their performance. Finally, we employ three interpretation methods to analyze the internal workings of our models, finding that LIME effectively identifies features similar to those used by experts for annotation.

Social animals can gain important benefits by inferring the goals behind the behavior of others. However, this ability has only been investigated in a handful of species outside of primates. In this study, we tested for the first time whether domestic horses can interpret human actions as goal directed. We used the classical “unwilling versus unable” paradigm: an experimenter performed three similar actions that have the same outcome, but the goal of the experimenter differed. In the unwilling condition, the experimenter had no intention to give a piece of food to a horse and moved it out of reach when the horse tried to eat it. In the two unable conditions, the experimenter had the intention to give the food to the horse but was unable to do so, either because there was a physical barrier between them or because of the experimenter’s clumsiness. The horses (n = 21) reacted differently in the three conditions: they showed more interest in the unable conditions, especially in the unable clumsy condition, than in the unwilling condition. These results are similar to results found in primates with the same paradigm and suggest that horses might have taken the experimenter’s goal, or even intentions, into account to adapt their behavior. Hence, our study offers more insights into horse interspecific social cognition towards humans.

When interacting with humans, domesticated species may respond to communicative gestures, such as pointing. However, it is currently unknown, except for in dogs, if species comprehend the communicative nature of such cues. Here, we investigated whether horses could follow the pointing of a human informant by evaluating the credibility of the information about the food-hiding place provided by the pointing of two informants. Using an object-choice task, we manipulated the attentional state of the two informants during food-hiding events and differentiated their knowledge about the location of the hidden food. Furthermore, we investigated the horses’ visual attention levels towards human behaviour to evaluate the relationship between their motivation and their performance of the task. The result showed that horses that sustained high attention levels could evaluate the credibility of the information and followed the pointing of an informant who knew where food was hidden (Z = − 2.281, P = 0.002, n = 36). This suggests that horses are highly sensitive to the attentional state and pointing gestures of humans, and that they perceive pointing as a communicative cue. This study also indicates that the motivation for the task should be investigated to determine the socio-cognitive abilities of animals.

The welfare of an animal is closely linked to their emotional experiences, making it essential to identify reliable indicators of these emotions. This study aimed to identify behaviours and facial movements in horses experiencing contrasting emotional valence, triggered by the anticipation of a positive condition (going to pasture) or a negative condition (going alone to a novel environment). Twenty horses were daily trained to wait in a starting box before being exposed to these two conditions. After one week of positive training or negative training, we analysed horses’ behaviours, cortisol variations, and facial movements while they waited in their starting box. First, we confirmed that the two conditions induced contrasting emotional valence, as evidenced by the shorter time taken to approach in the positive condition compared to the negative, and by the higher maximal heart rate and cortisol variation in the negative condition. Then using the Equine Facial Action Coding System (EquiFACS) and network analysis (NetFACS) we revealed distinct behaviours and facial expression profiles. In positive anticipation, the horses exhibited a greater range of behaviours, including shaking their heads from side to side, stepping back, sniffing, and pawing at the ground. Additionally, two distinct facial expression profiles were identified as specific to positive and negative anticipation. In positive anticipation, the horses displayed a higher neck, accompanied by a greater frequency of half-blinks and mouth movements. Conversely, in negative anticipation, the horses exhibited a medium neck, with ears backward accompanied by more flattened ears and expressed more nostril movements. The findings highlight the importance of these indicators in characterising horses’ emotions and emphasise their significance for assessing equine welfare.

Several studies have shown that horses have the ability to cross-modally recognize humans by associating their voice with their physical appearance. However, it remains unclear whether horses are able to differentiate humans according to different criteria, such as the fact that they are women or men. Horses might recognize some human characteristics, such as sex, and use these characteristics to classify them into different categories. The aim of this study was to explore whether domesticated horses are able to cross-modally recognize women and men according to visual and auditory cues, using a preferential looking paradigm. We simultaneously presented two videos of women and men’s faces, while playing a recording of a human voice belonging to one of these two categories through a loudspeaker. The results showed that the horses looked significantly more towards the congruent video than towards the incongruent video, suggesting that they are able to associate women’s voices with women’s faces and men’s voices with men’s faces. Further investigation is necessary to determine the mechanism underlying this recognition, as it might be interesting to determine which characteristics horses use to categorize humans. These results suggest a novel perspective that could allow us to better understand how horses perceive humans.

Social interactions shape both the physiological and behavioural development of offspring, and poor care/early caregiver loss is known to promote adverse outcomes during infancy in both animals and humans. How affiliative behaviours impact the future development of offspring remains an open question. Here, we used Equus caballus (domestic horse) as a model to investigate this question. By coupling magnetic resonance imaging, longitudinal biobehavioural assessments and advanced multivariate statistical modelling, we found that prolonged maternal presence during infancy promotes the maturation of brain regions involved in both social behaviour (anterior cingulate cortex and retrosplenial cortex) and physiological regulation (hypothalamus and amygdala). Additionally, offspring benefiting from a prolonged maternal presence showed higher default mode network connectivity, improved social competences and feeding behaviours, and higher concentrations of circulating lipids (triglyceride and cholesterol). The findings of the present study underscore the salient role of social interactions in the development of allostatic regulation in offspring. The present study shows that maternal presence beyond early life remains crucial for brain, behavioural and physiological development in young horses, highlighting the importance of the mother–offspring relationship during a childhood-like stage.

In the past 20 years, research focusing on interspecific sociocognitive abilities of animals toward humans has been growing, allowing a better understanding of the interactions between humans and animals. This review focuses on five sociocognitive abilities of domestic mammals in relation to humans as follows: discriminating and recognizing individual humans; perceiving human emotions; interpreting our attentional states and goals; using referential communication (perceiving human signals or sending signals to humans); and engaging in social learning with humans (e.g., local enhancement, demonstration and social referencing). We focused on different species of domestic mammals for which literature on the subject is available, namely, cats, cattle, dogs, ferrets, goats, horses, pigs, and sheep. The results show that some species have remarkable abilities to recognize us or to detect and interpret the emotions or signals sent by humans. For example, sheep and horses can recognize the face of their keeper in photographs, dogs can react to our smells of fear, and pigs can follow our pointing gestures. Nevertheless, the studies are unequally distributed across species: there are many studies in animals that live closely with humans, such as dogs, but little is known about livestock animals, such as cattle and pigs. However, on the basis of existing data, no obvious links have emerged between the cognitive abilities of animals toward humans and their ecological characteristics or the history and reasons for their domestication. This review encourages continuing and expanding this type of research to more abilities and species.