Dynamics of Cortical Networks Segregated into Layers and Columns
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
The neocortex covers 90% of the human cerebral cortex  and is responsible for higher cognitive function and socio-cognitive skills in all mammals. It is known to be structured in layers and in some species or cortical areas, in columns. A balanced network model was built, which incorporated these structural organizations and in particular, the layers, minicolumns and hypercolumns. The dynamics of eight different network models were studied, based on combinations of structural organizations that they have. The eigenvalue spectra of their matrices was calculated showing that layered networks have eigenvalues outside their bulk distribution in contrast to networks with columns and no layers. It was demonstrated, through simulations, that networks with layers are unstable and have a lower threshold to synchronization, thus, making them more susceptible to switch to synchronous and regular activity regimes . Moreover, introduction of minicolumns to these networks was observed to partially counterbalance synchrony and regularity, in the network and neuron activity, respectively. Layered networks, principally the ones without minicolumns, also have higher degree correlations and a reduced size of potential pre- and post- connections, which induces correlations in the neuronal activity and oscillations.
Place, publisher, year, edition, pages
2015. , 31 p.
dynamical, systems, neuroscience, computational, biology, modeling
IdentifiersURN: urn:nbn:se:kth:diva-176900OAI: oai:DiVA.org:kth-176900DiVA: diva2:868833
Subject / course
Master of Science - Machine Learning
2015-10-01, 527/F0, Lindstedtsvägen 24, Stockholm, 19:59 (English)
Kumar, Arvind, Assistant Professor
Lansner, Anders, Professor