Agreement between different sleep states and behaviour indicators in dairy cows

Methods used for estimating sleep in dairy cattle

Sleep serves several essential functions in all mammals including dairy cattle. Researchers are beginning to estimate sleep in dairy cattle using a combination of physiological measurements (e.g., polysomnography) as well as changes in behavior (e.g., different resting postures). Sleep may provide unique insight into how cows and calves respond to, and cope with, their environments, as a complement to other common measurements such as lying time. Although each of the methods to assess sleep in cattle has its advantages, there remain several challenges with each approach. The objective of this narrative mini-review is to describe current methods for estimating sleep in dairy cattle, including some of the advantages and limitations with each method. We will start with describing the research to date on adult cows, followed by preweaning dairy calves. We end the review with recommendations for researchers interested in assessing sleep in dairy cattle and ideas for future areas of research.

Shelter preference and behavior of dairy cows managed outdoors during calving in temperate winter conditions

This study aimed to assess the preference and use of an artificial shelter in dairy cows managed outdoors at calving during winter. We also evaluated whether this preference would be influenced by weather conditions (rain, air temperature, or wind) or the time of the day at calving. Two weeks before their expected calving date, 18 cows were paired based on body weight, body condition score, parity, and expected calving date. Each pair was assigned to an open paddock (42 m2/cow) with a bare soil surface, high mud content, and access to an artificial shelter until calving. The shelter consisted of a 6 × 3 m metal structure, with 3 sides covered with zinc sheets, a polycarbonate roof, and a thick layer of dry sawdust covering the ground surface. Shelter use and cow behavior inside it (lying and sleeping) were continuously recorded via video and analyzed using continuous observation. Observations were divided into 2 periods: the day before calving (24 to 48 h before calving) and the day of calving (24 h before calving until calving). Ninety-four percent (15/16) of the cows preferred to calve inside the artificial shelter, and this preference was not affected by the time of day or weather conditions. Regardless of the day of study (the day before calving or the day of calving), cows spent approximately 64% of their daily time inside the shelter, and once inside, they spent most of the time lying down. Our findings indicate dairy cows prefer using an artificial shelter as a calving and lying place, suggesting that opportunities for protection should be provided when they are managed outdoors in muddy paddock conditions.

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.

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.

Testing the Accuracy of Wearable Technology to Assess Sleep Behaviour in Domestic Dogs: A Prospective Tool for Animal Welfare Assessment in Kennels

Sleep is a physiological process that has been shown to impact both physical and psychological heath of individuals when compromised; hence, it has the potential to be used as an indicator of animal welfare. Nonetheless, evaluating sleep in non-human species normally involves manipulation of the subjects (i.e., placement of electrodes on the cranium), and most studies are conducted in a laboratory setting, which limits the generalisability of information obtained, and the species investigated. In this study, we evaluated an alternative method of assessing sleep behaviour in domestic dogs, using a wearable sensor, and compared the measurements obtained to behavioural observations to evaluate accuracy. Differences between methods ranged from 0.13% to 59.3% for diurnal observations and 0.1% to 95.9% for nocturnal observations for point-by-point observations. Comparisons between methods showed significant differences in certain behaviours, such as inactivity and activity for diurnal recordings. However, total activity and total sleep recorded did not differ statistically between methods. Overall, the wearable technology tested was found to be a useful, and a less-time consuming, tool in comparison to direct behavioural observations for the evaluation of behaviours and their indication of wellbeing in dogs. The agreement between the wearable technology and directly observed data ranged from 75% to 99% for recorded behaviours, and these results are similar to previous findings in the literature.

The Effects of Mild Disturbances on Sleep Behaviour in Laying Hens

The positive welfare of commercial animals presents many benefits, making the accurate assessment of welfare important. Assessments frequently use behaviour to determine welfare state; however, nighttime behaviours are often ignored. Sleep behaviour may offer new insights into welfare assessments. This study aimed to establish a baseline for sleep behaviour in laying hens and to then apply mild short-term disturbances and observe the subsequent effects. Twelve laying hens were divided into four batches and were surgically implanted with electroencephalogram (EEG) devices to record their brain activity. The batches were subjected to undisturbed, disturbed and recovery types of nights. Disturbed nights consisted of systematic sequences of disturbance application (wind, 90 dB noise or 20 lux light) applied one at a time for 5 min every 30 min from 21:00 to 03:00 (lights off period: 19:00-05:00). Sleep state was scored using EEG data and behaviour data from infrared cameras. Over all the types of night hens engaged in both SWS (58%) and REM sleep (18%) during lights off. When applied, the disturbances were effective at altering the amounts of wakefulness and SWS (Time × Type of Night, p < 0.001, p = 0.017, respectively), whereas REM sleep was unaltered (p = 0.540). There was no evidence of carry-over effects over the following day or night. Laying hens may be resilient to short-term sleep disruption by compensating for this in the same night, suggesting that these disturbances do not impact their long-term welfare (i.e., over days). Sleep behaviour potentially offers a unique means of assessing an aspect of animal welfare that, to date, has been poorly studied.

The development of sleep-like posture expression with age in female Holstein calves

Sleep plays an important role in cattle production and welfare. Therefore, this study aimed to investigate the development of sleep-like posture (SLP) expression in dairy calves from birth to the first calving as an indicator of their sleep behavior. Fifteen female Holstein calves were subjected. Daily SLP was measured eight times (0.5, 1, 2, 4, 8, 12, and 18 months and 23 months or 1 month before the first calving) using an accelerometer. Calves were kept in individual pens until weaning at 2.5 months old and then kept in the group. The daily SLP time decreased rapidly in early life; however, the rate of decrease gradually slowed and eventually reached a stable level of approximately 60 min/day after 12 months old. The daily SLP bout frequency showed the same change as the SLP time. In contrast, the averaged SLP bout duration slowly decreased with age. Longer daily SLP time during early life in female Holstein calves might be related to brain development. The individual expression of daily SLP time differs before and after weaning. Some external and/or internal factors associated with weaning may be involved SLP expression.

Sleep pattern in the dromedary camel: a behavioral and polysomnography study

Study Objectives

To investigate sleep patterns in the camel by combining behavioral and polysomnography (PSG) methods.

Methods

A noninvasive PSG study was conducted over four nights on four animals. Additionally, video recordings were used to monitor the sleep behaviors associated with different vigilance states.

Results

During the night, short periods of sporadic sleep-like behavior corresponding to a specific posture, sternal recumbency (SR) with the head lying down on the ground, were observed. The PSG results showed rapid shifts between five vigilance states, including wakefulness, drowsiness, rapid eye movement (REM) sleep, non-REM (NREM) sleep, and rumination. The camels typically slept only 1.7 hours per night, subdivided into 0.5 hours of REM sleep and 1.2 hours of NREM sleep. Camels spent most of the night being awake (2.3 hours), ruminating (2.4 hours), or drowsing (1.9 hours). Various combinations of transitions between the different vigilance states were observed, with a notable transition into REM sleep directly from drowsiness (9%) or wakefulness (4%). Behavioral postures were found to correlate with PSG vigilance states, thereby allowing a reliable prediction of the sleep stage based on SR and the head position (erected, motionless, or lying down on the ground). Notably, 100% of REM sleep occurred during the Head Lying Down-SR posture.

Conclusions

The camel is a diurnal species with a polyphasic sleep pattern at night. The best correlation between PSG and ethogram data indicates that sleep duration can be predicted by the behavioral method, provided that drowsiness is considered a part of sleep.

BOVIDS: A deep learning-based software package for pose estimation to evaluate nightly behavior and its application to common elands (Tragelaphus oryx) in zoos

Only a few studies on the nocturnal behavior of African ungulates exist so far, with mostly small sample sizes. For a comprehensive understanding of nocturnal behavior, the data basis needs to be expanded. Results obtained by observing zoo animals can provide clues for the study of wild animals and furthermore contribute to a better understanding of animal welfare and better husbandry conditions in zoos. The current contribution reduces the lack of data in two ways. First, we present a stand-alone open-source software package based on deep learning techniques, named Behavioral Observations by Videos and Images using Deep-Learning Software (BOVIDS). It can be used to identify ungulates in their enclosure and to determine the three behavioral poses "Standing," "Lying-head up," and "Lying-head down" on 11,411 h of video material with an accuracy of 99.4%. Second, BOVIDS is used to conduct a case study on 25 common elands (Tragelaphus oryx) out of 5 EAZA zoos with a total of 822 nights, yielding the first detailed description of the nightly behavior of common elands. Our results indicate that age and sex are influencing factors on the nocturnal activity budget, the length of behavioral phases as well as the number of phases per behavioral state during the night while the keeping zoo has no significant influence. It is found that males spend more time in REM sleep posture than females while young animals spend more time in this position than adult ones. Finally, the results suggest a rhythm between the Standing and Lying phases among common elands that opens future research directions.

Deep learning-based pose estimation for African ungulates in zoos

The description and analysis of animal behavior over long periods of time is one of the most important challenges in ecology. However, most of these studies are limited due to the time and cost required by human observers. The collection of data via video recordings allows observation periods to be extended. However, their evaluation by human observers is very time-consuming. Progress in automated evaluation, using suitable deep learning methods, seems to be a forward-looking approach to analyze even large amounts of video data in an adequate time frame.In this study, we present a multistep convolutional neural network system for detecting three typical stances of African ungulates in zoo enclosures which works with high accuracy. An important aspect of our approach is the introduction of model averaging and postprocessing rules to make the system robust to outliers.Our trained system achieves an in-domain classification accuracy of >0.92, which is improved to >0.96 by a postprocessing step. In addition, the whole system performs even well in an out-of-domain classification task with two unknown types, achieving an average accuracy of 0.93. We provide our system at https://github.com/Klimroth/Video-Action-Classifier-for-African-Ungulates-in-Zoos/tree/main/mrcnn_based so that interested users can train their own models to classify images and conduct behavioral studies of wildlife.The use of a multistep convolutional neural network for fast and accurate classification of wildlife behavior facilitates the evaluation of large amounts of image data in ecological studies and reduces the effort of manual analysis of images to a high degree. Our system also shows that postprocessing rules are a suitable way to make species-specific adjustments and substantially increase the accuracy of the description of single behavioral phases (number, duration). The results in the out-of-domain classification strongly suggest that our system is robust and achieves a high degree of accuracy even for new species, so that other settings (e.g., field studies) can be considered.

Machine learning prediction of sleep stages in dairy cows from heart rate and muscle activity measures

Sleep is important for cow health and shows promise as a tool for assessing welfare, but methods to accurately distinguish between important sleep stages are difficult and impractical to use with cattle in typical farm environments. The objective of this study was to determine if data from more easily applied non-invasive devices assessing neck muscle activity and heart rate (HR) alone could be used to differentiate between sleep stages. We developed, trained, and compared two machine learning models using neural networks and random forest algorithms to predict sleep stages from 15 variables (features) of the muscle activity and HR data collected from 12 cows in two environments. Using k-fold cross validation we compared the success of the models to the gold standard, Polysomnography (PSG). Overall, both models learned from the data and were able to accurately predict sleep stages from HR and muscle activity alone with classification accuracy in the range of similar human models. Further research is required to validate the models with a larger sample size, but the proposed methodology appears to give an accurate representation of sleep stages in cattle and could consequentially enable future sleep research into conditions affecting cow sleep and welfare.

Effects of acute lying and sleep deprivation on metabolic and inflammatory responses of lactating dairy cows

Dairy cows that are restricted from lying down have a reduced ability to sleep. In other species, sleep loss is a key risk factor for disease, mediated by changes in metabolic and inflammatory responses. The cumulative effect of lying and sleep deprivation on cow health is unknown. The objective was to determine the effects of lying and sleep deprivation on metabolic and inflammatory responses of dairy cows. Data were collected from 8 multiparous and 4 primiparous lactating cows (199 ± 44 d in milk, 77 ± 30 d pregnant; mean ± standard deviation) enrolled in a study using a crossover design. Each cow was exposed to 2 treatments meant to induce sleep loss: (1) human disturbance (imposed by researchers making noise or physical contact when the cow's posture suggested sleep) and (2) lying deprivation (imposed by a wooden grid placed on the pen floor). Cows experienced a 24-h baseline period (d -1) followed by a 24-h treatment period (d 0), with a 12-d washout period between treatments. Baseline and treatment periods were imposed from 2100 to 2059 h. Cows were housed in individual pens during the acclimation period (d -3 and -2), d -1, and d 0. Nonesterified fatty acid and glucose concentrations were measured at 0300, 0900, 1500, and 2059 h on d -1 and 0. Proinflammatory cytokine mRNA [tumor necrosis factor (TNF), interleukin-1B (IL1B), and interleukin-6 (IL6)] abundance in whole-blood leukocytes, both nonstimulated and stimulated with lipopolysaccharide, were assessed at 2059 h on d -1 (end of baseline) and d 0 (end of treatment). Nonesterified fatty acids and glucose varied by time of day but were not affected by treatment or day. The abundances of TNF and IL1B from both stimulated and nonstimulated cells were higher following 24 h of lying deprivation (d 0) compared with baseline (d -1). Abundance of IL6 was increased in nonstimulated cells after lying deprivation compared with baseline. In contrast, human disturbance for 24 h did not alter TNF, IL1B, or IL6 abundance relative to baseline levels. These results suggest that a short period of lying deprivation generally increases inflammatory responses but not metabolic responses.

Invited review: Lying time and the welfare of dairy cows

Adequate time lying down is often considered an important aspect of dairy cow welfare. We examine what is known about cows' motivation to lie down and the consequences for health and other indicators of biological function when this behavior is thwarted. We review the environmental and animal-based factors that affect lying time in the context of animal welfare. Our objective is to review the research into the time that dairy cows spend lying down and to critically examine the evidence for the link with animal welfare. Cows can be highly motivated to lie down. They show rebound lying behavior after periods of forced standing and will sacrifice other activities, such as feeding, to lie down for an adequate amount of time. They will work, by pushing levers or weighted gates, to lie down and show possible indicators of frustration when lying behavior is thwarted. Some evidence suggests that risk of lameness is increased in environments that provide unfavorable conditions for cows to lie down and where cows are forced to stand. Lameness itself can result in longer lying times, whereas mastitis reduces it. Cow-based factors such as reproductive status, age, and milk production influence lying time, but the welfare implications of these differences are unknown. Lower lying times are reported in pasture-based systems, dry lots, and bedded packs (9 h/d) compared with tiestalls and freestalls (10 to 12 h/d) in cross-farm research. Unfavorable conditions, including too few lying stalls for the number of cows, hard or wet lying surfaces, inadequate bedding, stalls that are too small or poorly designed, heat, and rain all reduce lying time. Time constraints, such as feeding or milking, can influence lying time. However, more information is needed about the implications of mediating factors such as the effect of the standing surface (concrete, pasture, or other surfaces) and cow behavior while standing (e.g., being restrained, walking, grazing) to understand the effect of low lying times on animal welfare. Many factors contribute to the difficulty of finding a valid threshold for daily lying time to use in the assessment of animal welfare. Although higher lying times often correspond with cow comfort, and lower lying times are seen in unfavorable conditions, exceptions occur, namely when cows lie down for longer because of disease or when they spend more time standing because of estrus or parturition, or to engage in other behaviors. In conclusion, lying behavior is important to dairy cattle, but caution and a full understanding of the context and the character of the animals in question is needed before drawing firm conclusions about animal welfare from measures of lying time.

The influence of concentrate-rich diet on daily sleep and lying postures of Japanese black cows

Sleep is one of the essential behaviors in mammalian health and welfare. In this study, we focused on how concentrate-rich diets (Conc) can influence the sleep of cattle. We hypothesized that Conc would suppress the daily sleep of cows because of behavioral frustration according to shorter eating and rumination. We subjected six Japanese black cows according to crossover experimental design. There were two diet treatments. In the Conc group, 70% of required TDN was supplied by concentrate and 30% by grass silage. On the other hand, in roughage fed group (Rough), all required TDN was supplied by grass silage. Maintenance behaviors, postures of cows, and sleep posture were measured by electronic devices. Cows fed with Conc spent less time in eating and rumination and spent more time in rest than Rough. However, contrary to our hypothesis, cows fed with Conc showed longer daily sleep posture duration (96.2 min/day vs. 69.1 min/day) and more frequent sleep posture bouts (15.4 times/day vs. 11.1 times/day) than Rough. We discussed about following possible explanations why cows fed with Conc slept longer as satisfaction, illness, displacement behavior, and shifting sleep posture.

Nightly selection of resting sites and group behavior reveal antipredator strategies in giraffe

This study presents the first findings on nocturnal behavior patterns of wild Angolan giraffe. We characterized their nocturnal behavior and analyzed the influence of ecological factors such as group size, season, and habitat use. Giraffe were observed using night vision systems and thermal imaging cameras on Okapuka Ranch, Namibia. A total of 77 giraffe were observed during 24 nights over two distinct periods-July-August 2016 (dry season) and February-March 2017 (wet season). Photoperiod had a marked influence on their activity and moving behavior. At dusk, giraffe reduced the time spent moving and increasingly lay down and slept at the onset of darkness. Body postures that likely correspond to rapid eye movement (REM) sleep posture (RSP) were observed 15.8 ± 18.3 min after giraffe sat down. Season had a significant effect with longer RSP phases during the dry season (dry: 155.2 ± 191.1 s, n = 79; wet: 85.8 ± 94.9 s, n = 73). Further analyses of the influence of social behavior patterns did not show an effect of group size on RSP lengths. When a group of giraffe spent time at a specific resting site, several individuals were alert (vigilant) while other group members sat down or took up RSP. Simultaneous RSP events within a group were rarely observed. Resting sites were characterized by single trees or sparse bushes on open areas allowing for good visibility in a relatively sheltered location.

Effects of acute lying and sleep deprivation on the behavior of lactating dairy cows

The objective was to determine the effects of sleep or lying deprivation on the behavior of dairy cows. Data were collected from 8 multi- and 4 primiparous cows (DIM = 199 ± 44 (mean ± SD); days pregnant = 77 ± 30). Using a crossover design, each cow experienced: 1) sleep deprivation implemented by noise or physical contact when their posture suggested sleep, and 2) lying deprivation imposed by a grid placed on the pen floor. One day before treatment (baseline), and treatment day (treatment) were followed by a 12-d washout period (with the first 7 d used to evaluate recovery). Study days were organized from 2100 to 2059. During habituation (d -3 and -2 before treatment), baseline (d -1), and trt (d 0), housing was individual boxstalls (mattress with no bedding). After treatment, cows returned to sand-bedded freestalls for a 7-d recovery period (d 1 to 7) where data on lying behaviors were collected. Following the recovery period, an additional 5-d period was provided to allow the cows a 12-d period between exposures to treatments. Daily lying time, number lying bouts, bout duration, and number of steps were recorded by dataloggers attached to the hind leg of cows throughout the study period. Data were analyzed using a mixed model including fixed effects of treatment (sleep deprivation vs. sleep and lying deprivation), day, and their interaction with significant main effects separated using a PDIFF statement (P ≤ 0.05). Interactions between treatment and day were detected for daily lying time and the number of bouts. Lying time was lower for both treatments during the treatment period compared to baseline. Lying time increased during the recovery period for both lying and sleep deprived cows. However, it took 4 d for the lying deprived cows to fully recover their lying time after treatment, whereas it took the sleep deprived cows 2 d for their lying time to return to baseline levels. Results suggest that both sleep and lying deprivation can have impact cow behavior. Management factors that limit freestall access likely reduce lying time and sleep, causing negative welfare implications for dairy cows.

The use of accelerometer to measure sleeping posture of beef cows

Sleep is one of the essential behaviors for mammals. The aims of this study were to validate the use of accelerometer for measuring sleeping posture of cattle. Duration of sleeping posture of seven Japanese Black cows from 19.00 to 07.00 hours was measured by both accelerometer and video, and a total of 67 accelerometer and video measurement sets were collected. We calculated Cohen's κ coefficient between accelerometer and video measurements and 91.5% of the κ-values were >0.80. Intra- and inter-observer coefficient of variance showed that specific acceleration waveform patterns of sleeping posture could be easily and accurately detected by independent observers. There were no significant differences in the frequency of sleeping posture occurrences between accelerometer and video measurements. We compared averaged sleeping posture bout, and the total sleeping posture time between accelerometer and video measurements using regression. In each trait, the slope was close to 1 and the intercept was not different from 0, which showed a strong agreement between accelerometer and video measurements. This shows that an accelerometer could accurately detect sleeping postures of cattle. We conclude that adequate measurements of sleeping postures can be made using an accelerometer.