 |
|
 |
|
|
|
|
||
|
ResearchExperimental and theoretical neuroscienceOur researchers explore the brain by means of experimentally recorded data from humans (using EEG) and from the visual cortex of mammals (highly parallel recordings) and also through computer simulations of realistic models of cortical microcircuits. In addition, we are developing new, non-conventional techniques of data analysis. The first important line of investigation is related to the analysis of neuronal data recorded from real brains. We explore both high-density EEG data recorded in humans, as well as paralell recordings from mammalian visual cortex. The second line of research investigates neurocomputational principles in simulations of large scale networks of spiking neurons. This includes studies ranging from new biologically-inspired computational models of vision to general studies related to the dynamics of large-scale neuronal microcircuits and issues regarding their computational simulation. Finally, we are trying to develop new and non-conventional analysis tools for spike-trains, LFP and EEG signals, as well as for the characterization of the complex, non-linear dynamics of neuronal populations. Experimental and Theoretical Neuroscience Laboratory NeuroboticsOur work in neurobotics and computational embodied neuroscience bridges computational neuroscience, biologically-inspired robotics and artificial intelligence, evolutionary / developmental / epigenetic robotics, and artificial life. We are interested in the understanding of self-organizational and developmental mechanisms underlying the robustness and flexibility exhibited by animals, and in their reproduction in artificial systems. We would like to uncover robust, adaptive mechanisms through which an artificial embodied cognitive agent can internalize and structure the flux of sensorimotor information it has access to, thus developing its own conceptualization of the external environment. Our methodology uses important principles revealed by the study of brain and body as a whole integrated system interacting with the world (embodiment, interactivism, constructivism), to build models that on one side are inspired by the functioning of brains, and on the other side are capable of reproducing complex behaviors in embodied systems (robots or robot-like simulations). Most of the models we develop are based on spiking neural networks. The results of this research may lead to applications in robotics, automation, and to new methods of information representation and processing in artificial systems.
|
|||
|
|
|
|