Pages

April 12, 2009

Connectomics Definition, History and Progress to Map of the human Brain


Brain mapping process and work from a 2008 research paper. High-resolution T1 weighted and diffusion spectrum MRI (DSI) is acquired. 66 cortical regions with clear anatomical landmarks are created and then (3b) individually subdivided into small regions of interest (ROIs) resulting in 998 ROIs

The Economist magazine describes the history and progress of Connectomics. From Wikipedia, Connectomics is a novel, high-throughput application of neural imaging and histological techniques in order to increase the speed, efficiency, and resolution of maps of the multitude of neural connections in a nervous system. The principal focus of such a project is the brain. The map produced by such a project is called a connectome.

Connectomics actually started before the word existed. In 1972 Sydney Brenner, a biologist then at Cambridge University, decided to work out the connections of every cell in the nervous system of a small nematode worm called C. elegans. He picked this animal because its body has a mere 959 cells, of which 302 are nerve cells. It is also a hermaphrodite, fertilising itself to produce clones. That means individual worms are more or less identical.

Dr Brenner and his team embedded their worms in blocks of plastic, sliced the blocks thinly and then stained each slice so its features would show up in an electron microscope. The resulting pictures let the path taken by each nerve cell to be traced through the worm’s body. They also revealed the connections between cells. Over the course of 14 painstaking years the team managed to map the complete nervous system of C. elegans, work for which Dr Brenner, too, won a Nobel prize.

The scale of that work, though, hardly compares with today’s quests to map the brains of mice and fruit flies. The cerebral cortex—the part of a mammal’s brain that thinks—is composed of 2mm-long units called cortical columns. Winfried Denk of the Max Planck Institute for Medical Research in Heidelberg, Germany, estimates that it would take a graduate student (the workhorse of all academic laboratories) about 130,000 years to reconstruct the circuitry of such a column. But efforts to automate the process are gaining ground.

Dr Brenner’s method used what is known as a transmission electron microscope. In this the electrons that form the image pass through the sample, so the individual slices have to be prepared and examined. Dr Denk is speeding matters up by using a scanning electron microscope instead. This takes pictures of the surface of an object. Dr Denk (or, rather, his graduate students) are thus able to load the machine with a chunk of plasticised brain and a slicer. Once the microscope has taken a picture of the exposed surface of the chunk, the slicer peels away a layer 25 billionths of a metre thick, revealing a new surface for the next shot. The slice itself can be discarded, so the process is much faster than using a transmission microscope.

Researchers have also devised sneaky ways to tag parts of the brain that are of special interest, so that they can be followed more easily from slice to slice. Dr Denk, for example, tracks the myriad branches of a single nerve cell using an enzyme from horseradish. This gets stuck on the cell’s surface and then reacts with a stain that is added to the sample.

It is also possible to trace neural pathways from cell to cell. Ed Callaway at the Salk Institute in La Jolla, California, does so using rabies viruses. Rabies hops between nerve cells as it races to the brain, which is why even an infected bite on the ankle will eventually drive someone mad. That ability to leap the gap between cells means the connections branching from a single cell can be mapped.

Even when the images are in, however, making a map from them is another matter. Dr Brenner’s team traced each cell by eye—matching shapes through hundreds of cross-sections. Sebastian Seung, a computational neuroscientist at the Massachusetts Institute of Technology, is working on a program that will automate this process, too. It will allow a computer to learn how to match cells from one slice to another by trial and error, as a human would, but with the infinite patience that humans lack.

The result of all this effort, it is hoped, will be precise circuit-diagrams of brains. The first brains to be mapped will probably have belonged to mice. Besides being cheap and disposable, a mouse brain weighs half a gram and packs a mere 16m neurons. Human brains (1.4kg and 100 billion neurons) will come later, when all the wrinkles have been ironed out in rodents, and proper methods devised to analyse the results.







The first complete high-resolution map of the human cerebral cortex identifies a single network core that could be key to the workings of both hemispheres of the brain. Completed in 2008.

Researchers from Indiana University, University of Lausanne, Switzerland, Ecole Polytechnique Fédérale de Lausanne, Switzerland, and Harvard Medical School created the first complete high-resolution map of how millions of neural fibers in the human cerebral cortex in 2008 -- the outer layer of the brain responsible for higher level thinking -- connect and communicate. Their groundbreaking work identified a single network core, or hub, that may be key to the workings of both hemispheres of the brain.

a team of neuroimaging researchers led by Hagmann used state-of-the-art diffusion MRI technology, which is a non-invasive scanning technique that estimates fiber connection trajectories based on gradient maps of the diffusion of water molecules through brain tissue. A highly sensitive variant of the method, called diffusion spectrum imaging (DSI), can depict the orientation of multiple fibers that cross a single location. The study applies this technique to the entire human cortex, resulting in maps of millions of neural fibers running throughout this highly furrowed part of the brain.

Sporns then carried out a computational analysis trying to identify regions of the brain that played a more central role in the connectivity, serving as hubs in the cortical network. Surprisingly, these analyses revealed a single highly and densely connected structural core in the brain of all participants.

"We found that the core, the most central part of the brain, is in the medial posterior portion of the cortex, and it straddles both hemispheres," Sporns said. "This wasn't known before. Researchers have been interested in this part of the brain for other reasons. For example, when you're at rest, this area uses up a lot of metabolic energy, but until now it hasn't been clear why."



High-Resolution Connection Matrix, Network Layout and Connectivity Backbone

FURTHER READING
MIT Technology Review in 2008discussed connectomics and Jeff Lichtman of Harvards work creating brainbows.



BrainBows: Genetically engineering mice so that their brain cells express different combinations of fluorescent colors reveals the brain’s complicated anatomy. In the image round green neurons are interspersed with diffuse support cells called astrocytes. Credit: Jean Livet


The overall goal of the Connectome project [Initiative in Innovative Computing (IIC) at Harvard has the connectome project as one of several computing initiatives] is to map, store, analyze and visualize the actual neural circuitry of the peripheral and central nervous systems in experimental organisms, based ona very large number of images from high-resolution microscopy. The proposingteam from the Center for Brain Sciences has already demonstrated its capacityfor, and expertise in, high-throughput imaging. The critical challenges arecomputational, as the total number of voxels needed to establish the Connectomeis ~10^14. The principal challenges are to develop: (a) algorithmsfor efficient 3D segmentation circuit identification (b) the ability to transfer, storeand analyze 3D images in multi 100GB range; and (c) scalable database techniques to store, manageand query multi-TB, multi-modal datasets.

blog comments powered by Disqus