Funded grants in 2017

Here is the full list of grants funded in the 2017 funding cycle

Projects

European Network in Brain Imaging of Tumour (ENBIT)

Anders Eklund
Linköping University, Sweden

Team

Dr L. Douw, Dpt of Anatomy and Neurosciences, VU University Medical Center in Amsterdam, Netherlands
Dr D. Marinazzo, Data Analysis dpt, University of Ghent, Belgium
Dr C. Pernet, Dr D Rodriguez and Dr D Job, Centre for Clinical Brain Sciences, University of Edinburgh, United Kingdom
Prof F-E Roux, CNRS - Cerco, Toulouse, France
Prof C Stippich, Dpt Diagnostic & Interventional Neuroradiology, University of Basel, Switzerland

Abstract

This proposal aims at the creation of a European Network in Brain Imaging of Tumour (ENBIT). The seed funding will help by allowing the creation of a basic data-sharing infrastructure (a repository with web access) with some data already available, and the organization of an open workshop. We believe that by pooling expertise, resources and data, we will be able to (i) accelerate progress in understanding how tumours affect the brain and therefore how to treat them, (ii) reduce the overall cost of research (iii) advance clinical practices and (iv) foster collaborations and provide a large resource usable for many purposes, including training and education. This is particularly relevance for neuroscience and neuroinformatics as it will provide a unique set of brain imaging data that can offer a window into brain plasticity due to the variety of tumors (slow vs fast growing, confined vs. invasive) and a resource for new clinically oriented informatics tools.


Workshop for Neuroinformatics in Aging

Eric Tatt Wei Ho
Universitii Teknologi PETRONAS, Malaysia

Team

Toshiharu Nakai (National Center for Geriatrics & Gerontology, Japan)
Hirohisa Watanabe (Nagoya University, Japan)
Fan-Pei Gloria Yang (National Tsinghua University, Taiwan)
Annabel SH Chen (Nanyang Technological University, Singapore)
Muzaimi Mustapha (Universiti Sains Malaysia)
Ahmad Fadzil Mohd Hani (Universiti Teknologi PETRONAS, Malaysia)

Abstract

We seek INCF funding for a scientific workshop on Neuroinformatics for Aging. The workshop is for planning a collaboration around the creation and application of neuroinformatics technologies to address clinical and wellness challenges in aging. We will organize a scientific symposium which will become an annual meeting between neuroscientists, neuroinformaticians, roboticists, neuropsychiatrists, neurologists and neurosurgeons. Forging international collaborations will be the primary goal at these meetings, as the means to (1) find and match expertise to project needs of each member group/country (2) promote the dissemination of tools, analysis techniques and knowledge of neuroinformatics (3) raise the visibility of neuroinformatics to researchers and clinicians from each member country through important use-cases in aging. The SIG will work towards 3 initial objectives which is (1) to establish a shared Asia Pacific neuroinformatics data repository and analytics workflows (to be seeded by collaborators from Japan, Taiwan, Malaysia & Singapore) (2) to define collaborative research projects and seek joint funding from national funders and (3) to develop a collaborative plan for neuroinformatics training for graduate & postdoctoral students in the Asia Pacific region.


Next-generation beamformer development in Python for human MEG/EEG oscillations

Karim Jerbi
University of Montreal, Department of Psychology, Canada

Team

Sarang S. Dalal, PhD, Neuroelectromagnetic Oscillations Lab, Aarhus University, Denmark
Denis Engemann, PhD, MNE-Python core developer, INRIA Saclay, France
Caroline Witton, PhD, Wellcome Trust Laboratory for MEG Studies, Aston University, UK
Alexandre Gramfort, PhD, MNE-Python core developer & founder, INRIA Saclay, France

Abstract

Beamformers are a class of high-resolution spatial filters that are popular for neural source reconstruction from MEG and EEG data in both epilepsy diagnostics and cognitive neuroscience. They are particularly suited to detecting weak neural oscillations. However, active beamformer development has mainly taken place in proprietary software or open-source toolboxes on proprietary platforms. This complicates analysis pipelines on computing clusters and hinders researchers with limited resources.
 
This project aims to port several recent developments in beamforming to the open-source MNE software. MNE is written in Python and has a rapidly expanding user base. It now aims to become a state-of-the-art platform for continued development of beamformer innovations.
 
Together with beamforming, circular statistics are crucial for analyzing neural oscillations that enable robust analysis of phase relationships between brain regions or across time. The project secondarily aims to extend Python support of circular statistics for the large scale of MEG/EEG datasets. This will make a compelling contribution beyond neuroscience to any field concerned with periodic signals.
 
By implementing both minimum norm estimates and beamformers in one consistent general source imaging framework, this project will furthermore catalyze systematic and fair comparisons between different source imaging methods to determine which one is more appropriate for a given app


Workshops

Creating online mouse whole-brain atlases from your own data

Daniel Fürth
Karolinska Institutet, Sweden

Partners

This is a workshop that will be organized by Karolinska Institutet (with Daniel Fürth as a Ph.D. student in Konstantinos Meletis and Marie Carléns groups at Karolinska Institutet, Sweden).

Purpose of workshop

This is a three day workshop that takes place in Stockholm, Sweden, at the Karolinska Institutet. During this workshop students and researchers are introduced to the open source software solution for whole mouse brain mapping efforts, called WholeBrain (http://wholebrainsoftware.org/cms) and it’s web-based framework for visualization and exploration of results called Openbrainmap (http://openbrainmap.org). Together these computational frameworks offer a wide range of tools: image processing pipelines for image segmentation and registration, Bayesian statistical packages handling nested hierarchical data, and a framework to produce interactive ‘Google-maps’-like neuroanatomical data. Users will receive detail instructions and demonstrations on how to use the tools all the way from sample treatment, microscopic imaging, setting up of a pipeline to high-level statistical analysis.


Travel

Human Atlasing Working Group

Alan Evans
McGill University, Canada

Partners

JB Poline jbpoline@gmail.com, D. Kennedy, S. Ghosh, G. Flandin(SPM)*, M. Jenkinson (FSL)*, L. Zollei (Freesurfer)*, Michael Halle (Slicer)*, L. Ng (Allen Brain Institute)*, R. Vincent (MNI). We hope to fund the participants(*) but will also keep online communication through github/slack with a larger group.

Purpose of travel

Researchers using neuroimaging techniques rely on some brain template space, such as the MNI 305 space, and one or several atlases providing anatomical or functional information for a given location in this space. Most atlases are constituted of either hard labels (zero, one, or several labels for a given position), or probabilistic information indicating the likelihood of a label. Atlases exist for both 3D and surface geometry and can represent various types of information (e.g. cytoarchitectonic or macroscopic anatomical or functional information, etc).
 
The use of atlases is ubiquitous and many have been proposed in the past. For anatomical structures only, Harvard-Oxford (HO), Automatic Anatomical Labelling (AAL), Freesurfer labels, JuBrain, or Talairach are amongst the most used ones. There is, however, no standard way of describing the information contained in an atlas. Even for very similar types of atlases such as the HO and AAL, similar information is encoded differently.
 
Some initial work has been done to propose a standard meta data description of Atlases (see https://github.com/INCF/HAWG-examples). We however need to finalize some choices, the implementation of the standard format for several well used atlases, and provide reference implementations. We propose to gather a 4-5 individuals in Montreal during the fall to achieve these goals.


Standards and algorithms for modelling functional electrical connectivity

Carlos H. Muravchik
Universidad Nacional de La Plata, Argentina

Partners

Dr. Alejandro Blenkmann; Prof. Anne-Kristin Solbakk, Department of Psychology, Oslo University, Norway

Purpose of travel

This is to help continue a long time collaboration with Dr. Blenkmann and initiate contact with the group of Prof. Solbakk, via a short-time research visit. We will define how to study local connectivity, using our current data with intracranial electrodes on patients with epilepsy. By observing signals at different scales -single neuron, local field potentials (LFP) and sEEG- we intend to determine (decide on methods or develop) suitable space-time algorithms to model effective and functional connectivity. We expect to be able to explore the relationship among regions of the epileptogenic network defined at a macroscale (such as epileptogenic, seizure onset, irritative, lesion, functional deficit zones) with mesoscale phenomena (such as wavefront propagation and 'ictal penumbra' in LFP) and even to the microscale (such as the so-called microseizures).
 
The meeting will be devoted to agree on procedures to deal with side-information (surfaces, conductivities, etc) and to develop signal processing algorithms providing the chosen connectivity measures. In order to expand the communication with the group in Oslo, a talk presenting the applicant recent results will be given. Moreover, connectivity measures are known to change in a task-relevant manner, reflecting changes in response to sensory, motor and cognitive events. This is expected to be of interest for other research topics by the Norwegian group.


Developing a novel detection method of dynamic cell assemblies in neural data

Masami Tatsuno
University of Lethbridge, Canada

Partner

Tomoki Fukai, RIKEN Brain Science Institute, Japan

Purpose of travel

The purpose of travel is to develop a novel detection method of dynamic cell assemblies in neural data such as multi-neuronal activity obtained by electrophysiological and imaging recordings. Evidence suggests that information processing in the brain is performed by complex dynamics of cell assemblies. It is therefore crucial to develop a powerful detection method of dynamically organized cell assemblies. To this end, the travel is aimed at promoting collaboration between the two institutions, Canadian Centre for Behavioural Neuroscience, the University of Lethbridge, Canada (UofL) and RIKEN Brain Science Institute, Japan (BSI). Tatsuno at UofL has extensive experience in multi-electrode recording from freely behaving rodents. Fukai at BSI is a world-leading computational neuroscientist who has developed various analysis tools such as the EToS spike-sorting software (http://etos.sourceforge.net/). His group is also developing a prototype of novel cell-assembly detection method based on edit similarity (CADES). During Tatsuno’s visit to Fukai’s lab, we will test the performance of CADES by applying it to Tatsuno’s electrophysiological data and seek a way to improve its performance. This development requires intensive discussion and software implementation. Therefore, in-person communication by a visit is crucial for this attempt to be successful. The developed software will be made freely available to neuroscience community.


Courses

Topics in brain-like computing

Pawel Herman
KTH Royal Institute of Technology, Department of Computational Science and Technology, Sweden

Team

The core team consists of 2 principal developers, Pawel Herman and Arvind Kumar, who are active researchers in computational neuroscience and neurocomputing in the Computational Brain Science group at KTH Royal Institute of Technology. They have international networks of collaborators in neuroscience, and are embedded in the interdisciplinary brain research environment in Stockholm region including Karolinska Institute and Stockholm University. Pawel and Arvind are teaching subjects in neuroscience, human perception, artificial neural networks and machine learning. The course organisation will be supported by other academics at KTH, e.g. A. Lansner and Ö. Ekeberg - pioneers of neurocomputing and computational neuroscience in Sweden. We will also rely on other competences in our School at KTH: autonomous agents (C. Peters), cognitive robotics (M.Björkman), deep learning (G. Salvi). The course is aligned with our recent initiative of building a Brain-IT community in Stockholm region.

Abstract

The course accounts for recent trends to exploit accumulating knowledge in brain science for devising the next generation computational intelligence paradigms. The vision is to draw from the information processing solutions adopted by brain to accelerate the development of novel approaches in popular areas of data analytics, machine learning, cognitive robotics, autonomous agents etc. With the support of interactive lectures and seminars we plan to organise (with publicly available teaching materials, i.e. slides, notes, videos) we will provide an interdisciplinary, blended learning environment for students and adepts of neuroinformatics, neuroscience, physics, math and computer science, among others. The aim is to increase the awareness of new opportunities emerging in the field beyond already established contribution of computational sciences to neuroscience, and facilitate discussions of the operating principles and functionality of the brain-like machine intelligence technology of tomorrow. The course will therefore complement the existing neuroinformatics teaching curricula. The intended impact is to introduce this new perspective of the brain-IT synergies, through the prism of brain-like computing, into the mainstream scope of neuroinformatics, and contribute to the education of future academic and industrial leaders engaging in long-term efforts towards embracing technological challenges ahead. This should in turn lead to a new wave of advancements in Neuroinformatics.


Computational modeling of neuronal plasticity

Florence I. Kleberg
Frankfurt Institute for Advanced Studies, Jochen Triesch lab of Neuroscience, Germany

Team

Florence I. Kleberg: Course creation
Prof. Jochen Triesch: Course co-creation and quality supervision
Gaby Ehlgen: Research funding officer

Abstract

The brain is highly complex, and how the interaction between neuronal activity and plasticity can lead to stable yet flexible brain function is an active topic of research. We propose an open-access, online course that will prepare students for addressing outstanding questions in brain plasticity through simulations of neurons and networks. The course will teach students to program their own spiking neuron model from scratch, empowering them to implement various types of plasticity and observe their interactions. Using an inverted classroom approach, students will self-study with the help of online content and advance through the course through different “levels” while a local instructor or tutor answers questions and monitors progress. Tutors will unlock new content for the student when an adequate solution has been provided to the previous problem. Students taking the course will consent into tutoring one student after them, so that participants without a local tutor will be able to ask questions and obtain study guidance, leading to a self-sustained study system. For additional support, a solution manual and Python source code for tutors will be added as well as an online discussion forum. The course is designed for bachelor and master students from various natural science backgrounds as a direct preparation for research in computational neuroscience, but it is also appropriate for enthusiasts from unrelated fields.


Lifecycle of human electroencephalography/event related potential data

Roman Mouček
University of West Bohemia, Department of Computer Science and Engineering, Czech Republic

Team

The team consists of the members (employees, Ph.D. students and master’s degree students – prospective Ph.D. applicants) of neuroinformatics laboratory at the Faculty of Applied Sciences, University of West Bohemia (UWB), Pilsen, Czech Republic, the UWB lawyer and the experts in electrophysiology (electroencephalography (EEG) and event related potentials – (ERP)) from the University Hospital (UH) in Pilsen. R. Mouček, UWB, researcher, project coordination, data and metadata standards, P. Brůha, UWB, researcher and Ph.D. student, data annotation, data curation, L. Vařeka, UWB, researcher and Ph.D. student, data analysis, K. Černá, UWB, Ph.D. student, statistical analysis, J. Vaněk, UWB, Ph.D. student, data standards, M. Navrátilová, UWB, EU legislation, I. Holečková, UH, expert in electrophysiology (EEG, ERP), V. Kraft, P. Šnejdar, V. Vacek, UWB, students of master’s degree, software and hardware infrastructure, data security, data storing, M. Minářová, UH, data collection.

Abstract

The goal of the proposed project is to develop a training material that describes the complete lifecycle of electroencephalography/event related potential (EEG/ERP) data obtained from human subjects. It will be beneficial to all participants round the world interested in EEG and ERP techniques, and collecting, annotating, standardizing, storing, processing, sharing and publishing data from electrical activity of the human brain. The training material will be data centric; it will guide participants from the basic characteristics of EEG/ERP data and metadata to their long-term preservation and publication. Data and metadata standards, standardization initiatives, data sustainability, and ideas of data sharing and data openness will be emphasized. The training material will be delivered in the form of 120+ minute long video tutorial consisting of six 20+ minute parts and covering e.g. description of the EEG/ERP technique, EEG/ERP data and metadata characteristics, design of EEG/ERP scientific project, hardware and software infrastructure for data collecting, legal issues related to data and metadata collection, storage and use, standards and tools for data and metadata annotation, storage, sharing and publication, and methods for data preprocessing and processing. The tutorial generally wants to help change the view of the importance of data in neuroscience and increase the efficiency and effectiveness of research in neuroscience.


Advanced Scientific Programming in Python Down Under

Juan Nunez-Iglesias
University of Melbourne, Melbourne Bioinformatics, Australia

Team

Juan Nunez-Iglesias, expert in scientific Python, co-author of the book Elegant SciPy, and previous faculty at ASPP.
Tiziano Zito, founder of the original ASPP, and frequent contributor to the SciPy ecosystem.
Jack Simpson, graduate student at the Australian National University, Software Carpentry (SWC) certified instructor and organiser of several SWC workshops at ANU.
Stéfan van der Walt, Fellow at Berkeley Institute of Data Science, creator of the scikit-image library, and frequent contributor to NumPy and SciPy, among others. Also an experienced faculty at ASPP.

Abstract

Software plays an increasingly important role in scientific research. This is especially true of neuroinformatics. However, most scientists are not trained in modern software engineering concepts and best practices. This is true even of those that regularly write scripts and software to achieve their research goals. This week-long course will teach attendees to use modern programming tools effectively. This will empower them to write clean, fast, and bug-free code that produces reliable scientific results, and that can be re-used by the community. The course uses Python as the teaching programming language because of its increasing prominence in scientific computing. In astronomy, where we have the best data on scientific publication, Python use has grown almost 20-fold over the past decade. In Neuroinformatics, too, Python is fast becoming the de-facto language of data acquisition, analysis and simulation. All the major simulators, such as NEST and NEURON, now offer Python as the preferred interface. ASPPDU will cover software development techniques such as version control and test-driven development, as well as advanced Python topics such as distributed computing, interfacing with C code, and advanced SciPy and NumPy applications. Students will be able to contribute to any major scientific software project, as well as develop their own.