A Review of Equine Sleep: Implications for Equine Welfare

Automated Observations of Dogs' Resting Behaviour Patterns Using Artificial Intelligence and Their Similarity to Behavioural Observations

Although direct behavioural observations are widely used, they are time-consuming, prone to error, require knowledge of the observed species, and depend on intra/inter-observer consistency. As a result, they pose challenges to the reliability and repeatability of studies. Automated video analysis is becoming popular for behavioural observations. Sleep is a biological metric that has the potential to become a reliable broad-spectrum metric that can indicate the quality of life and understanding sleep patterns can contribute to identifying and addressing potential welfare concerns, such as stress, discomfort, or health issues, thus promoting the overall welfare of animals; however, due to the laborious process of quantifying sleep patterns, it has been overlooked in animal welfare research. This study presents a system comparing convolutional neural networks (CNNs) with direct behavioural observation methods for the same data to detect and quantify dogs' sleeping patterns. A total of 13,688 videos were used to develop and train the model to quantify sleep duration and sleep fragmentation in dogs. To evaluate its similarity to the direct behavioural observations made by a single human observer, 6000 previously unseen frames were used. The system successfully classified 5430 frames, scoring a similarity rate of 89% when compared to the manually recorded observations. There was no significant difference in the percentage of time observed between the system and the human observer (p > 0.05). However, a significant difference was found in total sleep time recorded, where the automated system captured more hours than the observer (p < 0.05). This highlights the potential of using a CNN-based system to study animal welfare and behaviour research.

A field approach to observing changes in behavioural welfare indicators over 2 years in stabled horses

In stabled horses, several behaviours are considered to be important indicators of the state of welfare at the individual level: stereotypies, aggressive behaviours towards humans and the "withdrawn", alert, sternal, lateral and observation postures. Until now, these behaviours have been extensively studied in relation to different horse management practices. However, little is known about their changes or consistency over time. This study aimed to investigate differences in the expression of these behaviours assessed on 44 stabled horses during an initial 3-month period and then again on the same individuals 2 years later. Out of the six behaviours studied, two showed significant differences between the 2 years with medium effect sizes: the levels of aggressiveness towards humans increased (Wilcoxon signed-rank test: V = 65, P = 0.005) and those of recumbent rest during the day decreased (V = 416, P < 0.001) over time. The results also suggested limited evidence of major changes over time in the expression of stereotypies, the "withdrawn", alert and observation postures. However, the principal component analyses carried out on all the behaviours showed that alert and observation postures may slightly decrease over time for some individuals. Regarding stereotypies and the "withdrawn" posture, the results mainly suggested a change at the individual level in the expression of these behaviours over time. This study provided new insights into the dynamic nature of several behaviours when the horses' living conditions are not fundamentally altered and raised hypotheses about the state of welfare of stabled individuals over a 2-year period.

Sleep Pattern Interference in the Cognitive Performance of Lusitano Horses

Like most mammalian, polyphasic sleep, equine sleep can be divided into two phases: the REM (rapid eye movement) phase and the NREM (non-rapid eye movement) phase. For this study, a randomized crossover experiment was conducted using ten purebred Lusitano horses, all dressage athletes aged from three to seven years old. The horses were filmed before the intervention to characterize their sleep patterns. REM sleep deprivation was achieved by not letting the horses attain sternal or lateral recumbency for three consecutive days, totaling 72 h. A spatial memory task and a visual attention test were performed. A recording time of 48 h appeared to be long enough to characterize the sleep patterns of the stalled horses. The total recumbency time of the studied population was lower than that previously reported in horses. Although the recumbency times before and after the intervention were similar, there was a tendency shown by the delta (p = 0.0839) towards an increased time needed to resolve spatial memory tasks in the sleep-deprived group. Future studies may deepen the understanding of horse sleep requirements and patterns, and the effects of environmental changes on horse sleep.

Comparing lying behaviour of young riding horses on days in an individual indoor box, on an outdoor paddock alone, or in pairs and in the following night

BACKGROUND

Horses must lie down to go into vital rapid-eye movement (REM) sleep. If they are not lying down for sufficiently long periods they can become so sleep-deprived that they collapse uncontrollably, which results in a risk of injuries.

OBJECTIVES

To investigate how recumbency as a prerequisite to REM sleep on the experimental days and the following nights was influenced by changes in social and spatial environment throughout the day.

STUDY DESIGN

Cross-over design in which subjects experienced each experimental condition twice.

METHODS

Observations were conducted on a horse farm. Ten young horses in training were observed on days when they were alone indoors in a box, on an outdoor paddock alone, on the same paddock in pairs, and in the following night stabled alone. The number of lying bouts and the total lying duration throughout the day and night were assessed automatically using 3D-accelerometers and data were evaluated using mixed-models.

RESULTS

Horses had a higher number of lying bouts during the days (p = 0.05, by a factor of 1.21 [95% CI: 1.00-1.45]) and longer lying duration at night (p < 0.001, by a factor of 11.25 [6.47-18.40]). On average, the number of lying bouts and the duration of lying increased from being indoors alone, to being outdoors alone, and outdoors in pairs although this could not be statistically supported (bouts: p = 0.5, by a factor of 1.08 [0.84-1.36] and 1.17 [0.91-1.48]; p = 0.6, duration: by a factor of 1.39 [0.73-2.93] and 1.38 [0.68-2.78]).

MAIN LIMITATIONS

A small number of horses were observed and there was large variability between days within horses.

CONCLUSIONS

We found some indications that open space and a social companion throughout the day increased time lying down in the day as well as during the following nights thus allowing for more REM sleep.

Nocturnal behavioral patterns of African ungulates in zoos

Currently, most studies on ungulates' behavior are conducted during the daylight hours, but their nocturnal behavior patterns differ from those shown during day. Therefore, it is necessary to observe ungulates' behavior also overnight. Detailed analyses of nocturnal behavior have only been conducted for very prominent ungulates such as Giraffes (Giraffa camelopardalis), African Elephants (Loxodonta africana), or livestock (e.g., domesticated cattle, sheep, or pigs), and the nocturnal rhythms exhibited by many ungulates remain unknown. In the present study, the nocturnal rhythms of 192 individuals of 18 ungulate species from 20 European zoos are studied with respect to the behavioral positions standing, lying-head up, and lying-head down (the typical REM sleep position). Differences between individuals of different age were found, but no differences with respect to the sex were seen. Most species showed a significant increase in the proportion of lying during the night. In addition, the time between two events of "lying down" was studied in detail. A high degree of rhythmicity with respect to this quantity was found in all species. The proportion of lying in such a period was greater in Artiodactyla than in Perissodactyla, and greater in juveniles than in adults.

Towards an objective measurement of sleep quality in non-human animals: using the horse as a model species for the creation of sleep quality indices

Sleep disturbance is observed across species, resulting in neurocognitive dysfunction, poor impulse control and poor regulation of negative emotion. Understanding animal sleep disturbance is thus important to understand how environmental factors influence animal sleep and day-to-day welfare. Self-reporting tools for sleep disturbance commonly used in human research to determine sleep quality cannot be transferred to non-verbal animal species research. Human research has, however, successfully used frequency of awakenings to create an objective measurement of sleep quality. The aim of this study was to use a novel sleep-quality scoring system for a non-human mammalian species. Five separate sleep quality indices calculations were developed, using frequency of awakenings, total sleep time and total time spent in different sleep states. These indices were applied to a pre-existing data set of equine sleep behaviour taken from a study investigating the effects of environmental change (lighting and bedding) on the duration of time in different sleep states. Significant treatment effects for index scores both differed and aligned with the original sleep quantity results, thus sleep quality may be a useful alternative measurement of sleep disturbance that could be used to investigate impactful (emotional, cognitive) effects on the animal.

Racehorse welfare across a training season

Racehorse welfare is gaining increasing public attention, however scientific evidence in this area is lacking. In order to develop a better understanding of racehorse welfare, it must be measured and monitored. This is the first study to assess racehorse welfare using scientific objective methods across a training season. The aim of this study was threefold, firstly to investigate welfare measures which could be used in the first welfare assessment protocol for racehorses. Secondly, to understand the effect that a racing and training season had on individual racehorses and thirdly to identify risk factors for both good and poor welfare. Thirteen racehorse training yards were visited at the beginning and the peak of the racing season in England. Behavioral observations along with individual environmental and animal-based welfare measures were carried out on 353 horses in 13 training yards selected for variability. In our sample the horses were generally in good physical health: 94% of horses recorded as an ideal body condition score, no horses had signs of hoof neglect and 77.7% had no nasal discharge. The overall prevalence of external Mouth Corner Lesions was 12.9% and was significantly higher for Flat racing than Jump racing horses. The majority of horses (67.5%) showed positive horse human interactions. When stabled 54.1% horses had physical social contact and nasal discharge was not associated with increased physical contact. The training season significantly affected Human Reactivity Tests, Horse Grimace Scale scores and time spent resting and feeding. A total of 14.5% of horses displayed stereotypic behavior on at least two occasions. Horses with windows in their stables spent more time surveying their surroundings. Overall, in this population of racehorses, horses spent around a third of their daytime feeding (33.7%) followed by time spent standing resting (22.6%). The welfare assessment protocol used in this study is suitable for use in industry to collect welfare data on racehorses.

The Effect of Stabling Routines on Potential Behavioural Indicators of Affective State in Horses and Their Use in Assessing Quality of Life

Increasing interest in equine welfare has emphasised the need for objective and reliable behavioural indicators of horses’ affective state. However, research has yielded mixed results regarding behaviours suited for industry use largely because they are subject to anthropomorphic interpretation. Stabling is commonly used to manage domesticated horses despite research indicating that it can negatively impact horse welfare, but its effect on their affective state is yet to be quantified. Ten adult horses (11.8 ± 4.4 years) were observed either on a day- (DS) or night-stabling (NS) schedule over two consecutive 24 h periods. NS horses were kept confined for significantly longer (13.60 ± 0.04 h) than DS horses (7.73 ± 0.07; t7 = 5.70; p = 0.0004). Eight behaviours occurred significantly more often during NS than DS: forward ears (t7 = 3.32; p = 0.001), neutral ears (t7 = 3.47; p = 0.001), stepping forward (t7 = 2.62; p = 0.001), stepping laterally (t7 = 2.39; p = 0.001), sternal recumbency (t7 = 2.64; p = 0.001), yawning (t7 = 2.69; p = 0.001), non-nutritive chewing (t7 = 2.49; p = 0.001), and closing eyes (t7 = 2.71; p = 0.001). These behaviours may be candidates for indicators that can be used to determine the affective state in horses and subsequently be used to assess equine quality of life and to optimise individual horse welfare.