neuGRID

The aim of neuGRID is to deploy e-Infrastructures (EC-funded MammoGrid and AddNeuroMed's NeuBase) to enable the European neuroscience community to carry out research required for the pressing study of degenerative brain diseases and to build a new user-friendly Grid-based research e-Infrastructure (neuGRID), where the collection/archiving of large amounts of imaging data is paired with computationally intensive data analyses.

Project aims

neuGRID aims to become the "Google for Brain Imaging", providing a centrally-managed, easy-to-use set of tools with which scientists can perform analyses and collaborate.

 

Benefits of the research project

The project will initially provide neuroscientists with the capability to identify neurodegenerative disease markers through the analysis of 3D magnetic resonance brain images via the provision of sets of distributed medical and Grid services and later to provide a general, expandable infrastructure of services for other medical applications.

neuGRID will be compliant with acknowledged EU and international standards regarding data collection, data management, and Grid construction. Of the two deployed infrastructures, MammoGrid will provide knowledge related to the middleware and upperware that will allow applications to talk to the Grid, and AddNeuroMed that relates to the collection/archiving/retrieval of multicentre clinical data, biomedical images and computerized image analysis.

Key research challenges

Key research challenges will be the:

  • gridification of algorithm pipelines for brain image analysis
  • development of a mid-layer of services between user-facing and grid-facing services to make the infrastructure expandable to a number of algorithm pipelines
  • testing and validation of the prototype infrastructure

Project background

Brain imaging is currently one of the fastest expanding areas of biomedical interest for its role as a marker of brain diseases. Whereas the accessibility of other organs has facilitated the development of biomarkers (eg diagnosis and monitoring of a cancer through specific proteins present in the bloodstream), the brain is an exquisitely delicate organ and nature has thought fit to protect it tightly from outside influences – even those coming from the body itself.

Indeed, the whole metabolic brain machinery whereby the brain produces energy for its own functioning, structural molecules, and neurotransmitters, is protected from the rest of the body metabolism by the blood-brain-barrier, a sort of thin hole filter allowing only few small and selected molecules to pass from the bloodstream to the brain and vice versa. Not only the brain is protected from chemical insults by the blood-brain-barrier, but also from physical insults by the thick bone of the skull – the hardest and most shock resistant bone in the human body.

Development of brain disease markers

For these reasons, the development of brain disease markers has lagged behind that of other body organ diseases. The case of prostate is paradigmatic, where high amounts of a protein called Prostate Specific Antigen (PSA) can be readily detected in the blood and are used for screening and early diagnosis of prostate cancer and to test the efficacy of antineoplastic treatments.

Other brain diseases are unfortunately much more prevalent than multiple sclerosis for which presently we have no satisfactory treatment, such as Alzheimer's disease. Once believed to be an inescapable companion of older age, in the past 10 years we have learned that it is caused by "poisoning" of the brain by a small protein (Abeta). The increased pathophysiological knowledge has allowed developing antiamyloid drugs that are now in the pre-marketing phase of human testing. Early diagnosis and development of effective disease-modifying drugs would be greatly facilitated by the availability of an accurate (sensitive and specific cross sectionally and prospectively) disease marker.

How are computerized algorithms presently exploited for the study of the brain?

Current paradigm

The paradigm today is that of scientists physically migrating to imaging centres where they can find expertise and computational facilities carrying small personal datasets (a few hundreds of images at most). This is the case of image research partners of the present application (CO1 FBF and P5 VUmc) but also of countless other scientists. 

Typically, a research fellow or post doctorate spends three months or more at an image analysis centre – often thousands of miles from home – where he/she learns the use of the algorithms on personal image data, then returns to the original research group, where he/she can install all or part of the procedure and run jobs either in house or remotely on the image analysis centre servers or mainframes.

Without neuGrid

The combination of larger datasets and larger scientific community will make the present way of doing science unsustainable in the near future and make research environments necessary with archiving as well as computational facilities.

e-Infrastructures consisting of large repositories paired with high computational power to run computationally intensive algorithms on biomedical images and clinical data are hardly available.

Biomedical Informatics Research Network

The BIRN – Biomedical Informatics Research Network, funded by the US National Institutes of Health and other US federal agencies, is probably the effort most closely related to neuGRID. It aims to enable a software 'fabric' for seamless and secure federation of medical data across the network of participating centres and facilitate the collaborative use of domain tools and flexible processing/analysis frameworks for the study of brain diseases.

The BIRN architecture is Grid-based and made of about 20 Grid hosts around the US. Interestingly, the BIRN research environment has not yet had a huge impact on the neuroscience community despite having been in place for the past 6 years. The BIRN is involved in the management and maintenance of the North American ADNI.

Back to top