Neuroscientists are mapping out a complete atlas of connectivity in the human brain, but what’s emerging is a battle of scales.Mapping the Brain’s Highways
Wide Angle / by / August 11, 2009
“The human brain has been terra incognita for as long as we’ve known it,” says Olaf Sporns, a professor of neuroscience at Indiana University. In 2005, Sporns co-authored a paper attributing the large-scale shortcomings of comprehensive neuroscience research to a lack of a foundational, anatomical description of the brain: In order to properly navigate this “unknown land,” he said, we must first draw a map.
Sporns proposed calling this map the “connectome.” As a thorough atlas of the connections in the brain, the name deliberately conjures associations with the enormously successful human genome map that had been sequenced two years prior. Now, four years after Sporns’ initial paper, the National Institutes of Health Blueprint for Neuroscience Research is launching the $30 million Human Connectome Project (HCP) in hopes of creating a comprehensive map of a healthy adult brain by 2015.
The human brain is a composite organ, divided up into several hundred small areas with highly specialized functions. Viewed under a microscope, most of these centimeter-wide areas have visibly distinguishable cell patterns. Most importantly for the HCP, each of these areas is connected by millions of thread-like neuronal projections called axons that run together in parallel, winding to form long, bundled structures resembling thick fiber-optic cables. “These are the major highways, and they’re these beautiful spatially organized structures,” says Van Wedeen, an associate professor of radiology at Harvard Medical School, where he works on developing new brain imaging technologies. “How does the brain function?” Wedeen asks. “It’s all processes, and these processes occur because of connections between these specialized areas.”
The rationale behind the Human Connectome Project’s focus on these fiber bundles is that the different brain areas are believed to acquire their functional characteristics based on how they are connected to each other. “It’s the inputs and the outputs of a neuron that determine what its function is,” Sporns says. Thus, like so much in biology, structure directly defines function—neurons in close proximity will process the same kinds of information, and connectivity between these regions can inform us on how the broader processes operate.
While these brain sections and their connection pathways are visibly distinguishable, the fact that the structures of the brain overlap in three-dimensional space has made them almost impossible to model until very recently. Traditional observational techniques require using microscopes to view ultra-thin slices of tissue—messy business when trying to reconstruct three-dimensional structures in something as thick as a human brain. New imaging methods allow scientists to noninvasively observe the live brains of their human patients in two ways: as they perform tasks in functional imaging machines, enabling scientists to see which brain regions show simultaneous activation and thus imply connectivity, and in diffusion imaging scanners that model the pathways of the fiber bundles by recording water flow along the gradients of the cables. By attempting to match the data obtained from the two modalities, the neuroscientists can combine the correlational data with the anatomical data to, little by little, fit together the pieces of the puzzle.
“The reason we haven’t had this data before for the human is because we haven’t had the tools,” says Dr. Michael Huerta, associate director of the neuroscience division of the National Institutes of Mental Health (NIMH). “We’re at a really sweet spot because the technologies that exist right now are really just on the cusp of being able to do this in a systematic, high-throughput way.”
And yet, there are other neuroscientists for whom “connectome” has a different meaning altogether. While the NIH project attempts to map the large-scale connections between brain regions, some scientists want to pry deeper: They aspire to eventually map the human brain neuron by neuron, constructing a full “wiring diagram” of our most complex tissue at its most fine-grained. In some ways, the broader region-to-region approach of scientists like Sporns and Wedeen was a response to the daunting and, right now, technologically impossible goal of mapping the trillions of neuronal connections in the human brain. But some hard-liners went another route—sticking with the ambition of a neuron-by-neuron map, they’ve merely shifted their gaze to simpler organisms.
Thus far, the only complete neuronal connectome that has been mapped is that of the nematode Caenorhabditis elegans. Clocking in at 302 neurons, it is one of the simplest model organisms to possess a nervous system. The human brain, in contrast, is comprised of nearly 100 billion neurons. Technologies to visualize neurons in live subjects—as well as process such gargantuan volumes of data—do not yet exist, so only post-mortem studies in simpler organisms are even presently imaginable.
Scientists like Joshua Sanes, a professor of molecular and cell biology at Harvard, insist that despite these technological hurdles, the true answer to the connectome puzzle lies at the level of the neuron. Along with Jeff Lichtman, also a molecular and cell biology professor at Harvard, Sanes has pioneered some of the most detailed neuronal connectivity data to date. The two are currently working on new technologies for visualizing synapses in live animals. “Think of the connectome as an old radio: If you find out the amplifier is connected to the tuner, it doesn’t tell you how a tuner works—it just tells you what the tuner is connected to,” Sanes argues. “I think the ultimate goal of the cellular approach is to find out how the brain works in the sort of circuit diagram sense.”
But neuroscientists such as Sporns, while praising the accomplishments of Sanes and Lichtman as admirable, question the necessity of a neuron-to-neuron connectome. “If you want to describe the economy or some other complex social pattern out there, do you really need to know what every shopper is buying in the supermarket?” Sporns asks. For Sporns, the true value of the human connectome lies in the aggregate behavior of its parts.
Nevertheless, both parties agree that eventually the microscale and macroscale approaches will converge as one enormous data set, hopefully with the different layers of complexity informing each other. For now, the NIH hopes to accomplish its five-year goal of a completed region-to-region map by encouraging collaboration among neuroscientists, many of whom have spent years working on the problem in isolation, leading to some unverified techniques and little cross-checking of data.
“Before there was the Human Genome Project, people were studying gene sequences—they were just doing it a million different ways in a million different labs,” Huerta says. “The impact of the Human Genome Project cannot be overstated, and I think the Human Connectome Project will have a similar, transformative impact on neuroscience.”
Mapping the human connectome will rely on a descriptive rather than empirical approach to research, revitalizing classic debates of the value of induction versus deduction in the sciences. Regardless, these ambitious scientists firmly believe a completed map will provide an indispensable foundation for all future neuroscience research. “Yes it’s descriptive, yes it’s a fishing expedition,” Sporns says. “But that’s how you catch fish.”
Not So Fast
Featured Blogger / by / August 11, 2009
The NIH aims to map the connectivity of the human brain in five years. But a definitive atlas of the brain will remain out of our grasp for a long time.
Several weeks ago, the National Institutes of Health announced the Human Connectome Project, an ambitious $30 million five-year initiative, which aims to map the connectivity of the human brain.
Is this feasible? In short, the answer is no. The idea that a complete microscale map of the brain at the level of single neurons and synapses can be achieved within five years is unrealistic.
Firstly, the numbers involved are just too vast. The human brain is an incredibly complex structure, consisting of hundreds of billions of cells, which between them form something in the order of one quadrillion connections (synapses).
Secondly, it is technically impossible, at least at the moment. Brainbow, for example, the ingenious fluorescent protein-based method for visualizing neuronal connectivity developed in the labs of Jeff Lichtman and Josh Sanes, involves creating genetically engineered animals, and so is not applicable to humans.
A macroscale map of the connections between brain regions also poses problems. There is, for example, no universally accepted scheme for delineating the functional subdivisions of the human cerebral cortex. There is still much debate about exactly how—or if—tractography data from recently developed neuroimaging techniques such as diffusion tensor imaging are correlated to anatomy.
What might be possible is a connectivity map at a scale somewhere between these two levels of organization. Even so, drawing up such a map for the entire brain is still an enormous task. When the Human Connectome Project comes to an end in five years’ time, we may have a comprehensive map of the connections within a few of the brain’s subnetworks, such as the thalamocortical system, which alone comprises several hundred discrete brain regions and thousands of fiber tracts.
Thus, the connectome project is likely to lead to a rudimentary first draft, which, like that of the human genome, will undergo numerous subsequent revisions. The initiative to map the connectivity of the entire brain should be thought of as an open-ended one, which will be added to and refined for decades to come.
Undoubtedly, a whole-brain connectivity map will be useful to researchers once it is eventually completed. But what such a map can tell us about how the brain actually works is likely to be limited. This is because the connectome apparently ignores the phenomenon of neuroplasticity.
Plasticity refers to the brain’s ability to physically alter its structure in response to experience. Far from being immalleable, as was once thought, the brain is a highly dynamic organ. Neurons can sprout new connections within minutes of a given stimulus, and entire neural pathways can be rerouted so that function is recovered after a brain injury.
The connectome also disregards the functional importance of neuroglial cells, another class of cells which are found in the nervous system and which outnumber neurons by at least 10 to 1. Once thought to merely provide structural and nutritional support for neurons, glia have, in recent years, come into their own as key players in the brain. As well as performing the roles initially ascribed to them, glia carry out a whole host of other vital functions, including monitoring neuronal health, identifying damaged neurons, and regulating synaptic plasticity. They are also known to be capable of communicating not only with one another, but also with neurons. A map of brain connectivity cannot therefore be complete without taking glia into account.
Finally, although the large-scale connections are very similar among individuals, there are significant variations at smaller scales. And the fact that the large-scale connections change over time—from infancy, to adolescence, through to adulthood—complicates matters even more. A detailed, definitive description of brain connectivity will, therefore, remain out of our grasp for a long time yet.
Tags:
No comments:
Post a Comment