Activity Time Budgets—A Potential Tool to Monitor Equine Welfare?

Simple Summary Horses’ behavior is a good indicator of their welfare status. However, its complexity requires objective, quantifiable, and unambiguous evidence-based assessment criteria. As healthy, stress-free horses exhibit a highly repetitive daily routine, horses’ time budget (amount of time in a 24 h period spent on specific activities) can assist in equine welfare assessment. A systematic review of the literature yielded 12 papers that assessed equine time budgets for eating, resting and movement for a minimum of 24 continuous hours. A total of 144 horses (1–27 years old), 59 semi-feral and 85 domesticated horses, are included in this review. The reported 24 h time budgets for eating ranged from 10% to 66.6%, for resting from 8.1% to 66%, for lying from 2.7% to 27.3%, and for movement from 0.015% to 19.1%. The large variance in time budgets between studies can largely be attributed to differences in age and environmental conditions. Management interventions (free access to food, increased space, decreased population density) in domesticated horses yielded time budgets similar to semi-feral horses. The data support the importance of environmental conditions for horses’ well-being and the ability of time budgets to assist in monitoring horses’ welfare. Abstract Horses’ behavior can provide valuable insight into their subjective state and is thus a good indicator of welfare. However, its complexity requires objective, quantifiable, and unambiguous evidence-based assessment criteria. As healthy, stress-free horses exhibit a highly repetitive daily routine, temporal quantification of their behavioral activities (time budget analysis) can assist in equine welfare assessment. Therefore, the present systematic review aimed to provide an up-to-date analysis of equine time budget studies. A review of the literature yielded 12 papers that fulfilled the inclusion criteria: assessment of equine time budgets for eating, resting and movement for a minimum of 24 continuous hours. A total of 144 horses (1–27 years old), 59 semi-feral and 85 domesticated horses, are included in this review. The 24 h time budgets for foraging or eating (10–6.6%), resting (8.1–66%), lying (2.7–27.3%), and locomotion (0.015–19.1%) showed large variance between studies, which can largely be attributed to differences in age and environmental conditions. Management interventions in domesticated horses (ad libitum access to food, increased space, decreased population density) resulted in time budgets similar to their (semi-)feral conspecifics, emphasizing the importance of environmental conditions and the ability of time budgets to assist in monitoring horses’ welfare.

As welfare is a subjective individual experience, the animal welfare scientific community has moved progressively from input or resource-based measures, such as housing type or the amount of food provided to the animal, toward more output or animal-based assessments that not only evaluate the quality of the environment in which an animal is kept but also its physical and psychological condition and its ability to express the full repertoire of species-appropriate behaviors [6,9,11]. Therefore, behavior is increasingly used as an indicator of welfare because an animal's behavior can provide valuable insight into its subjective state [6,12,13]. The behavior of horses is defined by their niche as a prey species and highly social grassland dwellers with strong group fidelity who, in their natural habitat, are continual grazers with ultradian activity patterns roaming areas of land up to 78 km 2 [14][15][16][17]. Healthy, stress-free horses divide their time between activities that allow them to satisfy their basic requirements of food, movement, and rest, and exhibit a highly repetitive, individual, daily routine with almost identical time patterns of behavior from day to day [15,18,19]. Accordingly, the amount of time an animal engages in behavioral activities (time budget) is considered a very informative welfare indicator [6,9,15,[20][21][22][23][24]. To this end, domesticated horses are often compared to feral or wild conspecifics. Thereby, feral horses can be used as a benchmark for comparison, not as gold standard for optimum welfare [6,9,15,[20][21][22][23][24]. Indeed, domestic horses, given the opportunity to display species-appropriate behavior and an environment sufficiently reflecting a natural habitat, display time budgets similar to those of wild horses [4,15,23,25]. However, domestic horses are kept in a variety of housing systems that offer more or less adequate environmental conditions with differing levels of physical freedom, often with regimental feeding and limited foraging and social opportunities [14,15]. Thus, differences in the time budgets of domesticated horses compared to feral or wild conspecifics are currently used as a measure of compromised welfare [6,15,23,24,26,27].
As equine behavioral activity displays 24 h and circadian variation, measuring time budgets requires detailed surveillance over several days [9,17,28,29]. Traditionally, this has been done by direct observation which was prone to observer bias, often limited to daylight hours and too time and resource-intensive to be feasible for welfare assessment. However, recent advances in biotelemetry and artificial intelligence (AI) based on sensor (mostly accelerometer) or video analysis provide increased objectivity and enable remote recording of behavioral data, longer observation periods, and quantification of behaviors with higher accuracy and temporal resolution than the human eye allows [30][31][32]. Hence, time budget analysis, as an objective, quantitative measure of behavior, now has the potential to become a useful, reliable tool for on-farm assessment of equine welfare and comparison of welfare under different environmental conditions. Therefore, the present review aimed to provide an up-to-date analysis of equine time budget studies and synthesize evidence relating to the effect of housing and management systems on horses' behavioral time budgets.

Materials and Methods
This review was carried out in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [33].

Data Sources and Searches
Scientific peer-reviewed articles were identified through a systematic search in the PubMed (National Institutes of Health. PubMed (Database). Bethesda, MD, USA: National Library of Medicine; https://pubmed.ncbi.nlm.nih.gov, accessed on 25 February 2021) and Scopus (Elsevier, Amsterdam, The Netherlands; https://www.scopus.com, accessed on 25 February 2021) electronic databases searching for the terms "(("time budget" OR "activity budget" OR "activity tracking") OR ((sleep OR sleeping OR resting OR lying OR eating OR foraging) AND (behaviour OR behavior OR time OR video OR sensor OR gyroscope OR accelerometer))) AND (horse OR equine OR pony OR horses OR ponies OR "equus caballus")" in title or abstract, with no restriction on publication date, in November 2020. The study selection process was carried out by the first and last author following the procedure detailed in Figure 1, excluding papers that upon closer inspection, did not study the time horses spend on specific activities or did not include a minimum of 24 h uninterrupted observations. Any disagreement between the authors on the studies to be included in the review was resolved during a consensus meeting. All quantitative or qualitative randomized controlled trials, observational studies, and case series focused on equine time budgets or activity quantification written in English or German were included. The following exclusion criteria were set: (a) non-peer-reviewed publication, (b) conference/seminar abstract only published, (c) dissertation, thesis, review, commentary, or single case report, (d) the article was not written in English or German.

Data Extraction and Risk of Bias Assessment
Information on the population, intervention, comparison, outcome, and study design (PICOS) was retrieved from the articles, and the risk of bias of selected studies was assessed using a modification of the Evidence Project risk-of-bias tool [34][35][36].

Author(s) and
Przewalski horses kept under more confined conditions (44 ha semi-reserve) dedicated 29.8-46.4% of their time budget to eating or foraging, 36.4-36.6% to resting, 5.3% to lying, 7.4% to locomotion, and 10.2% to other behaviors such as drinking, self-grooming, and playing (Table 3) [28,39]. Behavioral analysis by automated tracking revealed a complex diurnal and ultradian rhythmicity and seasonal variations of activity patterns [28].
Four of the seven studies on domesticated horses [41][42][43]45] studied the effect of different feeding regimes on equine time budgets, and one study evaluated the effect of postoperative pain [40], stocking density [46] and turn-out management together with paddock size [44]. The time budget of domesticated horses was divided between 10-64% eating or foraging, 15.6-68% resting, 3-27.3% lying, 0.015-19.3% (in horses not confined to a stable: 2.5-19.3%) locomotion and 2-11.5% other behaviors such as drinking and self-grooming (Table 3) [40][41][42][43][44][45][46]. The large variance in activity time budgets can in part be explained by the age range of the horses included in the various studies, as young horses are resting more in a recumbent position than adults [43,46].
The feeding regime significantly affected the time budget of horses [41,45]. Regular feeding times and coordinated rations led to a significant reduction in the time spent on feeding. Offering the hay ration in hay bags increased the time spent feeding but to a lesser extent than ad libitum feeding [41,45]. The diurnal and ultradian rhythmicity of feeding of wild and feral horses could also be observed in domestic horses with more time spent feeding during daytime [42,44]. Paddock size and stocking density had no influence on feeding time but affected time for resting and locomotion with a decreased stocking density showing a positive correlation with locomotion, playing, and self-grooming [46] and a smaller paddock size being associated with decreased locomotion time [44].

Discussion
Aiming to provide an up-to-date analysis of equine time budget studies, this review included all studies that determined equine time budgets for eating, resting and locomotion over an observation period of at least 24 h [18,28,[37][38][39][40][41][42][43][44][45][46]. The reported time budgets of (semi-)feral and domesticated horses show large variances (Table 3). In (semi-)feral horses, the variation in time budgets was primarily attributed to age [18], sex [37] and seasonal influences with correspondingly variable food availability, temperatures, and insect pest densities [28,38]. At this point, it is important to take into consideration that the papers included in this review studied domesticated and semi-feral but not wild horses and hence may not entirely reflect the natural behavior of their wild ancestors.

Time Budget Measurement: Methods and Use
Unfortunately, even though Berger et al. already used a telemetric system in 1999 to monitor a group of horses over one year [28], many studies included in this review have not made optimal use of available technology (Table 1). Instead, they used direct observation or video recordings with scan or focal sampling which does not provide a continuous assessment of behavior and thus potentially misses or underrepresents important, infrequent behaviors. Using automated tracking methods instead of visual observation reduces observer bias, improves data resolution, and allows long-term tracking of larger group sizes. This may help to reveal behavioral patterns over more extended periods and determine the influence of environmental conditions on time budgets. Currently, most biotelemetry systems are not yet able to differentiate reliably between different gaits (walk, trot, canter) and between walking and static movements (e.g., stamping, twitching to ward off pests) and thus may determine erroneous gait patterns or too high movement values [32,47]. However, automated tracking methods have been validated in other species (seals, goats, pigs, birds) [30,31,[48][49][50] and show great promise for application in equine studies [28,44].
The time budgets of several welfare relevant behaviors, such as foraging, resting, and lying, can already be accurately determined with commercially available sensors and can be used as welfare indicators to identify welfare problems and monitor the success of interventions [32,44]. Furthermore, real-time analysis of equine behavior may also facilitate early detection of health problems, such as colic, lameness or other painful conditions and accelerate therapeutic interventions [4,40,[51][52][53][54][55]. Indeed, time budgets for specific behavior have been identified as more sensitive indicators of equine discomfort than repeated direct observation of specific events and postures and thus could facilitate rapid detection of painful conditions and objective, quantitative monitoring of the success of therapeutic interventions [40,54,55].

Time Budget for Feeding and Foraging
Horses' digestive physiology and anatomy have adapted to their natural diet that is rich in fiber and low in starch and energy [27,54,56]. Free-ranging horses devote the majority of their time to the search for and consumption of food, spending up to 18 h a day foraging, and rarely fast voluntarily for more than 2 to 4 h at a time [27,57]. Thus, the time budget for foraging in grazing (semi-)feral horses ranges from 50.82% to 66.6%, with circadian and seasonal variations depending on climatic conditions and food quality but also differences between daylight and darkness (Table 3) [18,28,37,38,58]. However, many domestic horses have limited access to roughage and are fed restricted amounts of hay and a commercial feed with higher caloric density. Thus, the determined time budget for foraging in the adult domestic horse varies greatly from 16% with rationed feeding [45] to 64% with ad libitum access to hay (Table 3) [41]. One of the papers included in this systematic review also found postoperative pain to reduce feeding time [40] but did not separate the effects of pain from those of general anesthesia. In contrast, another recent study investigating the influence of an iatrogenically-induced acute septic osteoarthritis on behavior found no effect of pain on feeding behavior [50].

Time Budget for Resting and Sleeping
Adult feral horses spend 12.9-29.3% of their day resting standing and 4.2-15.5% lying in a polyphasic pattern in multiple shorter periods (Table 3) [37,38,72]. However, the time budget for resting, which includes periods of inactivity and sleep, cannot be compared beyond doubt between studies, as there is disagreement whether vigilant standing is part of the resting behavior or activity behavior or forms an independent behavior category. Resting behavior can occur while standing as well as lying down in sternal or lateral recumbency and is age-dependent. Adult horses spend 80% of their resting time standing and only a relatively small proportion of their 24 h time budget recumbent, while foals under three months of age lie down for 70-80% of their resting time [18,37,38,52,73,74]. Young domestic horses are lying down approximately 25% of the day [43,46], which is significantly more than adults that are recumbent for approximately 5% of the day (Table 3) [40,42]. Domesticated horses, analogous to wild horses, lie down mostly between midnight and four in the morning [39,46].
Measuring lying behavior is an essential component of equine welfare assessment because horses usually fall asleep shortly after lying down; thus, recumbency is a reliable indicator of sleep. Four stages of horses' sleep-wake rhythm are typically differentiated and defined by specific cortical electronic activity and movement patterns: wakefulness, drowsiness, slow-wave-sleep, and paradoxical or rapid eye movement (REM) sleep, with the majority of sleep occurring between midnight and 5:00 am [39,44,46,60,72,75,76]. While sleep in horses is not uniquely associated with recumbency as horses can sleep standing, recumbency (sternal or lateral) is required for rapid eye movement (REM) sleep, which is vital for many physiological and cognitive functions [53,72]. Indeed, in other species including humans, REM-sleep deprivation has been linked to hyperalgesia and persistent chronic pain [53,77,78].
In horses, decreased lying time budgets were associated with inappropriate environmental conditions, stress, and painful musculoskeletal issues [4,40,50,52,53]. In contrast, increased lying times were observed in animals with higher social rank [79], larger stall size [80], straw bedding compared to shavings [81], and following administration of analgesics to horses suffering from orthopedic pain [4,52,53].

Time Budget for Locomotion and Movement
The time budget dedicated to movement or activity ranges from 4.3% to 13.4% in feral horses, with walk as the predominant gait and less movement at night than during the day (Table 3) [37][38][39]. In stallions, a greater need for rapid movement (trot, canter) has been reported [37]. The time budget for movement also includes, in addition to locomotion in the different gaits, play behavior which is most commonly observed in foals, and stereotypical movements such as weaving, crib-biting, and stall-walking. Horses with insufficient foraging opportunities and horses living in too densely stocked conditions show increased active locomotion patterns (Table 3) [4,26,27,42,45], confirming the value of locomotor activity as reliable indicators for horse welfare [82]. As horses also move while foraging, albeit slowly and less linearly, a clear distinction between foraging and movement behavior is required, but unfortunately this was not evident in all studies; thus, the time budgets for movement have to be interpreted with caution [9,15].

Conclusions
Activity time budgets allow an objective, quantitative on-farm welfare assessment and comparison of different management, feeding, and housing systems. In addition, changes in time budgets can be used to identify painful conditions and monitor the success of management interventions to improve equine welfare. However, the diversity of the horse populations, the environment the horses were living in, and the measurement methods leads to a great heterogeneity of the studies and is reflected in the highly variable time budgets. Thus, further studies of larger horse groups that live in clearly defined housing and management conditions, using modern observational technologies, such as biotelemetry or AI-based video analysis, are needed to further validate and establish the use of time budgets as a reliable indicator of equine welfare.