BioNEt/Institute of Neuroscience Workshop

Wednesday-Thursday 18-19 January 2006, Room 2.22, Research Beehive, Old Library Building, University of Newcastle. Attendance, refreshments and lunch are free. Register for event by completing the online registration form. Limited to 60 participants.

Integrative Neuroimaging:
Computing Challenges & Solutions

Programme

Day 1: Wednesday 18 January

10.30 Registration & Tea/Coffee

10.55 Welcome and Introductions

11.00 Lessons learnt in another e-science domain
Bob Mann, Astronomy & National e-Science Centre, University of Edinburgh
11.45 Title to be confirmed
Kai Alter, Neurology, Neurobiology & Psychiatry, University of Newcastle

12.30 Lunch

1.30 Challenges in neuroimaging computing: dealing with large datasets and dynamics
Gary Green, York Neuroimaging Centre, University of York
2.15 Neuroimaging and GRID computing: applications in pharmaceutical research
Goutham Edula, AstraZeneca R&D Lund
3.00 fMRI of macaque pushes the technology to its limits
Li Sun, Biology & Psychology, University of Newcastle

3.45 Tea/Coffee & Group Discussion
The following topics have been identified as pervading themes for registrants: (a) multimodal imaging techniques and image coregistration; (b) image data integration, management and analysis; (c) problems in human and macaque fMRI data processing, and (d) multiparametric fMRI: software and hardware issues. This session will discuss and add to these issues, and identify specific questions to be addressed during the breakout sessions on day 2.

4.30 Inference and the brain
Karl Friston, Institute of Neurology, University College London

7.00 Workshop Dinner - Details to be confirmed

Day 2: Thursday 19 January

9.00 Tea/Coffee

9.30 Dynamic causal modelling
Karl Friston, Institute of Neurology, University College London
10.15 Hardware and algorithms for simultaneous PET and MR Imaging
Richard Ansorge, BioPhysics Group, University of Cambridge

11.00 Tea/Coffee

11.30 Multimodal imaging using fMRI and MEG - Evoked/Induced oscillatory dynamics and the bold response
Krish Singh, Cardiff University Brain & Repair Imaging Centre (CUBRIC), Psychology, University of Cardiff
12.15 Registration for Breakout Sessions

12.30 Lunch

1.30 Breakout Session 1

2.15 Break - Change Sessions

2.30 Breakout Session 2

3.15 Tea/Coffee

3.45 The CARMEN neuroinformatics repository
Paul Watson, NERESC, University of Newcastle

Supported by BioNEt, CELS Ltd, European Regional Development Fund and Institute of Neuroscience

Abstract

Human, primate and small animal neuroimaging produces increasingly voluminous data, which is captured, registered, analysed and archived in a diverse range of structures and formats. And, specific to particular imaging technologies, e.g. fMRI, we have recently seen the evolution of organised resource interchange mechanisms, based on GRID computing, which are allowing neuroscientists to share data, software and equipment, and compute across economical, ubiquitous parallel computers developed by the UK e-Science Programme.

For researchers interested in imaging populations of neurones, and analysing their large scale neuroimage datasets, this has provided considerable scope for:

• Development of new methodologies to enhance algorithmic image registration, processing, analysis and comparison; taking advantage of functionally distributed computing.
• Formation and maintenance of very large scale data collections, making use of existing (and in some cases remote) date and equipment; facilitating a more significant understanding of inherent processes and trends, and providing opportunities for correlative analysis with other relevant datasets, e.g. genomics and population data.
• Propagation and enablement of diverse research community addressing a shared set of goals and challenges.

At the forefront of this research, neuroscientists wish to develop methods that can support and simplify quantitative integration of data produced by fMRI studies, and a range of complementary imaging techniques (EEG, MEG, PET etc), which augment fMRI by allowing the researcher to investigate known phenomena at a variety of levels of scale. As a result of the disparity of these techniques, experiments tend to produce data that is both extremely voluminous, and heterogeneous.

To extract best value, and optimum insight from this data, it is necessary to develop stable and adaptable computational methods that can support data integration, maximise utilisation of the computation available to the data, and underpin the development of generic tools. These objectives are the focus of the “Computing with Integrative Neuroimaging” workshop. The following topics will be covered:

• Current state of integrative neuroimaging in the UK, and identification of problems that can be addressed through development of new mathematics, software and data management strategies.
• Case studies profiling the way in which GRID computing and related computational technologies have been applied to neuroimage data by contemporary projects.
• Software technologies that have been developed within and outwith the neuroimaging community to integrate and analyse heterogeneous image data.
• Global features and implications of the “electronic laboratory environment”.

Over the course of two days discussion, a roadmap will be developed that will enable to the assembled community to develop and exploit new research opportunities.

© 2005 BioNEt - e-mail our webmaster