This title appears in the Scientific Report :
2019
Neural Network Mean-Field Analysis Toolbox
Neural Network Mean-Field Analysis Toolbox
In recent years, a lot of mean-field methods for calculating properties of commonly used random network models have been developed [1,2,3,4,5]. But, implementing them accurately can cause substantial work, just as it is to develop them. The availability of an easy-to-use implementation will foster t...
Saved in:
Personal Name(s): | Layer, Moritz (Corresponding author) |
---|---|
Senk, Johanna / Korvasová, Karolína / Helias, Moritz / Schuecker, Jannis / Bos, Hannah | |
Contributing Institute: |
Computational and Systems Neuroscience; INM-6 Computational and Systems Neuroscience; IAS-6 Jara-Institut Brain structure-function relationships; INM-10 |
Imprint: |
2019
|
Conference: | The 4th workshop on Advanced Methods in Theoretical Neuroscience, Goettingen (Germany), 2019-07-09 - 2019-07-12 |
Document Type: |
Poster |
Research Program: |
Doktorand ohne besondere Förderung GRK 2416: MultiSenses-MultiScales: Neue Ansätze zur Aufklärung neuronaler multisensorischer Integration Theory of multi-scale neuronal networks Theory, modelling and simulation |
Publikationsportal JuSER |
In recent years, a lot of mean-field methods for calculating properties of commonly used random network models have been developed [1,2,3,4,5]. But, implementing them accurately can cause substantial work, just as it is to develop them. The availability of an easy-to-use implementation will foster the wide use of these methods by a broad range of scientists, beyond the fraction of people who initially developed these method. The employed coarse-grained reductions enable insights into the mechanistic origin of network phenomena, they allow for targeted manipulations, and they enable the formulation of parameter constraints based on empirically observed activity; calculating network properties analytically is also often way faster than retrieving them from simulations. To support the widespread use of mean-field methods, we started collecting these implementations in an unified and easy-to-use framework of a python package. Currently, the focus is on random networks of leaky integrate-and-fire model neurons. In the future, we are planning to extend this package to include more tools and to support more neuron and network types. |