Researchers in France have delved into the complexities of learning in horses, unearthing fresh insights around the impact of stress, the effects of positive and negative reinforcement, and the influence of personality on the outcomes.
Mathilde Valenchon and her colleagues from the French National Institute for Agricultural Research (INRA) found that exposure to stress beforehand harmed learning performance, mainly under conditions of positive reinforcement.
But the findings from the study team painted a more complex picture.
They found that learning performance appeared to be differentially related to personality according to the type of reinforcement used and the presence of stress factors unrelated to the learning task.
Without these external stressors, the horses assessed as most fearful in personality tests were the best performers when they learned with negative reinforcement but were consistently the worst performers when they learned with positive reinforcement.
The researchers used 60 Welsh mares in the study, each of which had undergone a well-recognised personality test beforehand in which their fearfulness, gregariousness, reactivity to humans, level of locomotor activity and sensitivity to touch were assessed.
They were then assigned into four groups of 15 horses each. Each group received training in a task that involved the horses entering one of two compartments at the visual signal given by the experimenter.
Two groups learned the task with negative reinforcement and two learned it with positive reinforcement. However, one of the negative-reinforcement groups did so immediately after exposure to an external source of stress, as did one of the positive reforcement groups.
The stressors were either loud noises (for example, dog barks, bell ringing, people talking loudly), surprise movements (shaking of a tarpaulin, or a water or air jet emitted toward the horse), or the introduction of an unfamiliar object, such as a colourful cardboard box or balloons.
In the absence of stressors unrelated to the task, learning performance did not differ between negative and positive reinforcements, the researchers reported. However, the use of external stress sources was found to impaired learning performance.
“Interestingly, this learning deficit was smaller when the negative reinforcement was used,” they reported in the peer-reviewed open-access journal PLOS ONE.
Negative reinforcement, which is considered as a stressor related to the task, could have counter-balanced the impact of the external stressor by focusing attention toward the learning task, they suggested.
They also reported that learning performance appeared to differ between certain dimensions of personality depending on the presence of stressors and the type of reinforcement.
The results suggested that when negative reinforcement was used − that’s a stressor related to the task − the most fearful horses were the best performers when no external stressors were used. However, they were the worst performers when external stressors were present.
“On the contrary, when positive reinforcement is used, the most fearful horses appear to be consistently the worst performers, with and without exposure to stressors unrelated to the learning task.”
It was possible that the positive effect of negative reinforcement on performance, through the focus of attention toward the task, might have been accentuated in fearful individuals and could explain their improved performance in the absence of stressors unrelated to the task.
“By contrast, there were no stressors related to the learning task in the positive reinforcement group, and the fearful horses might have been more easily distracted by the external environment.
“When stressors unrelated to the task were added, learning performance of the most fearful horses appeared to be consistently impaired, independent from the type of reinforcement.
“The disruption of cognitive and attentional processes caused by the stressors unrelated to the task is likely to be more pronounced in fearful individuals who react more strongly to various forms of stressors.”
The study indicated that other dimensions of personality may also influence learning performance.
“Locomotor activity was positively related to performance in negative reinforcement learning, independent from exposure to stressors unrelated to the task,” they noted.
They hypothesized that locomotor activity may broadly reflect a tendency to initiate actions, since they previously observed such a positive effect with learning tasks that required a displacement of body position similar to the present task.
In addition, reactivity to humans was negatively related to learning performance only in the negative reinforcement and external stress group.
“The horses less frightened by humans might have been less affected by the stressors unrelated to the task because they were more attentive to human cues.”
They said that with 15 animals per group, only a few significant correlations between variables of personality and learning performance were revealed.
“To reveal the whole influence of personality on learning abilities, it would be ideal to perform this kind of study on a larger amount of subjects, maybe several hundred. Unfortunately, it is quite impossible to test this number of subjects within the equine species.
“However, we have to underline the fact that, taken together, all the experiments carried out on horses reveal consistent links between personality and learning abilities. It would mean that these correlations are not obtained by chance, but more likely reflect a real influence of personality.”
The researchers said they believed their study was the first to demonstrate in ungulates that stress affected learning performance differentially according to the type of reinforcement and in interaction with personality.
“Our study reveals the existence of relationships between personality and learning performance,” they said.
It contributed to a better understanding of the influence of stress on learning performance by showing the importance of the nature of the stress (related or unrelated to the task) and personality.
“Our results also provide important clues to more personalized training according to each animal’s needs.
“Indeed, the present study shows that positive and negative reinforcement may lead to equivalent learning performance in the absence of stressors unrelated to the task.
“Considering that previous studies highlighted long-lasting promising effects of positive reinforcement on training, welfare and relationships with humans, our results provide additional evidence for promoting the use of positive reinforcement in horse training, while the use of negative reinforcement still dominates traditional training methods.”
However, the present results also show that the use of food rewards as reinforcement may not be adequate in stressful conditions.
“Indeed, the loss of food motivation induced by stress could render the reinforcement inefficient and constitute an additional source of stress.”
They continued: “Our analyses including personality suggest that stress is a key factor in understanding how animals differ in learning performance according to their personality. Predicting how fearful animals react when they face a learning challenge therefore not only requires an evaluation of stress level but also of the nature of the stress (related or unrelated to the task) and their possible interactions.”
They said future investigations were needed to define more precisely when the switch from favourable to unfavourable conditions for learning occurred, according to the type and the intensity of a stressor and how it related to personality.
The study team comprised Valenchon , Frédéric Lévy, Chantal Moussu and Léa Lansade.
Valenchon M, Lévy F, Moussu C, Lansade L (2017) Stress affects instrumental learning based on positive or negative reinforcement in interaction with personality in domestic horses. PLoS ONE 12(5): e0170783. https://doi.org/10.1371/journal.pone.0170783