High-Technology Advances Spur Progress In Study Of Human Brain

Augmenting old devices and procedures with the latest computer-based techniques yields new opportunities for today's neuroscientist ST. LOUIS--In 1984, five fighter pilots spent three days hooked up to one of the world's most sophisticated machines for probing the brain's electrical impulses. But it was only last month that San Francisco neuroscientist Alan Gevins presented his results from that experiment. The project, hailed by colleagues as a synthesis of various research techniques, was an attempt to study cognition by combining mathematical modeling, computer graphics, signal processing, magnetic resonance imaging (MRI), and electroencephalography, or EEG recording. "His work represents the direction that brain science is going, and he represents the forefront," say Richard Dasheiff, director of the University of Pittsburgh Epilepsy Center. Yet the fact that it took Gevins six years is also an indication of how much further the field needs to advance to make efficient use of the available technologies. Leaders in neuroscience say technological advances over the past decade--especially in molecular and imaging approaches to the brain--have fueled an explosion in knowledge about the nervous system. But they lament the lack of adequate means for analyzing and putting together data generated through these technologies. A common complaint is that all these new findings have yet to yield a coherent picture of how the brain works and thinks. "The mathematical tools to understand this higher level of information don't exist," says Dasheiff. Gevins has spent almost 20 years creating tools that will let neuroscientists understand this information better. He presented his results and his technology as a special presidential lecturer at the 1990 meeting of the Society for Neuroscience held here earlier this month. More than 1,000 scientists heard Gevins, the head of EEG Systems Laboratory in San Francisco, describe how he monitored changes in the brain activity of these pilots. What seemed of special interest to those attending was how Gevins stretched existing technologies to their limits in producing his results. In interpreting the data, Gevins integrated methods from many fields to tackle the complexities of how the brain works, a difficult, but necessary, approach to studying higher brain functions, such as memory. "People have begun to look at smaller and smaller bits [of the system]," says Carlton Hunt, a neurobiologist at Washington University in St. Louis. "Sometimes, the integrative things get lost." But, Hunt adds, "Every once in a while some maverick will go in a totally different direction and open up things that nobody has looked at." Indeed, Gevins may be well positioned to be that sort of maverick, given his impressive background and his multidisciplinary research team. "In [studies of] the central nervous system, you need equipment and expertise from many fields," says Vernon Mountcastle, a neurobiologist at Johns Hopkins University. Gevins has received training in engineering, mathematics, and cognitive science, and his lab team includes a neuropsychologist, a cognitive psychologist, a neurosphysiologist, a mathematician, a physicist, computer programmers, and image processing, mechanical, and electrical engineers. Gevins is applying this mix of talents to revitalize EEG, an old technique for measuring electrical activity that is usually done with just a few electrodes. He increased the number of electrodes and harnessed the computer to analyze the data. Each of his pilot subjects was hooked up to 124 electrodes and spent 10 to 14 hours performing tedious mental tasks. He then used computer software to reconstruct the pilots' heads from MRI scans so he could track the path of the electrical impulses that reflected brain activity. And, although MRI is becoming a common tool in medicine, rarely has anyone tried to use it to re-create three-dimensional pictures. Nor has it been used in conjunction with EEG. Finally, to deal with all the data from the EEG and MRI, computer scientists at EEG Systems had to write a lot of complicated software. The use of 124 leads provides a "finer degree of spatial sampling," says Gevins. This sampling increases exponentially the amount of data generated and the degree of resolution; it also adds to the complexity of the data to be analyzed. That analysis requires sophisticated signal processing so that the weak signal generated by thought can be distinguished from stronger background signals, he says. Each of his experiments takes years to set up and analyze, says Gevins, who through the years has developed ever more sophisticated technology. In eight experiments he has looked at different aspects of cognition and brain function. Five have been written up in Science. Three, including the one involving the fighter pilots, were first reported at this fall's neuroscience meeting. Once the experiment was over, Gevins spent months figuring out how to filter out noise and how to compensate for "blurring" that occurs as the electrical signal passes through the skull. "That's a tough problem," he says. It also takes time to turn the data from the MRI scan into a three-dimensional model of the brain, and then to superimpose the EEG data. He is working on his fifth generation of software development and now calls his program MANSCAN (Mental Activity Network Scanner). The results are depicted in a series of images showing shifting networks of nerve cell activity. Arrows on these images represent thought processing. "The complexity is so great that the ability to communicate it [to colleagues] is very, very difficult," notes Gevins. As an example, he cites a presentation made five years ago in which he lacked three-dimensional color graphics to illustrate his results. His audience applauded politely. Three years later he presented the data again, this time in full color to enhance the visual representations of the data. "The people were standing on their chairs and congratulating me," he recalls. "He's getting to the area that we've always missed--thinking processes," says Richard Nakamura, chief of cognitive neuroscience at the National Institutes of Health's National Institute of Mental Health (NIMH). "His data tend to be highly processed, but they are exactly the sorts of things we need to know." At the same time, it's not yet clear what role EEG will play. EEG lets a scientist follow electrical impulses across the surface of the brain and observe changes over milliseconds of time. But EEG cannot probe deeper regions of this complex organ. In contrast, positron emission tomography (PET) lets one observe blood flow or metabolism in any part of the brain. Right now the neuroscience community seems to favor PET and another technology for measuring magnetic fields in the brain known as MEG. EEG needs to be better understood for it to catch on as a neuroscience research tool, says Nakamura. "Scientists have not had enough time to figure out how he is doing that [analyzing his data]," he says, "and until that's done, people are going to remain partially skeptical. People have to replicate the results." But Gevins is undaunted by such skepticism. He has started a business venture, SAM Technology, that will work to commercialize the lab's technology. Four grants awarded under the NIH and NIMH Small Business Innovative Research program have helped him to get started on turning his software into an off-the-shelf package that other scientists can use. And interest in his work seems to be growing. Investigators like Steven Petersen, a Washington University neurobiologist who uses PET, sees a role for both EEG and PET. "By integrating information from these different kinds of methodologies, you can help each other out," he explains. "The information we get [is] in ways complementary." But such integration takes money, Gevins complains. He had obtained a five-year NIH grant to do an experiment using various modalities, including MEG and PET, but yearly administrative cuts forced him to delete the PET portion of the experiment. Nakamura predicts that integration will someday occur. "What he [Gevins] was saying fits into the prejudices of many scientists [about how the brain works]," he says. "He's gotten the attention of a lot of people. A lot will be looking into that approach more carefully."

High-Technology Advances Spur Progress In Study Of Human Brain

Interested in reading more?

Become a Member of