May 5th, 2009
John Scalzi is one of the most popular SF bloggers on the Web, as well, of course, one of the most popular SF writers around. He’s also a tireless promoter of science fiction: not just his own, but everyone else’s. For several years he’s been running The Big Idea, a series of essays in which authors talk about the ideas behind their books. It’s proved so popular it will soon be its own spin-off website, but for now it’s still part of his blog, Whatever…and today he’s posted my 1,000 Big Idea essay about Terra Insegura.
Go check it out! And if you haven’t read Scalzi’s own books, you most definitely should.
Ah, the human brain. Seat of consciousness, miracle of creation or evolution (discuss amongst yourselves), able to jump to tall conclusions in a single bound, so incredibly complex that we’ll never be able to understand how it works.
Um, not so fast.
A year and a half ago, scientists at the Blue Brain Project in Switzerland announced they had successfully created an extremely detailed—down to the molecular level—model of the neurocortical column of a two-week-old rat…and that was just Phase 1 of their ambitious research effort aimed at nothing less than reverse-engineering the mammalian brain and recreating it in a computer.
The neurocortical column (NCC) is the basic unit of the neocortex, which in mammals is responsible for higher brain functions and thought.
“The thing about the neocortical column is that you can think of it as an isolated processor,” Blue Brain Project founder Henry Markram of the Brain and Mind Institute at the École Polytechnique (EPFL) in Lausanne, recently told BBC News. “It is very much the same from mouse to man—it gets a bit larger and a bit wider in humans, but the circuit diagram is very similar.”
At the Science Beyond Fiction conference in Prague last month Markram said the researchers are now taking their simulated NCC and inserting it into a “virtual reality agent”—a computer simulation of an animal. As the virtual animal moves around a virtual space, the researchers will observe the detailed activities in the column, and closely compare the results of their model with what is known about real NCCs in real rats, to ensure their model is accurate.
Once they’re satisfied with their model rat NCC, they can use it as a basic template on which to build models of NCCs from other species.
The computing power involved is, of course, immense. For its first phase, Blue Brain used an IBM Blue Gene supercomputer with more than 8,000 processors. It provided “only just enough” power: Markram estimates it can probably simulate about 50,000 fully complex neurons (brain cells) in close to real time. The rat NCC has about 10,000 neurons; at the upper end of the scale—that would be us—the NCC has 100,000 neurons.
And remember that the NCC is only the most basic element of the neocortex. The human neocortex has millions of NCCs. The hope is, however, that by creating an extremely detailed, accurate copy of a simple NCC at the molecular level, the researchers will then learn to create a simplified version that performs the same way but doesn’t require as much computing power.
However you slice it, they’re a long way now from where they’d like to get to…but remember, since 1958, computer power has doubled approximately every two years, and that trend is continuing. (This is called Moore’s Law, after Intel co-founder Gordon E. Moore, who noted the trend in 1965.)
Blue Brain is moving up to the next generation of supercomputer—and with it will be able to add in all the molecules and biochemical pathways in their model NCC, something they couldn’t do with their first machine.
All well and good, but what’s the point? After all, humans are already pretty good at making copies of the human brain: every child has one.
True, but it’s impossible to study a real human brain at this level. A computer simulation, unlike a brain, can be slowed down or even stopped so processes can be observed in detail. From that basic knowledge all sorts of benefits could flow.
For example, Markram believes that in a decade or two doctors may be able to draw on an enormous database to simulate patients individually to better predict how they’ll respond to a given drug or treatment.
More excitingly, Markram believes that as the computer simulation of the brain grows in complexity, “emergent properties” may arise…things like thought and creativity.
Will a computer-simulated brain someday create a new work of art or scientific invention?
Markram has a straightforward answer. “It’s not a question of years, it’s one of dollars. The psychology is there today and the technology is there today.
“It’s a matter of if society wants this. If they want it in 10 years, they’ll have it in 10 years. If they want it in 1,000 years, we can wait.”