ABSTRACT
Play behaviour can act as an indicator of positive animal welfare. Previous attempts to predict play behaviour in farmed calves are limited because of the classification methods used, which lead to overestimation, and the short time periods that calves are observed. The study aimed to automatically classify and quantify play behaviour in farmed calves using location data from ultra-wide band sensors and to investigate factors associated with play behaviour. Location data were collected from 46 calves in three cohorts for a period of 18 weeks. Behavioural observations from video footage were merged with location data to obtain a total of 101.36 h of labelled data. An AdaBoost ensemble learning algorithm was implemented to classify play behaviour. To account for overestimation, generally seen in low-prevalence behaviours, an adjusted count technique was applied to the outputs of the classifier. Two generalized linear mixed models were fitted to investigate factors (e.g. age, health) associated with duration of play and number of play instances per day. Our algorithm identified play behaviour with > 94% accuracy when evaluated on the test set with no animals used for training, and 16% overestimation, which was computed based on the predicted number of samples of play versus the number of samples labelled as play on the test set. The instances and duration of play behaviour per day significantly decreased with age and sickness, whilst play behaviour significantly increased during and after weaning. The instances of play also significantly decreased as mean temperature increased. We suggest that the quantification method that we used could be used to detect and monitor other low prevalence behaviours (e.g. social grooming) from location data, including indicators of positive welfare.
Subject(s)
Animal Welfare , Health Status , Cattle , Animals , Weaning , Farms , Algorithms , Behavior, AnimalABSTRACT
In this study, we assessed for the first time the use of a reticuloruminal temperature bolus and a thresholding method to detect drinking events and investigated different factors that can affect drinking behavior. First, we validated the detection of drinking events using 16 cows that received a reticuloruminal bolus. For this, we collected continuous drinking behavior data for 4 d using video recordings and ambient and water temperature for the same 4 d. After all the data were synchronized, we performed 2 threshold algorithms: a general-fixed threshold and a cow-day specific threshold algorithm. In the general-fixed threshold, a positive test was considered if the temperature of any cow fell below a fixed threshold; in the cow-day specific threshold, a positive test was considered when the temperature of specific cows fell below the threshold value deviations around the mean temperature of the cow for that day. The former was evaluated using a threshold varying between 35.7 and 39.5°C, and the latter using the formula µ-n10σ, where µ = mean of the temperature of each cow for one day, n = 1, 2, , 20, and σ = standard deviation of the temperature of each cow on that day. The performance of the validation of detection using each of the threshold types was computed using different metrics, including overall accuracy, precision, recall (also known as sensitivity), F-score, positive predictive value, negative predictive value, false discovery rate, false omission rate, and Cohen's kappa statistic. The findings of the first study showed that the cow-day specific threshold of n = 10 performed better (true positives = 466; false positives = 167; false negatives = 165; true negatives = 8,416) than using a general-fixed threshold of 38.1°C (true positives = 449; false positives = 181; false negatives = 182; true negatives = 8,402). With the information gained in this first study, we investigated the different factors associated with temperature drop characteristics per cow: number of drops, mean amplitude of the drop, and mean recovery time. For this, we used data from 54 cows collected for almost 1 yr to build a mixed-effect multilevel model that included days in milk, parity, average monthly milk production, and ambient temperature as explanatory variables. Cow characteristics and ambient temperature had significant effects on drinking events. Our results provide a platform for automated monitoring of drinking behavior, which has potential value in prediction of health and welfare in dairy cattle.
Subject(s)
Algorithms , Body Temperature/physiology , Cattle/physiology , Drinking Behavior/physiology , Reticulum/metabolism , Rumen/metabolism , Animal Husbandry/methods , Animals , Female , Lactation/physiology , Likelihood Functions , Milk , Parity , Predictive Value of Tests , Pregnancy , ROC Curve , Video RecordingABSTRACT
Canine atopic dermatitis (cAD) is a common hereditary clinical syndrome in domestic dogs with no definitive diagnostic tests, which causes marked morbidity and has a high economic impact internationally. We created a novel questionnaire for Labrador (LR) and Golden retriever (GR) owners to evaluate canine skin health with respect to clinical signs of cAD. 4,111 dogs had fully completed questionnaires (2,803 LR; 1,308 GR). 'Cases' (793) had a reported veterinary diagnosis of cAD, and 'controls' (1652) had no current or past clinical signs of cAD and were aged >3 years. Remaining dogs (1666) were initially categorised as 'Other'. Simulated annealing was used comparing 'Cases' and 'Others' to select a novel set of features able to classify a known case. Two feature sets are proposed, one for use on first evaluation and one for dogs with a history of skin problems. A sum for each list when applied to the whole population (including controls) was able to classify 'Cases' with a sensitivity of 89% to 94% and specificity of 71% to 69%, respectively, and identify potentially undiagnosed cases. Our findings demonstrate for the first time that owner questionnaire data can be reliably used to aid in the diagnostic process of cAD.
Subject(s)
Dermatitis, Atopic/diagnosis , Dog Diseases/diagnosis , Surveys and Questionnaires/standards , Animals , Dermatitis, Atopic/immunology , Dermatitis, Atopic/veterinary , Dog Diseases/immunology , Dogs , Female , Male , Reference Standards , Reproducibility of Results , Risk Factors , Sensitivity and Specificity , Skin/immunologyABSTRACT
Time constraints for dairy farmers are an important factor contributing to the under-detection of lameness, resulting in delayed or missed treatment of lame cows within many commercial dairy herds. Hence, a need exists for flexible and affordable cow-based sensor systems capable of monitoring behaviors such as time spent feeding, which may be affected by the onset of lameness. In this study a novel neck-mounted mobile sensor system that combines local positioning and activity (acceleration) was tested and validated on a commercial UK dairy farm. Position and activity data were collected over 5 consecutive days for 19 high-yield dairy cows (10 lame, 9 nonlame) that formed a subset of a larger (120 cow) management group housed in a freestall barn. A decision tree algorithm that included sensor-recorded position and accelerometer data was developed to classify a cow as doing 1 of 3 categories of behavior: (1) feeding, (2) not feeding, and (3) out of pen for milking. For each classified behavior the mean number of bouts, the mean bout duration, and the mean total duration across all bouts was determined on a daily basis, and also separately for the time periods in between milking (morning = 0630-1300 h; afternoon = 1430-2100 h; night = 2230-0500 h). A comparative analysis of the classified cow behaviors was undertaken using a Welch t-test with Benjamini-Hochberg post-hoc correction under the null hypothesis of no differences in the number or duration of behavioral bouts between the 2 test groups of lame and nonlame cows. Analysis showed that mean total daily feeding duration was significantly lower for lame cows compared with non-lame cows. Behavior was also affected by time of day with significantly lower mean total duration of feeding and higher total duration of nonfeeding in the afternoons for lame cows compared with nonlame cows. The results demonstrate how sensors that measure both position and acceleration are capable of detecting differences in feeding behavior that may be associated with lameness. Such behavioral differences could be used in the development of predictive algorithms for the prompt detection of lameness as part of a commercially viable automated behavioral monitoring system.