New York, NY (May 17, 2004) -- Imagine a puzzle made up of one hundred billion pieces, each reacting to the other, and you have a glimpse of the enormity of the challenge facing researchers bent on understanding how brain cells work together to create human perception, thought, and action.

Every day, over 50,000 neuroscientists around the globe collect data on just these types of neural interactions, publishing their collected facts and figures in over 300 journals and scientific assemblies worldwide. But the sheer quantity and scope of neuroscientific data means that individual researchers cannot hope to utilize but a small fraction of what is available.--

Many experts -- including Dr. Daniel Gardner, a Weill Cornell Medical College Professor of Physiology and Biophysics, and Director of the College's Laboratory of Neuroinformatics -- now believe the time has come to give this community of scientists a better means of accessing -- and re-analyzing -- this vital data.

"We wanted to develop methods that could allow people to take the data behind these figures and put them in a database, and then to allow other people to download those data from it, and analyze it in other ways," explained Dr. Gardner, who is also Professor of Physiology and Biophysics in Neuroscience at Weill Cornell.

The result, he said, is neurodatabase.org, an online resource for researchers looking for the kind of detailed data that fuzzy reproductions of tables and graphs found in print journals simply cannot provide.

Writing this month in a special issue of Nature Neuroscience, dedicated to the 10th anniversary of the Human Brain Project (HBP), Dr. Gardner outlined those components that make neurodatabase.org such an important tool.

"We have just finished, within the last couple of months, a very sophisticated data upload tool that permits people at their home computers to automatically put the data into neurodatabase.org, and to annotate the data as well," he explained.

Annotation is key: Because much of what's stored on neurodatabase.org consists of raw data on signaling activity between brain cells, Dr. Gardner said researchers who upload their results will be asked to provide specific "metadata" -- "where in the brain it is, what type of neuron it is, what type of activity an animal was doing or was not doing" when the readings were recorded. It is this type of information that website users will use as keywords whenever they perform searches for specific data, Dr. Gardner said.

Once the right data is located, a tool built into the website called a "virtual oscilloscope" can help website users view extended datasets -- providing them with the full range of data produced by a particular study, not just the summary versions published in journals. These datasets are fully downloadable and ready for re-use in ways the original researchers might not ever have imagined.

"Our thought was that people need to be able to take brain data and subject it to additional sorts of analysis that could not be done by the home lab -- either because they weren't interested in a particular question or because they didn't have the mathematical systems or sophistication to do so," Dr. Gardner said.

Ideally, researchers will be able to approach a research problem from various angles -- looking at data first from the molecular level, then from a "big picture" view of brain activity as recorded by fMRI. "So perhaps the question you have can be partially answered by data from one database type, and partially answered by data from another," Dr. Gardner explained.

Neurodatabase.org operates on tools written in Java, he noted, "so they can be used by anyone with any contemporary computer -- Mac, PC, or Unix." And for Dr. Gardner, there's no question as to who owns the rights to uploaded data. "When you upload data to our website, there's an acknowledgment that you retain those rights," he said. "We firmly believe that intellectual property rights belong to the people who create the data."

The neurodatabase.org project is funded by a grant from the National Institutes of Health as part of the Human Brain Project (HBP), which celebrated its 10th anniversary this week.

Like the Human Genome Project -- which has already completed its main task of mapping the human genome -- the HBP seeks to crack another "code" crucial to our understanding of what it is to be human.

"Every single signal thatâs ever come into your body, everything you know or experience, comes in to the brain as what's called an 'action potential,' a change in the electrical potential of a nerve cell," Dr. Gardner explained. Varying frequencies of these action potentials make up what's known as the "neural code."

"We don't understand the code," Dr. Gardner said. "We do not know the correlation between the signal itself and the experience." Cracking the neural code is proving much, much more difficult than understanding and mapping DNA, he added, since there are billions of brain cells involved in trillions of complicated relationships every second.

Still, online tools like neurodatabase.org may give the global community of neuroscientists a new weapon in that struggle.

"It's tremendously exciting," Dr. Gardner said. "What we're doing is using computers to help understand some of the techniques for brain study -- but what we're doing, as well, is understanding how the brain works as a computer."

# # #

Office of Public Affairs
NewYork-Presbyterian Hospital/Weill Cornell
525 East 68th Street, Box 144
New York, NY 10021