|AISC 2017 Home||Accomodation and Venue||Practical Information||Invited talks||Call for papers||Committees||Programme||Registration|
Lawrence W. Barsalou
Discrete theories of emotion typically assume that dedicated neural circuits/modules originated in evolution to produce basic emotions in a relatively ballistic and rigid manner (e.g., fear, anger, disgust, sadness, happiness). Three neuroimaging experiments challenge this view. In Experiment 1, different assemblies of neural resources represented the same emotion in different situations (e.g., fear under threat of physical harm vs. social evaluation). In Experiment 2, different groups of participants learned to experience fear and anger either in physical harm or social evaluation situations, and later, when asked to anticipate these emotions, assembled different neural resources. In Experiment 3, different populations (cancer survivors, experienced meditators, controls) assembled different neural resources in response to the same emotional situations. Consistent with constructivist theories of emotion, these results suggest that experiencing an emotion “soft assembles” relevant neural systems throughout the brain to produce emotion in the current situation. To the extent that the same areas are utilized across multiple instances of a common emotional situation, an entrenched pattern develops that functions as an emotion attractor on future occasions. These patterns can be viewed as situated conceptualizations that control emotion via their grounding in the brain’s systems for perception, action, and internal states.
The epistemic value of hybrid bionic systems for the discovery of neural and cognitive mechanisms
Hybrid bionic systems (HBSs), connecting computer and robotic devices with nervous systems, enable partial restoration of sensory and motor faculties in people suffering from amputation or various kinds of disability. They have also been claimed to constitute new and promising tools for neuroscientific and cognitive research. What kind of epistemic roles can HBSs play, and under what auxiliary methodological assumptions can they be sensibly used for the discovery of brain and cognitive mechanisms? These questions will be explored by identifying and exemplifying a taxonomy of bionics-based methodologies, differing from one another in the experimental procedure, in the nature of the research question addressed, and in the epistemic requirements needed to bring HBS behaviours to bear on the neuroscientific or cognitive hypothesis in question. The proposed methodological framework may be useful to set up HBS-supported investigations on neural and cognitive mechanisms, and to pursue finer-grained epistemological analyses of the roles of HBSs in neuroscience and cognitive science.
Soft robots inspired by animals and plants: towards new abilities for unstructured environments
Robots today are expected to operate in a variety of scenarios, being able to cope with uncertain situations and to react quickly to changes in the environment. In this scenario a strong relationship between Nature and technology plays a major role, with the winning approach of evaluating natural systems to abstract principles for new designs. Bioinspired soft robotics is a worldwide known paradigm to develop new solutions for science and technology, giving way to a series of innovative robotic solutions assisting and supporting today’s society. A bioinspired approach needs a deep understanding of the selected biological models in order to extract the key features relevant for designing robotic systems able to imitate the biological counterpart in some specific functions. Such biological principles traditionally originate from animal models for robots that can walk, swim, crawl, or fly. Recently, engineers, material and computer scientists have also increased their interest in plants, as a new model for developing robots and ICT solutions. Among living organisms, plants represent valuable biological models to illustrate physical principles or develop mechanical devices. Plants are based on a completely different paradigm with respect to the animal kingdom. They are networked, decentralized, modular, reiterated, redundant, and resilient systems. They can be considered as a great example of multi-functional distributed living beings able to exploit in toto the intrinsic properties of their own structures and the interaction with the environment.
In this talk I will compare ideas, biological features, and technological translations from these two Kingdoms to areas of interest in robotics: movement, sensing and control.
Evidence for the interactionist theory of reasoning
The interactionist theory of reasoning suggests that the main function of human reasoning is to exchange arguments and justifications in a social setting. Support for this theory has come so far from reviewing work in several domains of psychology. Here I will present novel experimental evidence supporting the theory’s predictions: 1) argumentative competence is universal and early developing; 2) the improvement of performance in group discussion stems from sound argumentative competence; 3) there is an asymmetry in the way people evaluate their own and other people’s arguments.
Collective narratives on social media
Do echo chambers actually exist on social media? By focusing on how both Italian and US Facebook users relate to two distinct narratives (involving conspiracy theories and science), we offer quantitative evidence that they do. The explanation involves users’ tendency to promote their favored narratives and hence to form polarized groups. Confirmation bias helps to account for users’ decisions about whether to spread content, thus creating informational cascades within identifiable communities. At the same time, aggregation of favored information within those communities reinforces selective exposure and group polarization. We provide empirical evidence that because they focus on their preferred narratives, users tend to assimilate only confirming claims and to ignore apparent refutations.
Neural Dynamics of Cognition
In this talk, I will give an overview of Dynamic Neural Fields theory, which bridges the dynamics of large neuronal populations with cognitive processes and behaviour, such as memory formation, decision making, learning, or language grounding. I will show how cognitive architectures can be developed in this neural-dynamic framework and can be instantiated in an embodied setting, i.e. to control a robotic agent. Furthermore, I will present our first steps towards using neuromorphic technology to leverage neural fields to a “programming language” for neuromorphic cognitive robots.
The critical question is “how (in the brain)?” rather than “where (in the brain)?”
Neuroimages are important because they were instrumental in showing that mental processes depend on cerebral processes. However, neuroimages were seriously misleading in conveying the impression that they were able to unveil the working brain. This is simply wrong. Neuroimages reveal how the blood flows in certain brain areas. They do not reveal how neural activity is distributed in the brain. Bridging the gap between cerebral blood flow and neural activity is not a straightforward operation, and, as such, it is open to a number of stumbles. In addition, results obtained with neuroimages must be viewed with much caution because they were often rendered void by gross statistical mistakes in the way data were analyzed.
A further weakness in neuroimages studies is that they rest on the assumption that mental processes are “localized” in the brain, meaning that they depend on the neural activity that takes place in restricted areas of the brain (so-called centers). In contrast, it appears that mental processes are “distributed” in the brain. That is, they are likely to depend on the activity of complex neural networks.
Even assuming that the difficulties I have just outlined are met, there remains what, in my view, is the main problem with neuroimaging as an instrument for exploring mental processes. Traditional neuroimaging studies are aimed at answering the question of “where in the brain” a given process occurs. This question is not very interesting for exploring how the mind works. We must move on to the much more interesting question of “how in the brain” a given process is implemented. Traditional neuroimaging studies cannot tackle this crucial issue.