Contatta l'autore

Michele Giugliano - Politecnico di Milano - [2001]

Preview

Indice

Bibliografia

Tesi completa: 198 pagine

Under the typical perspectives of the Computational Neuroscience and Bioengineering research fields, mechanistic and computational models of the nervous system can be classified according to two distinct approaches: at one end there are large-scale theories, where a large number of interacting simplified neurons interact and determine the collective behaviour, being much richer than those of the single units. At the other end, extremely detailed models of biochemical pathways and of the molecular, subcellular as well as synaptic mechanisms for information processing can be developed and studied. In the present thesis, an algorithmic strategy is developed in order to link those two quantitative description approaches, without making any unrealistic or simplistic assumption and with the aim of simulating large networks of synaptically interacting neurons, at low computational costs (i.e. memory and CPU time resources). This is strikingly relevant for the Neuroscience community, since synaptic chemical transmission and plasticity are nowadays believed to play a fundamental role in the adaptive properties characterizing the nervous system as a whole. Areas of related interest and applications of the topics presented in this thesis include the study of biophysical mechanisms for computation in neurons, the detailed computer simulations of neuronal circuits, the models of learning, the representation of sensory information in neuronal networks, systems models of sensory-motor integration, and computational analysis of problems in biological sensing, control and perception.