Applied Animal Behaviour Science最新文献

{"title":"Weaned piglet’s reactivity to humans, tonic immobility and behaviour in a spatial maze test is affected by gestating sows’ relationship to humans and positive handling at weaning","authors":"Mathilde Lanthony ,&nbsp;Emmanuelle Briard ,&nbsp;Marek Špinka ,&nbsp;Céline Tallet","doi":"10.1016/j.applanim.2023.106080","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106080","url":null,"abstract":"<div><p>Fear of human can impact negatively offspring’s behaviour and cognition whereas early positive handling has positive impacts. The aim of this study was to evaluate the crossed effects of trans-generational transfer and familiarisation with human on piglets’ behaviour. Sows were classified into two categories (Fearful, n = 11 and Docile, n = 13) based on whether they avoided human contact or not during a human approach test at 71 days of gestation. Fear of human was not related to salivary cortisol levels of the sows (F =2.72, P = 0.10). Fearful sows’ piglets stood immobilized longer (Chisq= 6.24, P = 0.01), as well females (8.70, P = 0.003) during a tonic immobility test (TIT) at 7 days but not at 15 days. The behaviour of two male and two female piglets per sow (n = 96) was studied after weaning (28 days). We found no effect of sow’s reaction to human on piglets’ emotionality expressed during an open-field test (OFT) at 32–33 days. However, during a Human voluntary approach test (VHAT) piglets born from fearful sows were more fearful than piglets born from docile sows (Chisq=8.23, P = 0.004,). Half of the piglets (n = 48, 1 male and 1 female per sow) were then familiarised with human while the others received minimum contact. We found no main or crossed effect of sow’s reaction to human and piglet’s treatment on piglets’ emotionality during an OFT, nor on their exploratory behaviour during VHAT at 46–47 days. Piglets’ born from fearful sows were less attracted to human during VHAT2 (Chisq=8.33, P = 0.004) and familiarised piglets were more attracted (Chisq=15.05, P &lt; 0.001). Piglets born from docile sows found the reward faster in a spatial maze test than those born from fearful sows. Both sows HAR and piglet familiarisation with human seemed to specifically affect human related behaviours without affecting their general emotionality, but also behaviour in a spatial maze test.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"268 ","pages":"Article 106080"},"PeriodicalIF":2.3,"publicationDate":"2023-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49710656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Longitudinal stability of detection dog behavioral assessment: A follow-up study of long-term working success","authors":"Lucia Lazarowski ,&nbsp;Bart Rogers ,&nbsp;Jordan G. Smith ,&nbsp;Sarah Krichbaum ,&nbsp;Paul Waggoner","doi":"10.1016/j.applanim.2023.106082","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106082","url":null,"abstract":"<div><p>Detection dogs are a critical asset to homeland security, law enforcement, conservation efforts, and more. However, due to the challenges faced in operational working roles and the behavioral requirements to succeed in such roles, many dogs bred and trained for detection tasks are ultimately deemed unsuitable. Therefore, reliable methods for selecting dogs with the highest chances of success are needed. Studies aiming to identify the behavioral traits associated with success and methods for assessment have focused on short-term outcomes related to completion of training or selection for working roles. However, intervening factors may emerge leading to attrition once in service. The goal of the current study was to determine the longitudinal stability of performance on a detection dog behavioral suitability test as well as its predictive validity in determining working success. We followed a cohort of candidate detection dogs (n = 61) to determine changes in number of dogs successfully completing training and still working two years later, as well as determine the long-term predictive validity of the behavioral test. We also repeated the test in a subset of dogs (n = 34) to determine the longitudinal stability of performance on the test from puppy to mature adult. We found that training completion success rates were similar to current working success rates (i.e., minimal attrition). Additionally, performance on aspects of the puppy test such as emotional reactivity conducted at 3 months was predictive of current status (<em>p</em> = .008, OR =.246), and significantly correlated with performance at 3 years (<em>r</em> = .330, <em>p</em> = .028). These findings demonstrate longitudinal stability in behavioral traits considered important for detection dog success, which will be valuable for improving behavioral assessment and selection measures.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"268 ","pages":"Article 106082"},"PeriodicalIF":2.3,"publicationDate":"2023-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49733152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lifting laboratory rats: A survey of methods, handlers’ reasons and concerns, and rat behavioural responses","authors":"Charlotte C. Burn,&nbsp;Trinity Camacho,&nbsp;Jo Hockenhull","doi":"10.1016/j.applanim.2023.106077","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106077","url":null,"abstract":"<div><p>Lifting mice by the tail is of animal welfare and scientific concern, but rat lifting methods are little researched, potentially differing from mice. Using an online questionnaire we explored different methods for lifting laboratory rats, alongside handlers’ reasons and concerns, and rat behavioural responses. We received 249 valid responses from self-selected rat handlers across research sectors, job roles, and 26 countries (39% UK). In this sample, eight different lifting methods, plus ‘other’, were used; Shoulder saddle (39%), Chest-and-bottom support (20%) and Tail lifting (11%) were most common. Regarding respondents’ reasons for using their main lifting method, those using Tail lifting selected ‘rat comfort’ significantly less frequently, and ‘quick’ more frequently, than did those using other methods. Most respondents had no concerns about their main lifting method, but those using Tail lifting were significantly more concerned about possible ‘stress to the rat’ than were those using Chest-and-bottom support. Concerns about speed, feasibility and scientific validity were rare. Rats reportedly defecated significantly more, and sniffed the hand less, when Tail lifted than when lifted using Chest-and-bottom support. Respondents who used Cupped hands reported rats to approach, sniff and climb onto their hand significantly more than respondents using certain other methods did. Notwithstanding potential sample bias, the findings suggest that lifting rats by the tail risks unnecessary suffering. It is also potentially concerning that the most common method, Shoulder saddle, showed no significant benefits over Tail lifting. Chest-and-bottom support, Cupping, Tunnel, and some rarer methods, may offer more refined methods for lifting rats.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"268 ","pages":"Article 106077"},"PeriodicalIF":2.3,"publicationDate":"2023-10-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49733148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating video telemedicine for providing virtual health care for cats via mock spay recheck examinations","authors":"Grace Boone ,&nbsp;Melissa Bain ,&nbsp;Janet Cutler ,&nbsp;Carly M. Moody","doi":"10.1016/j.applanim.2023.106061","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106061","url":null,"abstract":"<div><p>Cat caregivers face many potential barriers when accessing veterinary care, such as cat-related challenges with carriers, travel, and the clinic environment, as well as time away from home and/or work, distance to the veterinary clinic, and costs of veterinary care. A potential solution to overcoming some of these barriers may be the use of video telemedicine: using video technology to deliver health and behavior care, education, and information remotely. This study recruited companion cat caregivers and their cats (n = 30) to participate in two mock spay re-check examinations: an in-clinic appointment and a virtual video telemedicine appointment. During the mock examinations, cat behavioral and physiological responses were assessed. Order of appointment was counterbalanced, and caregivers completed a pre- and post- appointment questionnaire to assess attitudes and experience with the in-clinic versus video appointments before and after participation. The study results show that cats had a higher respiratory rate (p &lt; 0.001), larger pupil dilation (p &lt; 0.001), and more negative ear positions (p = 0.015) during the mock clinic appointment compared to the video appointment. Cats were also more likely to leave the handling area after the mock clinic examination compared to the video examination (p = 0.001). In addition, caregiver ratings of the video appointment were very positive (62.1%) or somewhat positive (34.5%), and participants stated it was very easy to access the technology (58.6%), internet (65.5%), and video conferencing platform (75.9%) needed. Post-survey participant ratings indicated that video telemedicine was perceived to be more helpful than in-clinic visits for reducing caregiver stress (p = 0.029) and cat stress (p &lt; 0.0001). There was also a post-survey difference in caregiver ratings on accessibility of veterinary care between video telemedicine and clinic appointments (p &lt; 0.0001), with results indicating that video telemedicine increases caregivers’ access to cat care. Thus, from cat caregiver perspectives, video telemedicine may reduce stress and increase accessibility to cat health care compared to clinic visits. Results also suggest that video appointments reduce negative responses in cats compared to in-clinic visits, and thus may be particularly useful for cats who show higher levels of stress in a clinic environment. More research is needed to evaluate cat responses during other appointment types for further assessment and replication of the study results.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"267 ","pages":"Article 106061"},"PeriodicalIF":2.3,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49710853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"First impressions: A preliminary study of domestic dog puppies’ responses to verbal cues issued by an artificial agent","authors":"Nicky Shaw,&nbsp;Lisa M. Riley","doi":"10.1016/j.applanim.2023.106063","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106063","url":null,"abstract":"<div><p>Domestic pet dog puppies (hereafter puppies) typically co-exist and develop within a technology-rich environment, spending a large amount of time in the home. An audio/ food dispensing automated device (an artificial agent, hereafter agent) was evaluated for enrichment potential. The agent issued pre-recorded owner-spoken verbal cues to puppies and food rewards for their correct behavioural responses. Puppies’ latencies to look up at the agent after it spoke their name, the word ‘hi’, and dispensed food rewards, and their abilities to respond correctly to its cues were measured. Welfare during interaction was examined using quantitative continuous behaviour sampling. In the first of two conditions, puppies (n = 8) were observed responding to verbal cues issued by the agent, with researcher and owner present (A+RO). The agent was remotely triggered to call each puppy by name and issue the cues ‘sit’ and ‘down’ in a repeated, randomised order totalling eight cues per puppy. In a baseline condition for comparison (owner, with researcher and agent present) (O+RA) the owner replaced the agent’s role. Attentional focus (looking up at the agent with direct eye contact) was achieved following the agent speaking the puppies’ names, the word ‘hi’ and dispensing food rewards (<em>M</em> = 33 sec). In the subsequent two test conditions, no significant difference in puppies’ correct first choice behavioural responses to cues issued by the agent versus owner was found (p = 0.609). Significantly longer response latencies to cues in the agent versus owner condition were revealed (p = 0.001). No significant differences were found in state behaviours ‘looking up’ (p = 0.069), ‘cued response behaviour’ (p = 0.12), ‘eating food rewards’ (p = 0.263) and ‘close proximity post-test’ (p = 0.612) between conditions. A significant difference was found between the two conditions for ‘tail wag bout’, greatest in O+RA (p = 0.028). Puppies responded correctly to the agent’s cues during their first exposure to this novel paradigm and indicators of motivation to interact were identified throughout testing.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"267 ","pages":"Article 106063"},"PeriodicalIF":2.3,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49724029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Brown and white layer pullet hybrids show different fear responses towards humans, but what role does light during incubation play in that?","authors":"Maëva W.E. Manet ,&nbsp;Saskia Kliphuis ,&nbsp;Rebecca E. Nordquist ,&nbsp;Vivian C. Goerlich ,&nbsp;Frank A.M. Tuyttens ,&nbsp;T. Bas Rodenburg","doi":"10.1016/j.applanim.2023.106056","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106056","url":null,"abstract":"<div><p>Good early life conditions are increasingly recognized as essential to animal welfare later in life. The use of light during incubation might improve coping capacities and welfare in later life in poultry, by more closely approximating chicken natural environments compared to the current conventional incubation in darkness. Previous studies showed that lighted incubation resulted in more lateralized chicks, a more pronounced daily behavior rhythm, earlier onset of melatonin rhythms, and lower stress reactions to various stressors after hatching. Most existing research, however, has been conducted on broilers, and little information on lighted incubation is available for laying hens. In the current research, Dekalb White and ISA Brown eggs were incubated in complete darkness or in a cycle of green 12 L:12D throughout incubation, and five fear of human tests were performed on the 387 chickens during the rearing phase. We expected dark-incubated chickens to show stronger fear responses than light-incubated chickens. That was only the case for one of 15 behavior measurements taken during the tests (<em>p</em> &lt; 0.05). In addition, white layer hybrids are known to be flightier and more fearful than brown hybrids. In this study, white chickens indeed showed stronger fear responses than brown chickens in 12 of the 15 behavior measurements (<em>p</em> ≤ 0.002). Furthermore, we expected light during incubation to have stronger effects on white chickens than on brown chickens, because of the stronger transmission of light through white eggshells. However, the interaction between hybrid and incubation was never significant (<em>p</em> ≥ 0.18). Finally, contrary to our expectations, there was no effect of the incubation treatments or the hybrid on plasma corticosterone responses to a manual restraint test (<em>p</em> ≥ 0.36). Since there was a hybrid effect on behavior in this test, it is reasonable to think that behavior reflected coping style, rather than fear level. To conclude, the light regime used in this study does not seem as promising as expected to improve laying hen welfare. Finally, the brown hybrid was usually less fearful than the white hybrid, though there were some exceptions depending on the stressor, and that should be taken into account in research and in laying hen management.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"267 ","pages":"Article 106056"},"PeriodicalIF":2.3,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49733309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated methods to study animal behaviour","authors":"Irene Camerlink,&nbsp;Péter Pongrácz","doi":"10.1016/j.applanim.2023.106084","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106084","url":null,"abstract":"","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"267 ","pages":"Article 106084"},"PeriodicalIF":2.3,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49710849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"What is a companion animal? An ethological approach based on Tinbergen's four questions. Critical review","authors":"Péter Pongrácz,&nbsp;Petra Dobos","doi":"10.1016/j.applanim.2023.106055","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106055","url":null,"abstract":"<div><p>Having one or more non-human animals in our company became a steadily growing custom around the world. The terminology that describes these non-consumable animals encompasses a wide range as well (e.g. ‘hobby animals’, ‘pets’, ‘exotic animals’). However, their customary dubbing as ‘companion animals’ raises several questions that warrant for a systematic approach. In this review, we based our approach on the fundamental four questions developed for studying behaviour by Tinbergen, through which behaviour-based descriptors can be found for the various criteria that consider an animal as being a companion to humans. There is little consensus about whether a companion animal would be (1) any animal that people keep for other than consumption or working purposes; or in a more narrow sense, (2) companion animals should provide hedonistic value through rewarding interactions with their owners; or in the strictest sense, (3) only those animals can be regarded as companion animals that stay with their owners of their own free will (without confinement). According to Tinbergen’s four recommended levels of behavioural research, we assessed the literature with regard to the (i) Mechanism, where we investigated what kind of behaviour is expected from a companion animal in the above defined clusters. Next, we surveyed (ii) the role of Ontogeny in causing potential changes in the behaviour of particular animals, thus making them more suitable for companionship with humans. The question about (iii) Evolutionary background of various companion animal types came next, whether and which kind of forces of selection were needed for their existence. Finally, (iv) regarding the Function of behaviour, our focus was directed to the advantages (adaptive value) that companionship with animals can provide to both parties if we consider the various definitions of companion animals. With this critical review our goal was to provide an ethology-based framework for the description of various approaches to this elusive human-animal relationship. Rather than relying on hard-to-quantify terms borrowed from human sciences (such as ‘attachment’, ‘bonding’ or ‘emotional closeness’), we believe that the essence of being a companion animal should be determined with the terminology of biology, more closely, with the help of behaviour.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"267 ","pages":"Article 106055"},"PeriodicalIF":2.3,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49710904","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Positive stimulus and sexual behavior of rams","authors":"Ana María García-Mahecha,&nbsp;Rodolfo Ungerfeld,&nbsp;Lorena Lacuesta","doi":"10.1016/j.applanim.2023.106058","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106058","url":null,"abstract":"<div><p><span>The reproductive result of commercial flocks is influenced by rams’ reproductive performance<span>, including their sexual behavior. Although there are pharmacological strategies for improving their sexual performance, organic production systems require hormone free strategies. Rams can be conditioned to positive stimuli such as food that triggers a response associated with positive affective state that could influence sexual behavior. Thus, a positive affective state could be an alternative strategy to improve sexual behavior in rams. Therefore, the aim of this study was to compare rams’ sexual behavior toward an estrous ewe after accessing or not (control) a highly motivating food item (fruit). Initially, the rams' preference to consume bananas or apples was determined, and corneal and rectal temperatures, and heart rate were recorded for 12 rams. The rectal and mean corneal temperature increased after the consumption of the fruits (38.8 ± 0.05 °C vs 39.0 ± 0.05 °C, P = 0.006 and 35.2 ± 0.17 °C vs 36.1 ± 0.16 °C, P = 0.0007 respectively), without significant effects on heart rate. Thereafter, a test was designed to confirm whether they were enough motivated to consume the fruit, when offered close to an unknown subject making deterring behaviors. All rams responded positively to the test, approaching the fruit and consuming it for at least 30 s uninterruptedly in all the tests. For 10 days, rams were daily conditioned to the place where their sexual behavior would be evaluated, by entering a pen to which they were conditioned, in which the ram remained for 20 min with two ewes in the luteal phase immediately after consuming their motivational fruit. Finally, the sexual behavior of each ram was tested with a previous known ewe (in estrous) in two conditions: in the pen where they were conditioned with prior availability of the motivational fruit or in a similar pen to which they were not conditioned, without access to food. In this way, each ram was tested in both conditions with a crossover design. Rams only tended to display more </span></span>flehmen after eating the positive stimulus than when tested without positive stimulus (P = 0.08). The previous consumption of palatable food did not modify the display of any other sexual behavior. It was concluded that the consumption of a highly motivating food as a positive stimulus did not modify the display of rams’ sexual behavior.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"267 ","pages":"Article 106058"},"PeriodicalIF":2.3,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49733373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Discovering the hidden personality of lambs: Harnessing the power of Deep Convolutional Neural Networks (DCNNs) to predict temperament from facial images","authors":"Cihan Çakmakçı ,&nbsp;Danielle Rodrigues Magalhaes ,&nbsp;Vitor Ramos Pacor ,&nbsp;Douglas Henrique Silva de Almeida ,&nbsp;Yusuf Çakmakçı ,&nbsp;Selma Dalga ,&nbsp;Csaba Szabo ,&nbsp;Gustavo A. María ,&nbsp;Cristiane Gonçalves Titto","doi":"10.1016/j.applanim.2023.106060","DOIUrl":"https://doi.org/10.1016/j.applanim.2023.106060","url":null,"abstract":"<div><p><span>The objective of this study was to define a more practical and reliable alternative to manual temperament classification methods that rely on the behavioral responses of animals individually subjected to various tests. Specifically, this study evaluated the correlation between facial image information and temperament based on deep convolutional neural networks (DCNNs) to predict the temperament of lambs based on their facial images. In the first phase, the lambs were categorized as to their temperament based on data acquired from a behavioral test to establish a ground truth for the temperament of the lambs. This enabled us to train (70%), validate (20%), and test (10%) deep-learning models in the second phase based on facial images and the corresponding temperament labels derived from the behavioral test. The performance of a custom deep convolutional neural network (C-DCNN) was compared to that of pre-trained VGG19 and Xception models for image classification. The Xception model achieved a training accuracy of 81%, which indicated that it learned well the underlying patterns in the data; however, lower validation (0.75) and test (0.58) accuracies indicate that it overfit the training data and did not generalize well to new samples. The VGG19 model, produced lower training (0.59), validation (0.46), and test (0.34) accuracies, which indicated that it did not learn the underlying patterns in the data as well as the Xception model. Furthermore, its precision (0.47), recall (0.42), and F1 score (0.41) indicated that the model performed poorly in identifying the classes correctly. The C-DCNN produced a moderate accuracy of 60%, which indicated that the model was able to predict the temperament traits of lambs with an accuracy of 60%, which was better than random guessing (33% accuracy), and demonstrated the potential of this approach in assessing temperament. The C-DCNN precision (0.69), recall (0.61) and F1 score (0.63) indicated that it had a moderate ability to correctly identify positive cases; however, the small size of the original dataset remains a limitation of the study because it might have caused the suboptimal performance of the models. To validate this approach, further research is needed based on a larger and more diverse dataset. We will continue to investigate the potential of deep learning and computer vision to predict animal personality traits from facial images based on large, diverse datasets, which might lead to more efficient and objective methods for assessing animal temperament and improving </span>animal welfare.</p></div>","PeriodicalId":8222,"journal":{"name":"Applied Animal Behaviour Science","volume":"267 ","pages":"Article 106060"},"PeriodicalIF":2.3,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49761503","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}