Abstracts & Slides
THE NEURAL BASIS OF EMPATHY IN HUMANS (slides)
Christian Keysers and Valeria Gazzola
Department of Neuroscience, UMCG, Groningen, The Netherlands & Netherlands Institute for Neuroscience, an institute of the Royal Dutch Academy of Arts and Sciences, Affiliations
c.keysers-at-nin.knaw.nl
Our capacity to smoothly and efficiently colaborate depends on the fact that we effortlessly understand what goes on in others humans around us although their goals and feelings are hidden neural states. How does our brain do that? I will show results from human fmri experiments and monkey single cell responses that suggest that while we view the actions of others, we activate our premotor, parietal and somatosensory cortices as if we were executing similar actions. While we view their sensations, our somatosensory areas are activated as if we felt similar sensations. While viewing their emotional facial expressions finally, we activate our premotor areas as if executing similar facial expressions and our insula as if experiencing similar emotions. Together, this set of activations plunges our brain into a state that resembles that of the people we observed, allowing us to ‘slip into their shoes’. Interestingly, these activations appear to be associated with what psychologists classically understand by empathy, as people that score higher on pen-and-paper empathy questionnaires activate their shared circuits more strongly. I will additionally discuss a Hebbian model of how mirror neurons could develop by observing ones own actions, and suggest that a similar architecture could possibly be used for artificial embodied agents to develop similar neurons. Finally, I will show that the human empathic system does respond to the sight of robots performing actions, without these robots needing to be particularly antropomorphic.
::
MEASURING STRATEGIC UNCERTAINTY AND RISK IN COORDINATION, ENTRY GAMES, AND LOTTERIES WITH fMRI
Andrea Brovelli (CNRS Marseille), Giorgio Coricelli (CNRS Lyon), Frank Heinemann (TH-Berlin), and Rosemarie Nagel (UPF-ICREA)
Neuroscientists and economists have recently begun to study jointly how strategic thinking regulate human individual and social behaviour. In this paper we measure strategic uncertainty in coordination games with strategic complementarities and substitutes in 2 person and large groups and risk within lottery setups, all framed in a similar way. The question we are addressing is whether the neural systems mediating decisions in individual and social context are distinct. More precisely, we are trying to identify whether risk and strategic uncertainty are mediated by different brain networks.
We found enhanced activity in bilateral anterior insula related to outcome uncertainty. We see clear differences in brain activity when comparing risk averse subjects and risk loving subjects playing lotteries and CC-games, but not when playing CS-games. This compliments our behavioral data which shows a strong correlation between risk attitudes and CC-games, but no correlation between risk attitudes and CS-games. Activity in the medial prefrontal cortex (mPFC), superior temporal sulcus, and temporo-parietal junction, and posterior cingulated cortex was related to playing in coordination and entry games. Increasing strategic uncertainty was correlated with neural activity in the mPFC.
Our results suggest that a common neural substrate (anterior insula) is shared in the individual and social contexts for the resolution of uncertainty. Moreover, the pattern of activity in the mPFC revealed the fundamental role of this area in strategic reasoning (Coricelli and Nagel, 2009).
TOWARDS THE GENETIC ARCHITECTURE OF COOPERATION
Since the publication, almost 150 years ago, of the Origin of Species the scientific study of the evolution of human-specific traits has been the focus of many efforts coming from very different areas of science. Nowadays, after a century an a half of research, impressive results have accumulated, particularly about those traits that, presumably, would “make us human”, setting us apart from the rest of primates, and about how these traits would have evolved. Over the last few years, a new area of research, genoeconomics, has started to make important contributions towards the study of hominization. Here, I review the foundations and promises of this new branch of science and discuss an ongoing Genome-Wide Association Study that we are conducting.
ON THE USE ARGUMENTATION IN AGREEMENT TECHNOLOGIES (slides)
Argumentation is one of the hot topics in current logic-based AI research. It is studied both as a form of (nonmonotonic) inference and as a kind of interaction between rational agents. In this talk the latter side of argumentation is discussed. An overview will be given of recent AI research on dialogue systems for argumentation, and of the the use of such systems in multi-agent interaction aimed at forming agreements.
REPUTATION AND CONFIDENCE FOR ARTIFICIAL [INTELLIGENT] ENTITIES, A COGNITIVE APPROACH (slides)
Since the moment you have autonomous entities interacting in open (often unstructured) environments, some kind of social control becomes a necessity to avoid chaos . The study of social control mechanisms that can be used in societies formed by artificial entities and how these mechanisms can be adapted to the specific characteristics of those societies has been a matter of study in different fields of Artificial Intelligence the last few years. Probably, one of the most studied mechanisms is reputation, attached to the notion of confidence (trust). However, in spite of the cognitive nature of reputation, until very recently, this nature has been substantially ignored when building models for artificial entities. We will present a cognitive approach to computational models of reputation and we will show how, by using this approach we are opening up a new range of possibilities for these models. The final goal of this research is to endow virtual entities with reputation models that can be used in environments where they have to interact not only with other artificial entities but also with human beings.
Adolf Tobeña
PSYCHOPHARMACOLOGY OF AGREEMENT (slides)
Adolf Tobeña
Department of Psychiatry and Forensic Medicine
Autonomous University of Barcelona.
Drugs have been tipically present on bargaining contexts either during negotiations (i.e. caffeine and other xanthines) or after achieving succesful pacts (i.e. alcohol). Cultures across the world have rich and varied traditions of using beverages or inhalations containing substances that may smooth and facilitate the interactive processes leading to fruitful agreements. Social rituals have a role, for sure, but specific molecules may also participate. In recent years experimental studies have documented solid effects of exogenously applied neuropeptides and steroids on economic games affecting trust, generosity and fairness/unfairness measures. An emphasis will be made on testosterone, dominance, status and bargaining data in different cooperative/competitive contexts. This evidence is opening new routes to explore the role of unsuspected factors operating during negotiations which may prove even useful for modelling efforts
OVERCOMING NOISE AND MISUNDERSTANDING IN SOCIAL DILEMMAS
Generosity, Empathy, and Trust
Paul A. M. Van Lange
VU University Amsterdam, The Netherlands
pam.van.lange-at-psy.vu.nl
Interpersonal misunderstanding is often rooted in noise, or discrepancies between intended and actual outcomes for an interaction partner due to unintended errors (e.g., not being able to respond to an email due to a local network breakdown). Incidents of noise are likely to give rise to misunderstanding (“he still has not responded to my email!”), which in the context of social dilemmas may evoke noncooperative rather than cooperative patterns of interaction (“next time, I will make him wait as well”). How can one effectively cope with incidents of negative noise in social dilemmas, situations in which self-interest and collective interests are conflicting? The major goal of the present paper is to review recent research, which seeks to examine whether (and how) generosity can reduce or eliminate the detrimental effects of negative noise on impressions and cooperation. The key message is that trust-building activities, through communicating generosity, helps people to maintain benign impressions of one another, and maintain high levels of cooperation – despite the detrimental effects of noise. Recent research complements this research by providing some evidence that empathy may enhance generosity, which in turn may help build trust.
PROFESSOR KRIPKE, LET ME INTRODUCE PROFESSOR NASH (slides)
Recent years have seen an enormous growth of interest in work at the intersection of logic and game theory. Forexample, researchers have used logic to make explicit the assumptions such as common knowledge that underpin many solution concepts, and have shown how solution concepts can be given a precise logical characterisation. From the perspective of computer science, these links open up the possibility of applying automated verification techniques, such as model checking, to the analysis and verification of computational mechanisms. We describe our work in this area, focussing on the use of game theoretic concepts in the specification and verification of coordination mechanisms for multi-agent systems.