[Paleopsych] New Scientist: The word: Connectome

Premise Checker checker at panix.com
Thu Nov 17 20:42:23 UTC 2005


The word: Connectome
http://www.newscientist.com/channel/opinion/mg18825251.000
>From issue 2525 of New Scientist magazine, 12 November 2005, page 62
[Happy to comply with your request for the full article, Brian.]

THE human brain is a fantastic maze of connections, a vast network of networks 
that circulates information and determines how we think and act. One of the 
many big puzzles left in neurology is working out which parts of the brain are 
connected - and how the networks function.

That's why top neuroscientist Olaf Sporns of Indiana University at Bloomington 
and his team are hoping for some lively debate about their new blueprint to map 
those connections. Sporns is calling it "the human connectome" after the 
billion-dollar human genome project, but it's bound to be far more 
sophisticated.

Why? Well, the genome is one-dimensional, while the connectome will be 
four-dimensional (three space, one time). The information needed to build the 
connectome is also far more elusive. Our brains contain roughly 10^11 neurons, 
with an estimated 10^14 possible connections. The magnitude of these numbers 
makes it impossible for the connectome to map the brain at the level of single 
neurons and synapses. Luckily, that may not be necessary because nerve cells 
tend to act in groups.

What will the connectome look like? At first, it will be a huge set of numbers 
from which cognitive patterns can be deduced. Input the coordinates of two 
brain regions and the connectome will give the probability of those two parts 
talking to one another. The coordinates refer to voxels (3D pixels), which is 
useful because it's how neuroscientists map the brain, so the connectome can be 
cross-referenced to other data. But in future, with the right technology, we 
could build a dynamic 4D model with the brain's connections operating in real 
time. Just think: one little thought, and the model would light up like a 
Christmas tree! "The brain is a maze of connections, a vast network of 
networks"

So how much is known already? There are a few precedents: large-scale 
connection patterns have been mapped for animal brains such as the macaque, cat 
and rat. But for human connections, researchers will need sophisticated imaging 
techniques that show how the brain's anatomy relates to its dynamic function. 
For that, they'll build on existing techniques such as diffusion tensor 
imaging, functional MRI and EEGs.

They'll also need the snazzy software used to map connections in the web and 
other large networks. Not to mention computers such as the giant 10 petaflop 
baby that Japan is planning.

The pay-off? Oh, just a few useful things such as finding out much more about 
how the different regions of the brain interact when we think or act, what 
injuries to particular regions do, and how our cognitive networks differ from 
those of other species.

Not surprisingly, the connectome will demand a worldwide effort by anatomists, 
brain imagers and computational scientists. And more billions than the genome 
project. Sporns thinks a first draft of the connectome could be ready in a few 
years. Watch this space.




More information about the paleopsych mailing list