Validation of a Behavior Test for Predicting Puppies' Suitability as Detection Dogs

Development of a modified C-BARQ for evaluating behavior in working dogs

Introduction

Current high demand for effective odor detection dogs calls for the development of reliable methods for measuring performance-related behavioral phenotypes in these highly specialized working animals. The Canine Behavioral Assessment & Research Questionnaire (C-BARQ) is a widely used behavioral assessment tool among working dog organizations with a demonstrated ability to predict success/failure of dogs in training. However, this instrument was developed originally to study the prevalence of behavior problems in the pet dog population, and it therefore lacks the capacity to measure specific behavioral propensities that may also be important predictors of working dog success. The current paper examines the factor structure, internal reliability, and content validity of a modified version of the C-BARQ designed to evaluate four new domains of canine behavior in addition to those encompassed by the original C-BARQ. These domains, labeled Playfulness, Impulsivity, Distractibility, and Basophobia (fear of falling), respectively, describe aspects of canine behavior or temperament which are believed to contribute substantially to working dog performance.

Methods

Exploratory factor analysis (EFA) of owner/handler questionnaire responses based on a sample of 1,117 working odor detection dogs.

Results

A total of 15 factors were extracted by EFA, 10 of which correspond to original C-BARQ factors. The remaining 5 comprise the four new domains- Playfulness, Impulsivity, Distractibility, and Basophobia- as well as a fifth new factor labeled Food focus.

Discussion

The resulting Working Dog Canine Behavioral Assessment & Research Questionnaire (WDC-BARQ) successfully expands the measurement capacities of the original C-BARQ to include dimensions of behavior/temperament of particular relevance to many working dog populations.

A preliminary analysis of the effect of individual differences on cognitive performance in young companion dogs

Many factors influence cognitive performance in dogs, including breed, temperament, rearing history, and training. Studies in working dog populations have demonstrated age-related improvements in cognitive task performance across the first years of development. However, the effect of certain factors, such as age, sex, and temperament, on cognitive performance in puppies has yet to be evaluated in a more diverse population of companion dogs. In this study, companion dogs under 12 months of age were tested once on two tasks purported to measure aspects of executive function: the delayed-search task (DST) and the detour reversal task (DRT). Owners also filled out the Canine Behavioral Assessment and Research Questionnaire (C-BARQ) to evaluate how temperament influenced task performance. Contrary to prior research, performance did not improve with age on either task. However, the lack of age effects was likely the result of small sample sizes and individual differences across other factors influencing performance. Specifically, temperament differences as measured by the C-BARQ subscales for nonsocial fear and excitability predicted task performance on the DST, but the effect of temperament on task performance differed between males and females. Excitability also predicted performance on the DRT, but the effect depended on the age of the dog. In addition, no correlations were observed between task measures, indicating a lack of construct validity. Overall, these findings provide a preliminary analysis of factors that appear to influence cognitive task performance in young companion dogs and highlight suggestions for future research evaluating the impact of individual differences on cognitive performance.

Transposons in the Williams-Beuren Syndrome Critical Region are Associated with Social Behavior in Assistance Dogs

A strong signature of selection in the domestic dog genome is found in a five-megabase region of chromosome six in which four structural variants derived from transposons have previously been associated with human-oriented social behavior, such as attentional bias to social stimuli and social interest in strangers. To explore these genetic associations in more phenotypic detail-as well as their role in training success in a specialized assistance dog program-we genotyped 1001 assistance dogs from Canine Companions for Independence®, including both successful graduates and dogs released from the training program for behaviors incompatible with their working role. We collected phenotypes on each dog using puppy-raiser questionnaires, trainer questionnaires, and both cognitive and behavioral tests. Using Bayesian mixed models, we found strong associations (95% credibility intervals excluding zero) between genotypes and certain behavioral measures, including separation-related problems, aggression when challenged or corrected, and reactivity to other dogs. Furthermore, we found moderate differences in the genotypes of dogs who graduated versus those who did not; insertions in GTF2I showed the strongest association with training success (β = 0.23, CI95% = - 0.04, 0.49), translating to an odds-ratio of 1.25 for one insertion. Our results provide insight into the role of each of these four transposons in canine sociability and may inform breeding and training practices for working dog organizations. Furthermore, the observed importance of the gene GTF2I supports the emerging consensus that variation in GTF2I genotypes and expression have important consequences for social behavior broadly.

An artificial intelligence approach to predicting personality types in dogs

Canine personality and behavioural characteristics have a significant influence on relationships between domestic dogs and humans as well as determining the suitability of dogs for specific working roles. As a result, many researchers have attempted to develop reliable personality assessment tools for dogs. Most previous work has analysed dogs' behavioural patterns collected via questionnaires using traditional statistical analytic approaches. Artificial Intelligence has been widely and successfully used for predicting human personality types. However, similar approaches have not been applied to data on canine personality. In this research, machine learning techniques were applied to the classification of canine personality types using behavioural data derived from the C-BARQ project. As the dataset was not labelled, in the first step, an unsupervised learning approach was adopted and K-Means algorithm was used to perform clustering and labelling of the data. Five distinct categories of dogs emerged from the K-Means clustering analysis of behavioural data, corresponding to five different personality types. Feature importance analysis was then conducted to identify the relative importance of each behavioural variable's contribution to each cluster and descriptive labels were generated for each of the personality traits based on these associations. The five personality types identified in this paper were labelled: "Excitable/Hyperattached", "Anxious/Fearful", "Aloof/Predatory", "Reactive/Assertive", and "Calm/Agreeable". Four machine learning models including Support Vector Machine (SVM), K-Nearest Neighbour (KNN), Naïve Bayes, and Decision Tree were implemented to predict the personality traits of dogs based on the labelled data. The performance of the models was evaluated using fivefold cross validation method and the results demonstrated that the Decision Tree model provided the best performance with a substantial accuracy of 99%. The novel AI-based methodology in this research may be useful in the future to enhance the selection and training of dogs for specific working and non-working roles.

Digitally-enhanced dog behavioral testing

Behavioral traits in dogs are assessed for a wide range of purposes such as determining selection for breeding, chance of being adopted or prediction of working aptitude. Most methods for assessing behavioral traits are questionnaire or observation-based, requiring significant amounts of time, effort and expertise. In addition, these methods might be also susceptible to subjectivity and bias, negatively impacting their reliability. In this study, we proposed an automated computational approach that may provide a more objective, robust and resource-efficient alternative to current solutions. Using part of a 'Stranger Test' protocol, we tested n = 53 dogs for their response to the presence and neutral actions of a stranger. Dog coping styles were scored by three dog behavior experts. Moreover, data were collected from their owners/trainers using the Canine Behavioral Assessment and Research Questionnaire (C-BARQ). An unsupervised clustering of the dogs' trajectories revealed two main clusters showing a significant difference in the stranger-directed fear C-BARQ category, as well as a good separation between (sufficiently) relaxed dogs and dogs with excessive behaviors towards strangers based on expert scoring. Based on the clustering, we obtained a machine learning classifier for expert scoring of coping styles towards strangers, which reached an accuracy of 78%. We also obtained a regression model predicting C-BARQ scores with varying performance, the best being Owner-Directed Aggression (with a mean average error of 0.108) and Excitability (with a mean square error of 0.032). This case study demonstrates a novel paradigm of 'machine-based' dog behavioral assessment, highlighting the value and great promise of AI in this context.

Puppies in the problem-solving paradigm: quick males and social females

We report an observational, double-blind study that examined puppies’ behaviors while engaged in solving an experimental food retrieval task (food retrieval task instrument: FRTI). The experimental setting included passive social distractors (i.e., the dog’s owner and a stranger). The focus was on how the social and physical environment shapes puppies’ behaviors according to sex. The dependent variables were the number of tasks solved on an apparatus (Performance Index) and the time required to solve the first task (Speed). Sex and Stress were set as explanatory factors, and Social Interest, FRTI interactions, other behavior, and age as covariates. The main findings were that male puppies solved the first task faster than females. On the other hand, females displayed significantly more social interest and did so more rapidly than males. Males showed delayed task resolution. This study demonstrates sex differences in a problem-solving task in dog puppies for the first time, thus highlighting that sexually dimorphic behavioral differences in problem-solving strategies develop early on during ontogenesis.

The Use and Potential of Biomedical Detection Dogs During a Disease Outbreak

Biomedical detection dogs offer incredible advantages during disease outbreaks that are presently unmatched by current technologies, however, dogs still face hurdles of implementation due to lack of inter-governmental cooperation and acceptance by the public health community. Here, we refine the definition of a biomedical detection dog, discuss the potential applications, capabilities, and limitations of biomedical detection dogs in disease outbreak scenarios, and the safety measures that must be considered before and during deployment. Finally, we provide recommendations on how to address and overcome the barriers to acceptance of biomedical detection dogs through a dedicated research and development investment in olfactory sciences.

Artificial Selection Drives SNPs of Olfactory Receptor Genes into Different Working Traits in Labrador Retrievers

Labs as guide dogs or sniffer dogs in usage have been introduced into China for more than 20 years. These two types of working dogs own blunt or acute olfactory senses, which have been obtained by artificial selection in relatively closed populations. In order to attain stable olfactory attributes and meet use-oriented demands, Chinese breeders keep doing the same artificial selection. Though olfactory behavior is canine genetic behavior, genotypes of OR genes formed by breeding schemes are largely unknown. Here, we characterized 26 SNPs, 2 deletions, and 2 insertions of 7 OR genes between sniffer dogs and guide dogs in order to find out the candidate alleles associated with working specific traits. The results showed that there were candidate functional SNP alleles in one locus that had statistically severely significant differences between the two subpopulations. Furthermore, the levels of polymorphism were not high in all loci and linkage disequilibrium only happened within one OR gene. Hardy–Weinberg equilibrium (HWE) tests showed that there was a higher ratio not in HWE and lower FST within the two working dog populations. We conclude that artificial selection in working capacities has acted on SNP alleles of OR genes in a dog breed and driven the evolution in compliance with people's intentions though the changes are limited in decades of strategic breeding.

Advancing Genetic Selection and Behavioral Genomics of Working Dogs Through Collaborative Science

The ancient partnership between people and dogs is struggling to meet modern day needs, with demand exceeding our capacity to safely breed high-performing and healthy dogs. New statistical genetic approaches and genomic technology have the potential to revolutionize dog breeding, by transitioning from problematic phenotypic selection to methods that can preserve genetic diversity while increasing the proportion of successful dogs. To fully utilize this technology will require ultra large datasets, with hundreds of thousands of dogs. Today, dog breeders struggle to apply even the tools available now, stymied by the need for sophisticated data storage infrastructure and expertise in statistical genetics. Here, we review recent advances in animal breeding, and how a new approach to dog breeding would address the needs of working dog breeders today while also providing them with a path to realizing the next generation of technology. We provide a step-by-step guide for dog breeders to start implementing estimated breeding value selection in their programs now, and we describe how genotyping and DNA sequencing data, as it becomes more widely available, can be integrated into this approach. Finally, we call for data sharing among dog breeding programs as a path to achieving a future that can benefit all dogs, and their human partners too.

Expert Perspectives on the Performance of Explosive Detection Canines: Performance Degrading Factors

The explosive detection canine (EDC) team is currently the best available mobile sensor capability in the fight against explosive threats. While the EDC can perform at a high level, the EDC team faces numerous factors during the search process that may degrade performance. Understanding these factors is key to effective selection, training, assessment, deployment, and operationalizable research. A systematic description of these factors is absent from the literature. This qualitative study leveraged the perspectives of expert EDC handlers, trainers, and leaders (n = 17) to determine the factors that degrade EDC performance. The participants revealed factors specific to utilization, the EDC team, and the physical, climate, operational, and explosive odor environments. Key results were the reality of performance degradation, the impact of the handler, and the importance of preparation. This study's results can help improve EDC selection, training, assessment, and deployment and further research into sustaining EDC performance.