guest ::
| Home > Publications database > RateML: A Code Generation Tool for Brain Network Models |
| Home > Publications database > RateML: A Code Generation Tool for Brain Network Models |
| Poster (After Call) | FZJ-2022-05434 |
2022
Please use a persistent id in citations: http://hdl.handle.net/2128/32941
Abstract: It is argued that the relation of contemporary theoretical approaches to consciousness to the neurophysiology of the brain, can best be studied via detailed whole-brain modeling[1]. Often neuroimaging modalities such as fMRI, EEG and PET are applied to relate activation of brain regions or brain structure with consciousness[1-3]. In order to contribute to the discussion on consciousness, The Virtual Brain[4] (TVB), a neuroinformatics platform for full brain network simulations using biologically realistic connectivity, can be employed to relate brain structure to dynamics. In order to reduce workflow complexity for whole brain network models used by simulators such as TVB, a tool called RateML can be utilized. It enables scientist to generate models from a high level XML file containing constructs corresponding to aspects of the whole-brain model, abstracting the concern for the actual implementation. Its output is a Python model which can be used in the graphical interface of TVB and a model and simulator object for the Compute Unified Device Architecture (CUDA) parallel computing platform, enabling extreme parameter space exploration by making use of the highly parallel architecture of the Graphical Processing Unit (GPU). RateML has its foundation in Low Entropy Model Specification (LEMS)[5] which has been used to build NeuroML2. Nowadays, it is impossible to separate computational neuroscience from the study on consciousness, RateML makes the computational infrastructure accessible to scientists to make stronger and statistically significant claims about the brain.
|
The record appears in these collections: |
PDF
Fulltext by OpenAccess repository