The past near 20 years of research at ELTE’s Department of Ethology has confirmed our hypothesis that the numerous behavioural similarities between humans and dogs are the result of a convergent evolutionary process through which dogs have been able to adapt to the human environment. The functional similarities between the behavioural repertoires of the two species exist not just superficially but also within their components, on the level of ‘social primitives’. Domestication and directional selection has brought forth such variability that provides us with a unique opportunity to analyse underlying biological mechanisms, making the dog a promising model for studying the evolution of human social cognition.
We hope to further expand the understanding of the dog model through an interdisciplinary approach. To do so, we utilize a combination of two scientific approaches, resulting in an innovative synergic strategy that has not yet been applied elsewhere. The first approach is based on the traditional analysis of social behaviour and focuses on comparing human attachment, aggression, and socio-cognitive interactions to their analogues in canines. We plan to assess not only the behaviours themselves, but to also discover their genetic backgrounds through the most modern non-invasive methods known, including mapping the full genome, functional behavioural-genetic methods, and measuring brain activity by fMRI.
The second approach applies a more synthetic method whereby we hope to prove that the previously described human-analogue social behaviours demonstrated by dogs can also be applied to non-biological agents, for example social robots, to improve their abilities to interact with real humans. Much like how dogs had to learn to understand our complex social and verbal communication styles when they were domesticated, robots too will need to be able to interpret our social behaviours accurately.
With the cooperation of our national and international partners, we hope to lay down the basis of a natural animal model that will not only provide valuable information to the understanding of human evolutionary biology, but will also further and give new direction to robotics research.
The MTA-ELTE Comparative Ethology Research Group has been conducting canine research in two phases at ELTE’s Department of Ethology since 1998. Our initial goal of developing a more complete human evolutionary model through studying canine behaviour becomes more attainable each day.
Social robotics is a thriving field in building artificial agents. The possibility to construct agents that can engage in meaningful social interaction with humans presents new challenges for engineers.
In our opinion all social robots should be seen as companions and more conceptual emphasis should be put on the inter-specific interaction between humans and social robots. We argue that companion robots should be designed with a broad range of not necessarily sophisticated (human-like) social skills, which correspond to the expected function and the level of social interaction with humans. We identify problems in social robotics from an ethological point of view, and argue that a closer look at human-animal interaction, especially the detailed investigation of the social relationship between humans and dogs, may provide important insights for designing believable social behaviours for social robots.
We suggest that dog-human interaction provides a rich source of knowledge for designing social robots that are able to interact with humans under a wide range of conditions. Dogs are especially promising for inspiration because they share numerous social niches with humans. It is very important, however, that companion robots should not be a copy of dogs but their behaviour should delineate the general design features of social behaviours, which play an essential role in interactions that develop between dogs and humans.
The design of the robots should be maximally functional and only their behaviour should reflect the specific aspects of the dog behaviour complex that can emerge within the ranges of their functionality. Depending on their function, social robots should display the relevant subset of dog behaviours that suit best the actual social relationship.
In a previous large-scale EU project (EU FP-7-NEST (FP7-215554):“Living with Robots and Interactive Companions”, http://lirec.eu/) we worked in a collaboration of ten European partners. The goal of this project was to create interactive artificial companions, which are able to convey emotions and intentions in a socially acceptable and believable way towards the human partners, and which are capable of long-term relationships with humans.
Dogs, as a result of the domestication process, have adapted to the human social environment and evolved a set of social skills which make them an exceptionally suitable model species in studying human social behaviour. Besides, the fact that the canine genome is already described and many identified polymorphisms can be found also in the human genome, provides further support for using dogs as a model in the research of behaviour genetics.
Social Behaviour – similarly to several other phenotypes – is a complex trait. This means that both environmental factors and numerous genetic components contribute to the development of the given phenotype. Consequently the effect of a single gene is usually rather small and its identification is a great challenge, it is however important in understanding the molecular background of neural processes.
We investigate the genetic variations – polymorphisms and mutations – of candidate genes potentially related to social behaviour in dogs. Polymorphic variants of candidate genes are identified in silico employing public databases, as well as in vitro by direct sequencing of PCR-amplified genome regions of dogs and wolves DNA, purified from buccal cells. Our main targets are the coding regions and the regulatory un-translated sequences of the candidate genes. The identified polymorphisms are subsequently genotyped involving large populations representing different dog breeds. High-throughput techniques for large-scale genotype analysis include real-time polymerase chain reaction (PCR), allele-specific PCR, PCR combined with sequence specific digestion and downstream electrophoretic analysis.
The phenotypes are described through a set of behaviour tests developed for measuring dogs’ social behaviour, then genotype–phenotype association is assessed by statistical approaches. Biological effects of polymorphisms of interest are also studied by in vitro techniques.
Candidate gene analysis assumes that the phenotypic trait is determined to some extent by genes that have detectable effects. Candidate genes are usually related to the neurotransmitter and hormonal systems and the aim is to find a significant association between variation in the phenotype and the allele polymorphism.
Previously, we reported an association between the D4 dopamine receptor (DRD4) gene polymorphism and the activity-impulsivity (Héjjas et al., 2007a) and social impulsivity behaviour trait (Héjjas et al., 2009). Besides we identified anad analysed new variable number of tandem repeats polymorphisms (VNTRSs) in the genes of the dopaminergic neurotransmitter system of dogs and wolves (Héjjas et al., 2007b). Additionally, we found that tyrosine hydroxilase (TH) polymorphism was related to activity-impulsivity and to an ADHD-related trait assessed by a novel test battery in pet German Shepherd Dogs (Kubinyi et al., 2012). High scores in the test battery was characterized by high motor activity, high latency of laying down on the side, frequent vocalization and looking in the direction of the owner during separation from him or her. Recent results on sled-pulling Siberian huskies confirmed that both DRD4 and TH polymorphisms are associated with acitivity-impulsivity (Wan et al., 2013).
Dogs live with humans – we have shared a very similar social environment for tens of thousands years. But we know very little about how the dog brain encodes social and emotional signals of humans and other dogs. How similar or different are the neural mechanisms in dogs and humans that help to recognize conspecifics and other species, and that help to understand others’ emotions?
We train dogs to lay motionless in an fMRI scanner. This makes it possible to run the very same neuroimaging experiments on dog and human participants. Apart from a few studies on monkeys, ours is the first attempt to directly compare the neural processing in humans and a non-primate species. Dogs take part readily in these experiments, thus this species could become the non-human reference species for such comparative investigations – opening up the space for a completely new branch of comparative neuroscience.
Being mammals, humans and dogs have a similar brain structure, but very little is known about the localisation of specific functions in dogs. This comparative and non-invasive method, using the knowledge already established for humans, offers a nice way of finding out where specific neural processes take place in the dogs’ brain. There are many claims about the complexity level of mental processing in animals. The fMRI method offers a new tool for finding functional similarities between dogs and humans. In the long run, this research helps to achieve a deeper understanding of dogs’ mental skills.
In addition to investigate similarities and differences in the brain functions of dogs and humans with fMRI, we aim to perform a systematic analysis of neural regulation and gain finer temporal resolution of responses by using sleep EEG/ERP.
Our novel methods offer a totally new way of looking at the neural processing in dogs. We believe that our research will provide the first step to understand how dogs are able to tune into their owners’ feelings, and what makes the alliance of the two species so effective.
Our first MRI examinations were started in 2005 at the Institute of Diagnostic Imaging and Radiation Oncology, University of Kaposvár using a 1.5T scanner (Siemens). The aim of these studies was to develop a method to examine the neurological processes behind the dogs’ specific socio-cognitive behaviours with a modern neuroimaging technique. We managed to take appropriate anatomical images of awaken dogs’ brain; there was no significant difference between the images taken in the alert and anesthetized state. In our pilot fMRI study activation to images of a food reward from a beamer was obtained at the nucleus caudatus (part of the rewarding system in humans), and also at the lobus occipitalis (visual cortex). In case of somatosensory stimulation (stroking) activation appeared in the right somatosensory cortex and in the nucleus caudatus.