drupal statistics module

Machines Like Us

The brain race: can giant computers map the mind?

Friday, 22 February 2013
by Charles Watson

In the past month, we have seen two major announcements of huge projects to map the brain – the European Human Brain Project (HBP) and the Obama Brain Activity Map (BAM).

What you may not have noticed is a third, much more promising project announced by the Seattle-based Allen Institute for Brain Science to do similar things – but more on this later on.

Of the first two, the European HBP will give €1 billion to the Lausanne-based research group headed by Henry Markram. Markram is a brilliant salesman whose ambitious plan to make a working computer model of the cerebral cortex (“the Blue Brain Project”) has been strongly supported by IBM since about 2005.

The fact the Blue Brain project has not produced any significant breakthroughs in recent years does not seem to have worried the European funding agencies. Apparently they like the idea of Markram building a monster computer to lead Europe into the future of brain research.

The US plan is just as ambitious, but its aims seem to be more commercial and political than scientific. Obama hopes that companies such as Google and Microsoft will combine with universities and drug companies to lead the way to curing diseases such as Alzheimer’s.

No start-up funds have so far been allocated, but the plan clearly centres on the building of a massive computer network to simulate brain activity.

Obama sees the project as putting the US first in what he calls the “brain race” – just as Kennedy drove the space race competition with the Russians. Of course, this kind of announcement makes great political sense, but in my opinion it may be another case of the Emperor’s New Clothes.

Baby steps for the brain

The harsh truth is that brain research is still in its infancy, and big computers cannot replace our fundamental lack of understanding of how brains work.

I have watched the Markram project over the past five years and have been underwhelmed by the insights it has generated. For a start, Markram’s work focuses on a model of a tiny piece of rat cerebral cortex, which ignores the fact that the most important parts of the brain, in terms of survival, are outside the cerebral cortex.

We know that subcortical structures such as the hypothalamus can manage eating, drinking, reproduction, nurturing of offspring and defence all by themselves, but we are not even close to understanding the complex networks that make these basic systems work.

It is true the cerebral cortex of humans is awesomely powerful, but if we cannot even understand the basic survival functions of the brain, I think it’s a very long shot to predict that we can make an electronic cerebral cortex with a big computer.