NESTML – Creating a Neuron Modeling Language and Generating Efficient Code for the NEST Simulator with MontiCore
NESTML – Creating a Neuron Modeling Language and Generating Efficient Code for the NEST Simulator with MontiCore
Neuroscientists use computer simulations as one way to research the brain and brain activity. They developed and published numerous neuron and synapse models with different levels of detail to be used in simulations of single neurons or large biological neuronal networks. Besides the neuron and syna...
Saved in:
Personal Name(s): | Ippen, Tammo (Corresponding Author) |
---|---|
Contributing Institute: |
Computational and Systems Neuroscience; IAS-6 Computational and Systems Neuroscience; INM-6 |
Published in: | 2013 |
Imprint: |
2013
|
Physical Description: |
145 p. |
Dissertation Note: |
RWTH, Masterarbeit, 2014 |
Document Type: |
Master Thesis |
Research Program: |
Addenda |
Subject (ZB): | |
Publikationsportal JuSER |
Neuroscientists use computer simulations as one way to research the brain and brain activity. They developed and published numerous neuron and synapse models with different levels of detail to be used in simulations of single neurons or large biological neuronal networks. Besides the neuron and synapse models, the neuroscience commu- nity has developed several simulators with different scope and, mostly, incompatible model description languages. This makes it hard to develop and publish new neuron and synapse models and even harder to compare and verify findings across simulators, since the model must be implemented and adjusted for every simulator.This thesis describes the design of NESTML and its development with the MontiCore framework. NESTML is an extendable modeling language for the neuroscience do- main. It allows modeling spiking point neuron models in a clean and concise syntax. An associated processing tool performs static analysis on NESTML models to check for programmatic correctness and, thus, supports neuroscientists in creating new neuron models. Furthermore, it generates efficient code for the NEST simulator and the NEST module infrastructure, which allows to easily compile and load the generated code into NEST. This reduces the work to create and to maintain neuron models for NEST and, by adding more simulator targets in the future, across simulators. |