This title appears in the Scientific Report :
2013
A multi-area network model of the visual cortex - simulation andtheoretical description
A multi-area network model of the visual cortex - simulation andtheoretical description
We present a model comprising the 32 areas of the macaque cortex associated with visual processing, where the individual areas are based on a model of 1mm2 cortical patch. Combining a simple neuron model with complex connectivity enables us to study the infuence of the structural connectivity itself...
Saved in:
Personal Name(s): | Schücker, Jannis (Corresponding author) |
---|---|
Schmidt, Maximilian / Helias, Moritz / Diesmann, Markus | |
Contributing Institute: |
Computational and Systems Neuroscience; IAS-6 Computational and Systems Neuroscience; INM-6 |
Imprint: |
2013
|
Conference: | Jülich-Torino Workshop on Computational Neurosciences, Turin (Italy), 2013-10-04 - 2013-10-04 |
Document Type: |
Talk (non-conference) |
Research Program: |
Brain-inspired multiscale computation in neuromorphic hybrid systems Supercomputing and Modelling for the Human Brain Helmholtz Alliance on Systems Biology Signalling Pathways and Mechanisms in the Nervous System |
Publikationsportal JuSER |
We present a model comprising the 32 areas of the macaque cortex associated with visual processing, where the
individual areas are based on a model of 1mm2 cortical patch. Combining a simple neuron model with complex
connectivity enables us to study the infuence of the structural connectivity itself on cortical dynamics. Due to
the sheer size of the network the exploration of its parameter space comes with high computational costs, which
handicap the search for a good working point with realistic asynchronous activity at low ring rates. Here,
we present a reduction of the high-dimensional spiking model to a simpler system, replacing the dynamical
variables of each single neuron by mean ring rates of populations. This approach is based on mean-eld
techniques (Brunel, 2000; Fourcaud & Brunel, 2002) and enables us to theoretically predict the mean activity
in the network model. We present results from the simulation and compare them to the predictions of the
mean-eld theory. The mean-eld theory will provide further insights into the complex dynamics of the model,
revealing mechanisms which lead to oscillations in the system. Here we show how a stability analysis around
the mean activity state can expose the mechanisms that lead to slow oscillations on the timescale of seconds.
In a further step we aim to understand fast oscillations emerging in the macroscopic network model. For this
purpose we have to determine how single neurons transfer synaptic input to spiking output, i.e. the transfer
function. We here show how this transfer function can be determined empirically for neurons embedded in the
macroscopic model and we compare our results to analytical predictions. |