High-performance computing in neuroscience - from physiologically realistic neurons to full-scale brain models

Chairs: Wolfram Schenck, Alex Peyser, Markus Butz (JSC)

Supercomputing is increasingly available in neuroscience and boosts the ability to create models with a degree of detail and biological realism never seen before. By the recently available computational power, single cell models can now represent a highly detailed neuronal morphology, compartmentalized functional interactions between synapses on a single dendritic branch and even molecular processes on a sub-synaptic scale.

Biological neuronal network models, too, become more realistic as representing a large amount of different cell types in a realistic layered cortical organization predicting dynamics of spike trains in cortical networks. Different forms of synaptic and structural plasticity can be combined in one model allowing us to study interfering activity and connectivity dynamics on different spatio-temporal scales. The ultimate goal is to generate full-scale brain models on the world’s high-end supercomputers. The hope is that physiologically realistic brain models will provide us deeper understanding of the healthy and diseased brain and offer novel tools to design new treatment strategies after brain lesions and for neurodegeneration.

The aim of this workshop is therefore to bring together the leading developers of high performance simulation and hardware tools in neuroscience with users from experimental fields and to demonstrate potential applications of the new techniques. Workshop speakers from the HPC domain are encouraged to present how their tools contribute to the scientific progress in neuroscience while experimentalists should point out the need for HPC resources in their workflow.

This workshop will be complemented by a special issue in Frontiers in Neuroanatomy entitled “Anatomy and plasticity in large-scale brain models”, https://www.frontiersin.org/research-topics/3644/anatomy-and-plasticity-in-large-scale-brain-models.