Department of Physics, Chemistry and Biology (IFM)

The factor Friendliness was calculated with the scores of questions referring to antagonistic behaviors in the context of interactions with both donkeys and humans. Nervousness on the other hand incorporates questions that classified the individuals on their comfort with the environment in challenging situations. However, these factors are interrelated, and both refer to broader personality traits, deeming the classification ambivalent. 

It is predicted that the results reported here would be different if tests were conducted in an isolated and more neutral environment, especially if we take into consideration that donkeys might be more sensitive to weather changes than other equids. Younger individuals possibly have different time budgets and display differing behaviors (in this case observed with Vigilance scores), when compared to adults - which was not controlled for in this study. Additional research is necessary to further develop these questions, as equid literature is very scarce on the topic. 

An earlier study with donkeys revealed that their short-term memory extended for thirty seconds. The memory tests reported here indicate that miniature donkeys are capable of retaining information about the location of hidden objects, within a fairly complex experimental set, for at least one minute. In fact, the actual extent of their prospective memory is likely to be much bigger, and the reason why only half of the individuals successfully finished the task might be related to environmental distractions or problems with the experimental set up. These results are in concordance with findings about short-term memory abilities of horses.

The correlation found between Predictability and Memory may suggest that when individuals are able to better recall events, they are also more predictable. As most unpredictable behaviors, in this context, were reported to be related to fear responses, low scores for Predictability were most likely attributed to insecure individuals that, despite their progress during training on one day, did not increase their confidence and comfort for the next days, in the experimental setting. If true, the latter hypothesis might find its explanation in the individual’s memory capacity to register information and retrieve it later (e.g. the fact that the experimental procedure is nonthreatening rewarded). Additionally, fearful horses have been found to perform worse than nonfearful horses, in a memory task, under stressful conditions - which would explain the results in this study, if assumed that unpredictability was related to fear. 

As predicted, donkeys were able to learn to discriminate between a positive and a negative stimulus. The removal of food in the bucket has proven to be enough to promote discrimination learning, deeming unnecessary to use negative reinforcement for this kind of tests. Although not quantified in this study, time spent at the positive location during negative and ambiguous trials, suggested that spatial location cues might play a more important role than colour discrimination. Though horses have dichromatic vision and are able to distinguish brightness levels, additional studies would be necessary to confirm that colour discrimination was even incorporated in the study donkeys’ cognitive process to solve this task.

Individuals differed from each other in their judgement bias, as exhibited in their different latencies to approach an ambiguous stimulus. Since there were no treatment differences, the variation found between individuals must be attributed to either life history, current housing situation, personality, or a combination of these. The positive correlation found between Patience and the Judgement bias score indicates that the individuals classified as more patient were also more pessimistic. Patience however, as interpreted by the observers, might instead reflect reduced nervousness or low exploration rates, for example. There is also a possibility that this personality factor reflects that individuals are generally slow when solving new challenges, as observed during the Detour task. From a different perspective, expectations towards some stimuli are perhaps reducing the individual’s motivation to interact with them, originating the passive behavior we understand as patience. The latter has been studied as an adaptive choice of renouncing immediate benefits to acquire more valuable future rewards. Maybe this behavior is not only related to the expectation of a better future reward but also, in some cases, with the expectation of no reward at all. Moreover, if we assume that individuals with generally lower activity levels may have been classified by the observers as more patient, this could partially explain why they react less promptly to an unfamiliar stimulus. An underlying link between personality traits related to interactions with unfamiliar environments, and the individuals’ expectations towards unknown situations might be present, as observed in pigs during a social isolation and novel object task (Asher et al., 2016). 

The fact that donkeys show different judgement bias, which are not correlated with their life history or living conditions, indicates that optimism/pessimism might not always be a consequence of acute stress or terrible welfare. Instead, it suggests that a trait-like feature might be conditioning these behaviors. It is important to define how these bias affect animals’ behavior and cognition, in order to improve our interaction with them and to broaden our notion of animal wellbeing.

Results from the U-task suggest that miniature donkeys have cognitive spatial abilities that allow them to detour around a symmetrical and an asymmetric obstacle. The expected strategy to solve the task, for D1 and D2 trials, would be to take the shorter route (i.e. walk around the shorter side of the barrier), yet this was not found to be the general response. Most subjects did not use reasoning and showed no spatial learning during this task, as previously observed in horses. Instead, this experiment becomes the second report of lateralized behavior in donkeys.

One individual revealed the ability to discover new and more expeditious ways to solve the test, as it switched from not having a preferred detour side during the symmetric trials and D1, to choosing the shorter detour during D2, when the asymmetry was highest. It would not be surprising that behavioral lateralization surmounts abstract thought in such a simple task, as it has been reported to be an adaptive strategy for coordination and facilitating action patterns. It is yet to be described how laterality affects donkey cognition, with studies in other species suggesting that it could be matched with impaired reasoning , and others proposing that enables behavioral function in the context of ecology . Although this study found no interactions between Laterality and other variables, it has also been suggested to potentially expose the valence of emotions.

The fact that more individuals preferred to detour around the left could be commented on the light of these findings, that indicate a higher involvement of the right hemisphere of the brain in the processing of response to novelty and flight behavior, food reward and positive social situations. As the test required a bucket to be moved in the presence of the animal, flight behavior was common during habituation and not completely eliminated during trials. More specialized tests are required to investigate how a nervous individual would process such a challenge when fear responses are absent, and when compared to a less nervous individual. Accordingly, when Baragli et al. (2011) suggests that non-lateralized horses might have higher spatial reasoning abilities, a question arises on whether these individuals displayed generally higher cognitive skills, or relatively lower fear-related responses, reducing the effects of lateralized processing of cognition. The same arguments apply to the positive correlation found between Concentration scores and the mean latency to approach the goal throughout U-task trials, since this trait is also closely related to the processing of daunting situations. The ability to concentrate in the tests required the animals to be relaxed, since potentially frightening stimuli were often present (e.g. noise, cars, unfamiliar human presence) and were the primary object of attention.

Similar studies would benefit from a bigger sample size and longer training/habituation periods, provided that the study subjects are under the same circumstances. Fear levels during the tests were not measured and fear responses to the study apparatus were not completely eliminated before the animals were tested, which could have affected the results. An important improvement to make would be to add controls for arousal, motivation, and distraction confounders. Given the kinship between the donkeys, a familial link could also be considered in future analysis. Regarding the personality questionnaire, some questions that proved to be highly correlated might have been redundant or prone to ambiguous interpretation by the observers. The same way, it was not perfectly clear what each gradient referred to in the scale of the described behavior for each question, and the personality factors discriminated do not always correspond to personality traits described in similar studies for other species. Hence, the application of validated behavioral tests with the study subjects would probably have been a more accurate approach to characterize their personalities. Another solution would be to validate a new questionnaire that allowed the observer to characterize donkey personality within the scope of well-studied personality traits that are common to all animal species – facilitating any results to be compared with the results of other studies. Finally, it is important to consider that each individual was only scored in one particular time of their lives, and that the observers did not spend more than a maximum of three hours per day with the animals, on a regular week.

In conclusion, the results reported here indicate that judgement bias might vary between individuals living under the same good conditions, suggesting once more that optimist/pessimism might have a component that is intrinsic to personality. Although not corresponding to the proposed hypothesis, it was shown that specific personality factors might predict the valence of judgement bias. No evidence was found to attribute causality between these and the individual differences in cognition, but unexpected links between personality factors and performance during the cognitive tests were observed. This study represents the second report of judgement bias experiments in donkeys, and the first referring to a possible link between these and personality in this species.

The emerging study of intrinsic individual differences in judgement bias - possibly linked to well-known personality traits, is of great importance to understand animal welfare, and to further develop animal behavior studies. Transposable to other species as they likely are, these findings could also yield knowledge about how much cognitive bias experienced by humans are inherent in personality, and independent of circumstances – contributing to the study of human psychology.