The researchers will develop computer software that will allow medical researchers and physicians around the world to share information and images online. The images will assist them in better understanding brain disorders such as Alzheimer's and autism.
"Rutgers is very proud to direct this collaboration by outstanding information scientists, mathematicians and psychologists to further medical research," said Joseph J. Seneca, university vice president for academic affairs. "This project focuses high-powered information analysis with behavioral science in creative ways that ultimately can improve the quality of life."
Funded by the NSF's Cross-Cutting Initiative in Information Technology Research, the Rutgers project is headed by Paul Kantor, professor of library and information studies science at Rutgers' School of Communication, Information and Library Studies (SCILS) in New Brunswick and Stephen Hanson, associate professor of psychology and co-director of the Advanced Imaging Center at Rutgers-Newark.
To study brain disorders, doctors can take pictures of patients' brains using a highly sophisticated instrument called functional magnetic resonance imaging (fMRI).
The technique produces millions of bytes of data that are transformed into moving images, allowing doctors to see how brain signals move across the brain.
"Scientists would greatly benefit from accessing these images, to look for patterns that might provide clues about the differences between a diseased brain and a normal brain," Kantor said.
Currently, however, the data are indexed by little more than the patient's vital statistics. "If we have better ways of indexing and retrieving images, it will be helpful to not only researchers but also physicians," he said.
The Rutgers project aims to solve this problem through a collaboration of information scientists, cognitive scientists and computer scientists and engineers. Years from now, the new software will allow a neurologist to take dynamic brain scans, tap into the online archives to find patients around the world with similar images, and pull up the case records of those patients, Kantor said. Their work will combine cutting-edge mathematical and computational analysis with a file-sharing system at Rutgers called "RUMBA." Similar to Napster, which allows the sharing of data through the Internet, this system enables physicians and scientists to maintain their own data files at hospitals and laboratories, yet share them online.
The new search tools that will be developed are based on the idea of dynamic brain activity. When something happens in the brain, signals go from place to place, creating a pathway through neurons that can be seen by brain imaging techniques such as fMRI.
This activation is like a spreading vine that expands through the brain over time, Kantor explained. But the underlying signals are hidden by the complex chemistry of the brain, much as the leaves of a vine hide the stalks.
The Rutgers research aims to find the vine amidst the leaves, Kantor said. This requires filtering away the secondary signals and "noise," to see the original path of activation. A second component of the research could be called "vine matching." This is the mathematical problem of figuring out when one vinelike pattern is similar to another.
Other co-principal investigators on the project include Deborah Silver, associate professor of electrical and computer engineering, Rutgers-New Brunswick; Ben Martin Bly, assistant professor of psychology, Rutgers-Newark; Sven Dickinson, associate professor of computer science, University of Toronto; Lawrence Shepp, professor of statistics, Rutgers-New Brunswick; and Jonathan Cohen, professor of psychology and director of the Center for the Study of Brain, Mind and Behavior, Princeton University.
The brain image indexing project originated in the Rutgers Distributed Laboratory for Digital Libraries at SCILS, which grew out of Rutgers' strategic program to stimulate new research. That program, called the Strategic Reallocation Opportunity Analysis, has brought a tenfold return on investment to the university.
The project was also jump-started by a pilot grant from Rutgers' Information Sciences and Technology Council, headed by Saul Amarel, the Alan M. Turing emeritus professor of computer science, and Casimir Kulikowski, the Board of Governors Professor of Computer Science.
Paul Kantor, 732-932-7500, extension 8216, or 732-545-7939;
Stephen J. Hanson, 973-353-5440, extension 5095.