- Title: Plasticity and Learning in Cortical Networks
- Principal Investigator:
- Funding agency: European Union
- Grant type: Marie Curie Excellence Center
- Contract No.: MEXT-CT-2006-042484
- Duration: 2006-2010
- Amount: 1.2 Mio. Euro
Over the last years much progress has been made in characterizing different forms of plasticity that are thought to underlie the brain's remarkable learning abilities. Synaptic mechanisms have been the focus of much research. Intrinsic plasticity mechanisms, which change the nonlinear properties of individual neurons, have been less well studied in the past but are receiving more attention recently. While theoretical studies have led to an improved understanding of the computational properties of these plasticity forms, up to now they have mostly been studied in isolation. We currently simply do not know how these forms of plasticity might interact and what computational properties may arise from such interactions. We have recently suggested that the interaction of different forms of plasticity can give rise to qualitatively new and very interesting computational and learning properties. The goal of this proposal is to systematically investigate this exciting issue with the help of computer models and theoretical analysis. To this end, we study the learning abilities and dynamics of artificial neural network models which combine different plasticity mechanisms. These models will help to clarify how the brain can form effective internal representations of its physical and social environment and how it learns to use these internal representations to intelligently interact with the environment. This in turn may incite new developments in education, health care, and the design of artificial cognitive systems.