NIH Grant Supports Research Into Mapping the Brain With Machine Vision
For as much as we know about the human body, the inner workings of the brain remain elusive. This lack of understanding has precluded the development of a neuroprosthetic device, or brain implant, that is truly biocompatible with brain tissue. A multidisciplinary team of researchers led by Professor Badri Roysam has secured a $2.4 million grant from the National Institutes of Health (NIH) to develop a new, open source software toolkit for analyzing 3-D multichannel brain images — a critical step toward the development of a viable brain implant.
Such a permanent or long-term brain implant could benefit any number of individuals who have suffered brain trauma, but so far, researchers have been unsuccessful in developing neuroprosthetics that are not ultimately rejected and rendered useless by the surrounding brain tissue.
The multiyear project aims to arm biologists and other brain experts with the ability to better map, understand, and quantify the complex interactions between the components of brain tissue and their different functions, said Roysam, professor of electrical, computer, and systems engineering.
“One of the holy grails of this business is figuring out how to create a better neuroprosthetic device that is compatible with tissue and lasts longer. What makes this difficult is that brain tissue is, indisputably, the most complex tissue that we know of,” Roysam said. “We’re planning to look at the different types of cells that make up brain tissue in extraordinary detail, and then quantitatively map the interactions and relationships among these cells and neighboring blood vessels.”
While current optical microscopes are able to capture 3-D images of brain tissue and biochemical markers in brain tissue, current software tools are not able to analyze large volumes of data, requiring a researcher to manually look over each image. Roysam’s team is working to develop software that uses machine vision and artificial intelligence to automatically analyze these large, complex data sets, with the goal of mapping the structural and functional relationships and identifying interactions among brain tissue components.
Roysam said this next-generation image analysis tool, dubbed Fluorescence Association Rules for Multidimensional Insight, or FARSIGHT, will be specifically designed to meet the needs of tissue-scale biology and will accelerate the discovery and development of longer-lasting neuroprosthetic implants. The software also will benefit research involving the analysis of complex and dynamic microenvironments of cells, including stem-cell niches, stroke research, engineered tissue, and tumor microenvironments.
“Our goal is to provide the entire community of biologists, particularly neurobiologists, with software that doesn’t currently exist,” Roysam said. “Because researchers with different backgrounds and expertise will be using our software, we need to make sure that it’s easy to use.”
FARSIGHT will be open source, meaning that Roysam and his team will make available the software’s source code to the public, and encourage others to tweak the program and experiment with new applications. Roysam said FARSIGHT will be designed using multiplatform development tools so that the software will work on Windows, Linux, and Macintosh computing platforms.
“One of the holy grails of this business is figuring out how to create a better neuroprosthetic device that is compatible with tissue and lasts longer. What makes this difficult is that brain tissue is, indisputably, the most complex tissue that we know of,” Roysam said.
“Open source is an effective and efficient way to collaborate,” he said. “It can potentially lead to better software, because it continually benefits from peer review. We can leverage other people’s work, and other people can use our software as a launching pad for new innovations.”
Looking for associations between multiple sets of highly complex data requires massive computing power, and Roysam said the new research project will leverage Rensselaer’s newly established Computational Center for Nanotechnology Innovations (CCNI).
The supercomputer will significantly trim the time it takes the research team to develop FARSIGHT and assemble a new 3-D map of the brain.
“We would like to be able to map the whole brain in more detail than ever before,” he said. “This requires massive amounts of data that would have previously taken far too long to process. But with CCNI, we can do things that we couldn’t even think about doing before.”
Along with Roysam, researchers working on the project include William Shain of the New York State Department of Health’s Wadsworth Center; Sally Temple, adjunct professor of biomedical engineering at Rensselaer; professor Kevin Eliceiri from the University of Wisconsin; William Schroeder of the Clifton Park, N.Y., software firm Kitware Inc.; and tumor biologist Rakesh Jain from Harvard Medical School at Massachusetts General Hospital. Roysam said his team of graduate students, undergraduates, and postdoctoral researchers also continue to make important contributions toward the FARSIGHT project.
SOURCE: Rensselaer Polytechnic Institute
CONTACT:
Inside Rensselaer
Strategic Communications and External Relations
1000 Troy Building
110 Eighth Street
Troy, N.Y. 12180
leibat@rpi.edu.
###