Travel grant reports 2010
Marc de Kamps, University of Leeds, UK
Arrangement of a Grant Application Meeting, 18 August 2010, Leeds UK
Activities and accomplishments:
On July 13, 2010, we received confirmation of a travel grant from the INCF for the purpose of constructing a consortium for a proposal submission of the FET Open Scheme (http://cordis.europa.eu/fp7/dc/index.cfm?fuseaction=usersite.FP7DetailsCallPage&call_id=319).
The meeting was attended by:
• Eduardo Fernandez, University of Elche, Spain
• Kevin Gurney, University of Sheffield, UK
• Jim Austin, Cybula Ltd and University of York, UK
• Ingo Bojak, University of Nijmegen, Netherlands
• Marc de Kamps, University of Leeds, UK
• Karim Djemame, University of Leeds, UK
• Matthia Hovestadt, TU Berlin, Germany
• Gaute Einevoll, UMB, Norway
The consortium agreed to submit a proposal on using novel modeling techniques linking the neuronal level (Fernadez) with the population level (de Kamps). A novel forward modeling approach (Einevoll) would be used to predict imaging signals (Bojak) from attention experiments (Gurney, de Kamps). The considerable computational requirements would be handled by a cloud-based infrastructure on CARMEN (Austin, Djemame) with commercial spin-offs of the new technology (Hovestadt, Djemame).
We agreed a five-page submitted the proposal on 1 December 2010. Unfortunately on 17 March 2011 we heard we had not passed the first submission stage. The feedback on the proposal was negative in the sense that the project seemed plausible, but did not present the required breakthrough ideas required for FET Open. We will look for resubmission in specific calls on Cloud computing because the feedback was such that we do not believe that FET Open is the right call at the moment. The consortium, however, will pursue further grant applications.
Gaute Einevoll T, Mathematical Sciences and Technology, Norwegian Univ of Life Sciences, Aas, Norway
Ingo Bojak, Donders Institute for Brain, Cognition and Behaviours, Radboud University Nijmegen, The Netherlands. (May 2010)
Purpose and Accomplishments:
The main purpose was to meet with representatives of the Dutch INCF Node to discuss future collaborations, in particular modelling of extracellular potentials and specifically how to add this component to large-scale network models to allow the prediction of EEG signals.
Furthermore, during the visit a talk was given at Donders Institute for Brain, Cognition and Behaviours, Radboud University Nijmegen.
The purpose was to attend the Methods in Computational Neuroscience (MCN) course at the Marine Biology Laboratory, Woods Hole, MA, USA during August 2010. For this summer school a program was put together which focused on examining neural processes using theoretical interpretations that were explainable by biological mechanisms. Like the faculty, the students came from diverse backgrounds, from physics, mathematics, neurophysiology and systems biology, and hailed from many universities internationally. As a student who worked with mainly theoretical models, I was able to speak to faculty and fellow students who came from a strong experimental background, and theorists who worked in interpreting biological findings. During the four week course, students had the opportunity to engage with the faculty and guest lecturers. Finally, we had the chance in the final weeks to propose and work on individual projects under the guidance of faculty members.
Each of the four weeks had an overall theme ('Neural Coding', 'Dynamical Systems', 'Learning and Plasticity', 'Large-Scale Brain Models'), and talks were structured to develop and reflect multiple aspects of each theme. For the first two weeks, the typical format was a talk by one or two guest speakers in the morning, with time for discussion and with a strong degree of interaction between speaker and students. Afternoons and evenings were either spent working in groups on problem sets or with additional lectures or tutorials, with sufficient time for interaction with faculty and other students. During the last week and a half, in addition to morning lectures, time was devoted to individual projects which each student proposed and developed with the assistance of a faculty member and tutor, before presenting their project on the final day.
My specific aims in attending the course were to consolidate my knowledge and to place my own work and understanding within the broader framework of computational neuroscience. My own work deals with explaining the impact that structutal inhomogeneities have on dynamics, and I hoped to hear from people that were researching topics thought to be strongly related with my own, notably optimal network design and oscillations. In addition, I was also enthusiastic about hearing from the people who themselves made the steps forward, while meeting researchers with similar research interests. Both of these were fulfilled: I was able to meet researchers in a uniique environment, where the students were able to engage with faculty. Most importnantly I was able to deepen and complement my knowledge of several key themes in computational neuroscience, by hearing how researchers had tackled problems using alternative approaches, such as using techniques from information theory to understand grid cell coding.
For my individual project, I chose to implement and extend a model of spike timing-dependent plasticity (STDP). This was not related to my normal PhD work, but I wanted to investigate something that we had learned during the lecture and to take advantage of the presence of a faculty member who initially published the model.
I would like to thank the Directors of the MCN for putting together such a challenging but rewarding programme, and the INCF for the travel grant which assisted me to participate.
Man Yi Yim (Bernstein Center Freiburg, Germany)
Attend the FENS/IBRO Summer School “Cognition and action: Systems neuroscience approaches to understanding complex behaviour” on June 05-11, 2010 in Dubrovnik, Croatia
The main aim to participate in this FENS/IBRO summer school was to learn more about system neuroscience in relation to cognition and action from diverse approaches, including imaging, pharmacology, computational modeling and neurophysiology. In the poster session, I presented a poster on my work and discussed with other scientists to learn more about the ongoing research in the field. Last but not least, I was keen on meeting researchers with the same research interest and building up collaborations when interests matched.
Each day of the course started with lectures by faculty with time for discussion. Afternoons were open, allowing students to interact with faculty and discuss papers in more detail as well as working on group projects related to the lectures. Additional faculty lectures were held in the late afternoon, followed by dinner and evening poster sessions, in which students presented and discussed their work with the faculty. On the last day, students presented their work on the projects associated with the faculty's expertise.
This was a very fruitful summer school, in terms of both scientific knowledge and discussions with researchers. I learned diverse aspects of system neuroscience from the presentations by faculty. As a PhD student working on spiking neural network of the striatum, I was particularly interested in the experimental findings and the proposed functions of the striatum. Several speakers talked about the functional segregations of the dorsal and the ventral striatum, the direct and the indirect pathways in the basal ganglia and the effects of dopamine on cognitive tasks. Besides, I was fascinated by the modeling approaches using probabilistic learning and Bayesian inference, which account for the stochastic nature of the neuronal activity and the uncertainty in our perception and learning. In my poster session, I discussed my work with many researchers and got feedback from them. One of the faculty kindly provide me with some neuronal data of the striatum which will be very useful for my future work. During the school, I discussed with faculty and students who are working on the striatum and basal ganglia, and I am looking forward to the exchange of ideas and collaborations in the future. As for the student project, my group had enjoyable and fruitful time discussing the papers and talking with the professors in the field of probabilistic perception and probabilistic learning. On the last day I presented two models for the group project, namely the probabilistic population codes and the Sampling based representation. I definitely would like to learn more about these kinds of theoretical and computational modeling in the future.
Donders Inst. for Brain, Cognition and Behavior; University Medical Centre St. Radboud; Nijmegen, The Netherlands.
The aim of the 9 day visit in Nijmegen, 5-14 May 2010, was to:
1.Create mechanism allowing for sharing data between Scalable Brain Atlas and 3d Brain Atlas Reconstructor by developing suitable converters and overcoming technical difficulties.
2.Discuss possibilities of using integrated tools in long term research projects.
The visit was very fruitful as both of the goals were accomplished:
1.Creating bidirectional data converters: Scalable Brain Atlas can be now supported with data generated by 3d Brain Atlas Reconstructor. Moreover, it is possible to convert any atlas available on scalablebrainatlas.incf.org/ into three-dimensional model using 3d Brain Atlas Reconstructor. Because the 3dBAR uses data downloaded directly from Scalable Brain Atlas, it is assured that 3D models are always generated using the most recent set of slices.
2.Extending 3d Brain Atlas Reconstructor data storage format by new metadata elements (spot labels, comment labels, etc.) and adding mechanism for detecting unwanted identical structures. Both improvements simplify reviewing reconstructed models and spotting potential errors.
3.Creating a parser for colour-coded bitmaps: Parser allows to convert bitmaps (in which structures are stored in areas with uniform colour) to 3dBAR file format and then create 3D model.
Long term cooperation:
With enhanced features of 3d Brain Atlas Reconstructor and Scalable Brain Atlas it was possible to define and initiate long term project involving both tools. The purpose of the project is to transform three-dimensional models of segmented structures from Paxinos and Franklim mouse brain atlas into Waxholm space and then take coronal slices from the warped geometries. In order to achieve this goal regular on-line meetings are planned within the next months. We hope that a (preliminary) working version would be demonstrated as a demo during the 3rd INCF Congress in Kobe.
Ben Torben-Nielsen, Theoretical and Experimental Neuroscience Unit, Okinawa Institute of Science and Technology, Japan
Volker Steuber. Biocomputation group, Science and Technology Research Institute, University of Hertfordshire, UK
Kick-off a collaboration with the group headed by Dr. Steuber at the University of Hertfordshire. The project focusses on the the morphological specification of neurons to perform a particular function (here pattern recognition). Hence, its relation to INCF's 'Ontologies of Neural Structures' program.
- Delivered specialized, custom built software for this project.
- Implemented project-specific parts of the software. Obtained initial results.
- Discussed further actions to be undertaken by project partners in Okinawa and Hatfield.
Computational Neuroscience Group, Norwegian University of Life Sciences, As, Norway
The main goal of the visit was to work together on two scientific projects.
1) I wanted to learn about the software tools (LFPy) used to study extracellular potentials of model neurons. This software is developed by the Computational Neuroscience Group.
2) We also planned to work on a Matlab GUI tool for the "inverse CSD" method of analyzing LFP data from multi-electrode arrays.
The second goal was to prepare a grant application for a longer-term visit.
1) I have talked to the developers of LFPy and learned how LFPy can be used
a) to create and control NEURON simulations of morphologically reconstructed neurons,
b) to gather the voltage and trans-membrane currents data, and
c) to analyze the extracellular field.
We discussed applicability of these tools to my research projects.
2) We wrote the first version of a simple GUI tool which allows to visualize 2-dimensional LFP data, and to apply different variants of current source density analysis.
We will further work on this tool; we hope it will be useful for the neuroscience community, as the multielectrode arrays become more popular and the 2-dimensional recordings of LFP become common. We intend to make the software available in near future through the INCF software center.
3) We prepared a grant application for a 10-month research stay in Norway. The application will be submitted to the Research Council of Norway in February 2010.
4) Additionaly, the visit was an opportunity to present my recent research to the members of the Computational Neuroscience Group (a talk entitled 'Extracting functional components of neural dynamics with Independent Component Analysis and inverse Current Source Density').