Human Brain Project & EBRAINS at FENS 2020 Virtual Forum

Ida Aasebø (University of Oslo)

EBRAINS Data sharing is part of the EBRAINS Research Infrastructure, developed by and delivered to the research community the EU Human Brain Project. Learn more about what EBRAINS data sharing at ebrains.eu, what it means for you as a researcher and the underlying tools and services.

<iframe allowfullscreen="true" frameborder="0" height="250" src="https://www.youtube.com/embed/Lih-V_jxBNg" style="maxwidth=100%" width="450"></iframe>

Contact: curation-support@ebrains.eu

Maja Puchades (University of Oslo)

QUINT workflow for atlas based quantifications of features in 2D histological series.
In this session, QUINT workflow for atlas based quantification and spatial analysis of features like labelled cells or tract-tracing signals was presented. The workflow is  tailored for analysis of rodent 2D brain section images and is based on image segmentation, registration to a reference atlas and spatial analysis. All tools are open access as well as the instructional QUINT collaboratory.

Video on this session: Atlas based spatial analysis of histological images from rodent brain

QUINT Workflow Collab

M. Morgan (Swiss Federal Institute of Technology Lausanne)

You can now access the recording of this session via Facebook, Twitter or Youtube.

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/9RN3Y3rVHow" style="max-width:100% " width="450"></iframe>

Marie-Luise Neitz (Technical University of Munich)

You can now access the recording of this session via Facebook, Twitter or Youtube.

CEoI presentation pdf (601.2 KB)

CEoI presentation pptx (585.2 KB)

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/Z8MT9VbP7FY" style="max-width:100% " width="450"></iframe>

Anne Nahm (FZ Juelich) & Alex Upton (CSCS)

This presentation offered an introduction to the Fenix infrastructure which is providing compute, cloud and storage resources to neuroscientists and highlights a few neuroscience projects that are using the resources provided by Fenix.

You can now access the recording of this session via Facebook, Twitter or Youtube.

<iframe allowfullscreen="true" frameborder="0" height="250" src="https://www.youtube.com/embed/xie-lrsLXiA" style="max-width:100% " width="450"></iframe>

Maja Puchades (University of Oslo)

QUINT workflow for atlas based quantifications of features in 2D histological series.
In this session, the QUINT workflow for atlas based quantification and spatial analysis of features like labelled cells or tract-tracing signals was presented. The workflow is  tailored for analysis of rodent 2D brain section images and is based on image segmentation, registration to a reference atlas and spatial analysis. All tools are open access as well as the instructional QUINT collaboratory.

Video on this session: Atlas based spatial analysis of histological images from rodent brain

QUINT open collab

Steve Furber, Andrew Rowley (University of Manchester), Johannes Schemmel (Heidelberg University)

Presenting SpiNNaker and BrainScaleS for neuroscience. 

You can now access the recording of this session via Youtube and Twitter.

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/5Y5zsQhN_MI" style="max-width:100% " width="450"></iframe>

Q&A:

Q: What are the big breakthroughs you're hoping for with this architecture?

Johannes Schemmel: We hope to find the learning rules that give biological brains their  superior learning capabilities and apply them to novel computing  technologies, creating a generation of computers that is more energy  efficient and better suited to cope with our natural environment.

Steve Furber: Our aim is to contribute new insights to brain science and/or to bring insights from brain science across to mainstream AI. Significant progress along either of these directions would constitute a “big breakthrough” for us!

Q: What is the power consumption of the neuromorphic compared to normal one?

Johannes Schemmel: Due to the analog nature of our circuits the calculation of the effect of  a single spike on the neuron is several orders of magnitude more energy  efficient compared to a numerical calculation, always assuming  comparable silicon technologies.

Steve Furber: A measure of energy efficiency is the energy per synaptic event (SE) - the energy used to pass one spike through one synapse in a simulation. Biology delivers 10^14  SE/J. A very detailed supercomputer model such as those produced by the Blue Brain Project deliver 1 SE/J; a simplified supercomputer model might deliver 10^4 SE/J. SpiNNaker can deliver up to 10^8 SE/J, SpiNNaker-2 up to 10^9 SE/J, BrainScaleS, True North and Loihi around 10^10 SE/J. Biology wins handsomely!

For further questions please contact: neuromorphic@humanbrainproject.eu

Ingrid Reiten & Ulrike Schlegel (University of Oslo)

EBRAINS Data sharing is part of the EBRAINS Research Infrastructure, developed by and delivered to the research community the EU Human Brain Project. Learn more about what EBRAINS data sharing at ebrains.eu, what it means for you as a researcher and the underlying tools and services.

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/Lih-V_jxBNg" style="max-width:100% " width="450"></iframe>

Contact: curation-support@ebrains.eu

Viktor Jirsa, Spase Petkoski (Aix-Marseille Université) & Petra Ritter (Charité)

The Virtual Brain (TVB) is an open source platform for constructing and simulating personalised brain network models. End-to-end workflow for creating personalised brain models to running multi-scale brain simulations with The Virtual Brain are now available as integrated software solutions on the Human Brain Project’s  EBRAINS cloud.
• Find TVB tools and data via the Knowledge Graph
• Use TVB workflows via the Collaboratory
• Run TVB workflows in the EBRAINS Cloud or on HPC backend

Integrated TVB workflows on EBRAINS enable the processing of large cohort data sets - solving the problem of small numbers preventing generalizable results, a precondition for clinical translation.

You can now access the recording of this session via Facebook, Twitter or Youtube.

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/gPbFV81805M" style="max-width:100% " width="450"></iframe>

Karin Grasenick (convelop)

Ingrid Reiten & Ulrike Schlegel (University of Oslo)

EBRAINS Data sharing is part of the EBRAINS Research Infrastructure, developed by and delivered to the research community the EU Human Brain Project. Learn more about what EBRAINS data sharing at ebrains.eu, what it means for you as a researcher and the underlying tools and services.

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/Lih-V_jxBNg" style="max-width:100% " width="450"></iframe>

Contact: curation-support@ebrains.eu

Karin Grasenick & Julia Trattnig (convelop)

You can now access the recording of this session via Facebook, Twitter or Youtube.

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/UdAYk_mW82c" style="max-width:100% " width="450"></iframe>

Anne Nahm (FZ Juelich) & Alex Upton (CSCS)

This presentation offered an introduction to the Fenix infrastructure which is providing compute, cloud and storage resources to neuroscientists and highlights a few neuroscience projects that are using the resources provided by Fenix.

You can now access the recording of this session via Facebook, Twitter or Youtube.

<iframe allowfullscreen="true" frameborder="0" height="240" src="https://www.youtube.com/embed/AGtAmBD4R_c" style="max-width:100% " width="450"></iframe>