At the 19th ISES Conference in New Zealand, Dr. Cathrynne Henshall presented where we are on Detecting positive emotions in horses to assess their overall quality of life. As an expert in equine behaviour and welfare, Dr. Henshall’s talk covered how subtle emotional cues in horses can reveal a lot about their well-being. Her presentation focused on the importance of understanding horses’ subjective experiences and how recognizing these emotional states can enhance the way we approach equine care and management.

This article covers topics Dr. Henshall discussed, such as the subjective experience of horses, emotions, moods and arousal, indicators of positive/negative emotions, the role of agency and free choice and AI in emotion detection.

The Subjective Experience of Horses

All humans and animals experience life subjectively, seeking positive experiences and avoiding negative ones. This understanding is key to modern welfare assessments, which look further than just good health and nutrition to include the type, intensity, and duration of experiences. These assessments focus on the balance between positive and negative experiences to work out whether a horse has a “good life.” Researchers can evaluate this balance by determining the ratio of positive to negative experiences.

Dr. Cathrynne Henshall states, “Welfare is not just about good health and nutrition, but about the subjective experience they have.”

Emotions and Moods: Short and Long-Term States

Emotions and moods in horses are usually determined as either positive or negative. Emotions are short-term psychological states with a few elements: a subjective experience (how the emotion feels), a behavioural aspect (how the emotion is expressed), and a cognitive component (how it influences decision-making). These emotional states can change rapidly, which influences how a horse responds to situations.

For example, a horse might show signs of fear, happiness, or excitement in the moment, which can be observed through their behaviour, but it can be challenging for researchers to accurately link behaviours to specific emotional states as horses can’t self-report, unlike humans who can communicate via verbal communication. However, horses can still make decisions based on their emotional states, such as avoiding or approaching certain situations based on past experiences.

Moods, on the other hand, are longer-term psychological states that accumulate over time because of multiple emotional experiences. In recent research, optimism and pessimism have been explored as indicators of mood in horses. Horses that are optimistic tend to approach uncertain situations because they expect there to be a positive outcome, whereas pessimistic horses tend to avoid them because they are expecting a negative outcome.

Horses in heightened states of arousal during activities like jumping may be mistakenly assumed to be enjoying the experience via indicators such as ears pricked forwards or increasing speed towards the jump, even though these behavioural indicators could reflect their responses to the physical or psychological stress of the situation, rather than pleasure.

Moods, Emotions, and Arousal in Horses

The connection between moods, emotions, and arousal is a key aspect of understanding a horse’s psychological state. Arousal refers to a physiological response to stimuli and is linked to emotions, but it does not have a particular valence (positive or negative). For example, high arousal can indicate fear or positive excitement, while low arousal might indicate depression or simply lethargy.

While researchers have made huge progress in understanding individual emotions at certain points, it is more challenging to work out how the accumulation of emotions and moods over long periods of time impacts a horse’s quality of life. This requires further research, especially as horses’ emotional states may not be clear from their behaviour.

Detecting Positive Emotions: The Challenges

Dr. Henshall said one of the biggest challenges in equine research is the detection of positive emotions, which are often subtle and harder to identify than negative emotions. While emotions like fear or aggression are easier to observe, positive emotional states may be presented through more subtle behavioural cues.

Despite existing research, there are not currently widely accepted, validated indicators of positive emotional valence in horses. The subtlety of these emotions can lead to potential misinterpretation. For example, horses in heightened states of arousal during activities like jumping may be mistakenly assumed to be enjoying the experience via indicators such as ears pricked forwards or increasing speed towards the jump, even though these behavioural indicators could reflect their responses to the physical or psychological stress of the situation, rather than pleasure.

The difficulty of detecting different emotional states in horses was demonstrated by images of a horse taking part in Dr Henshall’s recent experiment exploring facial indicators of emotions in Thoroughbred horses. In this experiment (seen in figure 1 below), the horse was trained to touch a plastic square, triggering a sensor that dispenses food. After extensive training, the horses underwent an extinction procedure in which they were no longer rewarded with food if they touched the panel. In these images, the horse first touches the square, but then remembers that no food will come, so he simply rests his nose there for a while (A). In the next picture (B), the response is reinstated as the horse anticipates food, preparing to touch the sensor, followed by the consumption of the reward. The last image captures the moment right after the food is given (C).

Subtle differences in the shape of the musculature around the horse’s eyes, ear position and orientation, as well as the position and shape of the lips, can be used to infer different emotional states; however, this requires access to recording equipment and the time to spend analyzing the data. This highlights the difficulty of detecting emotional valence in real-time situations.

Three pictures of a grey horse.

Figure 1: Pictures from a trial during Dr. Henshall’s experiment (Photo credit: Dr Cathrynne Henshall).

The Role of Agency and Free Choice in Emotional Expression

Agency, or the ability of horses to make choices freely, plays a significant role in how they express emotions. Studies have shown that when horses have more control over their environment or actions, they are more likely to display behaviours that researchers can link to positive emotions. For example, play behaviour is often seen as an indicator of positive emotional states. However, when horses’ choices are restricted in research settings, their natural behaviour may be suppressed, making it harder to detect true emotional responses.

Indicators of Positive and Negative Emotional States

Researchers use several methods to detect emotional states in horses, ranging from facial expressions to body language and vocalizations. One of the most reliable methods for identifying subtle behavioural indicators from which the underlying emotional state can be inferred is the observation of facial expressions. Tools like EquiFACS (Equine Facial Action Coding System) allow researchers to record and analyze small changes in facial expressions, which can provide valuable insights into the emotional state of a horse.

Whole-body behaviours, prosocial behaviours, and vocalizations such as whinnies have also been studied. Recent research suggests that facial expressions are sensitive to quick emotional changes, making them one of the most effective tools for detecting how a horse feels in real time.

Positive Emotions in Ridden Horses: The “Holy Grail”

Detecting positive emotions in ridden horses is an area of great interest for researchers and equine professionals. The question of whether certain ridden activities — like jumping or dressage — induce positive emotions is still unresolved. While exercise can lead to the release of neurochemicals associated with positive affect, such as dopamine and endorphins, researchers are cautious about assuming that all ridden activities are inherently enjoyable for horses.

Dr Cathrynne Henshall states, “Exercise may induce positive affect, but the intensity, duration, and perception of control over the exercise are key factors that determine whether the experience is positive for the horse.”

This means that while exercise may trigger positive neurochemical responses, whether the horse perceives the activity as enjoyable remains an open question.

The Potential of AI in Emotion Detection

Artificial intelligence (AI) holds high potential for improving the detection of emotions in horses. By analyzing large datasets such as video footage, AI can detect subtle emotional indicators that might usually be missed. This could significantly enhance the accuracy and sensitivity of emotion detection in equine research.

The absence of negative emotions does not mean that a horse is experiencing happiness or enjoyment.

However, there are limitations and concerns with AI, especially when it comes to explainability. Many AI models function as “black boxes,” meaning that researchers may not fully understand how the AI is making its decisions. This lack of transparency raises questions about the reliability of AI in detecting emotional states, particularly when subtle behaviours are involved. Additionally, if the data used to train AI models is flawed or incomplete, the results may be inaccurate.

Applying the Concept of Telos

To truly understand and detect positive emotions in horses, researchers need to consider the concept of telos — the essence of what it means to be a horse. This concept focuses on the natural characteristics, behaviours, and preferences of horses, and aligns research methods with their inherent needs and instincts.

By applying the telos lens, researchers can create environments and activities that facilitate natural behaviours, which in turn may help identify when a horse is more likely to experience positive emotions. Understanding the unique characteristics of horses, such as their ethological, cognitive, and physiological traits, provides a strong foundation for detecting emotional states in various contexts.

Conclusion: Are Horses Enjoying Themselves?

While significant progress has been made in identifying and reducing negative emotional states in horses, the detection of positive emotions remains more complex. The absence of negative emotions does not mean that a horse is experiencing happiness or enjoyment.

The challenge lies in interpreting subtle behavioural cues and ensuring that research methodologies are sensitive enough to capture positive states.

Ultimately, the question remains: Do neutral emotional states during activities like training and competition mean that horses have good welfare while taking part in those activities? Should we aim to create conditions that actively promote positive emotions in horses? Understanding how to detect and enhance positive emotional experiences in horses will be crucial for improving their welfare and ensuring they lead fulfilling lives.

Become a member of the International Society for Equitation Science for access to the entire conference and exclusive equitation science content.