Understanding how the brain generates the thoughts, feelings and impulses that make up human consciousness will require a reliable brain-wiring diagram of almost unimaginable complexity. Scientists at the Society for Neuroscience meeting discussed emerging technologies that suggest a world of possibility for such brain mapping.
A clear picture of how neural networks link to one another and coordinate mental activity amounted to science fiction until recent years, and it still remains decades from fruition. But researchers are optimistic.
“I marvel at the power of the arsenal of tools our field is creating,” H. Sebastian Seung, a professor of computational neuroscience at the Massachusetts Institute of Technology, said in a lecture titled “The Once and Future Science of Neural Networks.”
“In the future scientists will create vast databases telling us how neural networks are created, what signals they send to each other, what genes they express. Our possession of these tools allow us to dream today of what I never could have imagined in my youth.”
H. Sebastian Seung envisions using powerful computers to map and investigate the brain in new ways. (Photos courtesy of SfN)
The vast amounts of data will produce a new field Seung calls connectomics, which will involve using powerful computers to chart and analyze the brain.
“Analyzing that information will be one of the greatest computational challenges of all time,” Seung said.
One way of coping with this flood of data will be through neuroinformatics, a relatively new field that uses computers for mining and sharing data, said Mark Ellisman, a professor of neuroscience and bioengineering at the University of California at San Diego, who delivered a lecture on “Brain Research in the Digital Age.”
Ellisman envisions the creation of a “Whole Brain Catalog” that would consist of a massive computer database, allowing scientists to deposit information they accumulate about the brain into a centralized bank from which all scientists could draw. “The storage of data is not the problem,” he said. “Managing the data is the problem.”
Neuroinformatics will play a pivotal role in accelerating progress in neuroscience in the 21st century, Ellisman said. The National Institutes of Health, he said, is developing the Biomedical Informatics Research Network (BIRN) to enable researchers to exchange data on animal models of multiple sclerosis, Alzheimer’s disease and Parkinson’s disease. And the Cell-Centered Database (CCDB) is providing scaffolding for knowledge about molecular processes, organelles and other sub-microscopic processes in the brain.
Detecting Connection Patterns
New imaging techniques will help create a detailed wiring diagram of the brain, scientists agree. A type of magnetic resonance imaging called diffusion tensor imaging uses the movement of water in the brain to infer the location of tracts of fibers called axons that link various regions, provides vivid insights into how the brain is wired, said Heidi Johansen-Berg of the Center for Functional Magnetic Imaging of the Brain at the University of Oxford.
Until recently such wiring patterns were discovered only by dissecting brains, or by performing invasive tracer studies in living animals. Diffusion tensor imaging allows researchers to see the living human brain in action noninvasively.
For example, in the ventricles of the brain, which are filled with cerebrospinal fluid, the water diffuses evenly in every direction, but in the presence of the myelin-coated axons, water tends to flow parallel to these fibers. By measuring this motion, it becomes possible to map those tracts, thereby revealing how brain regions are connected.
“The changes in connection patterns also hint at boundaries between regions,” Johansen-Berg said in her lecture, “Imaging Human Brain Connections.”
Johansen-Berg cited as an example the work of researchers who have been able to divide the thalamus, in the center of the brain, into regions that are not visible on magnetic resonance imaging scans, and then trace how those various regions connect to the motor cortex, the prefrontal cortex, the visual system and other areas. Such wiring diagrams provide invaluable information about how the brain modulates its functions.
“By making use of connectivity data, we can determine structures not previously available to us,” she said.
Karel Svoboda of the Howard Hughes Medical Institute at Janelia Farm Research Campus in Virginia has been using technology to reveal the workings of synapses, which are the communication links among neurons. This knowledge is essential for understanding how the brain changes in response to experience.
Using a technique known as two-photon laser scanning microscopy, Svoboda has produced time-lapse images that reveal how synapses communicate.
Karel Svoboda of the Howard Hughes Medical Institute is employing a technique called two-photon laser scanning microscopy to study how synapses communicate. (Courtesy of SfN)
“With its high resolution, it is possible to image individual synapses,” he said during his lecture, “Imaging Synapses in Their Habitat.”
“It’s very specific—it can measure the distribution and trafficking of molecules. It allows us to detect stable versus plastic synapses.”
Although most synapses are stable, some persisting for a lifetime, others change vigorously in response to experience, Svoboda said.
“At the fast end of the spectrum, neurotransmitter release can be modulated in as little as a few milliseconds, a phenomenon that is thought to contribute to short-term memory,” Svoboda said. “On the slow end of the spectrum, changes in synaptic strength are believed to encode long-term memories, implying that the structure and function of some synapses might be maintained over years.”
Despite such advances, communication within the brain remains largely a mystery, Svoboda noted.
“We really have no idea what happens at synapses when they change,” he said. “Imaging individual synapses during plasticity would be a huge step forward, and would remove many ambiguities.”
Mriganka Sur of the Picower Institute for Learning and Memory at MIT has used another imaging technique, multi-photon laser scanning microscopy, to obtain images of the function and structure of individual neurons, their synapses, and the supporting cells known as astrocytes, whose function remains tantalizingly unknown.
“We have found that neurons in the visual cortex are clustered according to their ability to detect different properties, such as the vertical edge of an object, or the horizontal edge,” said Sur. “This sensitivity to the orientation of the edge of light that defines an object is, at its most basic level, how we see all forms and shapes.”
In his book On Intelligence, Jeffrey Hawkins, developer of the PalmPilot and the Treo “smartphone,” asserts that the brain is a pattern-recognition device with memory capacity that is dedicated to making predictions about the environment. Hawkins argued in his presentation at the meeting that such a view of intelligence applies to artificial as well as human intelligence, and he studies neuroscience, he said, in an effort to understand how to make computers smarter.
Seung turned that idea around.
“I say, what can computer science do for neuroscience?” he asked during his lecture. “Philosophers love to ponder whether the human brain is complex enough to understand itself, but the new question is different: When will the artifacts developed by man become sophisticated enough to understand the human brain? I do not know when this will happen, but I am sure it will eventually come to pass.”