Podium: How the brain copes with language



From the Michael Faraday Award lecture given by the director

of the Royal Institution to the Royal Society

WHAT IS it about the human brain that makes it so special and so different? If we look at language, might we get a clue as to what makes it so special? To do that we don't have to stay in this century; as Newton said, "we will stand on the shoulders of giants" - we shouldn't think that everything was invented just five years ago.

Paul Broca was a physician in France in 1860 and earned his place in the history books because one day he examined a patient named Leborgne - nobody knows him by that name, but by his nickname, Tan. He was called Tan because that is all he could say. He had a severe speech impediment and could not articulate at all. Tan earned his place in the history books by dying, which allowed Broca to look at his brain and discover a big hole in Tan's brain now called, for obvious reasons, Broca's area. Broca deduced that it was the damage to this area that had caused Tan's severe speech impediment.

Does this mean, then, that the Numbskull idea is right? The Numbskulls, of course, were the mainstay of the Beezer, who lived in a man's head and operated in rigid compartments in which they had specific jobs to do. Even some neuroscientists will speak about the "centre" for this, as if the brain had autonomous little regions operating like independent mini-brains. So can we can talk about a "centre" for language? Something is damaged and there is a speech impediment - surely that proves the point?

But, within 10 years, along came a German physician, Carl Wernicke, who showed that the situation is not so simple. Wernicke reported on patients who had speech problems, but theirs were slightly different. When asked to describe a picture, someone with so-called Wernicke's aphasia said "you go the people go in and once more there you'd have one your bees your bees oh and then come out in a few minutes. How many how many beeses. Eight many beeses. Ira comes the best and she in gone the pesh..."

It is jargon - absolute rubbish; "jargon aphasia" is a genuine clinical condition. And, importantly, the areas of the brain damaged in jargon aphasia are different from those damaged in Broca's aphasia. So already we know that we can't just talk about a single speech centre.

A technique that has spread through the neuroscience community like wildfire over the last decade or so, is brain imaging. This can be described as a way of seeing what we know and can learn nowadays about the localisation of language ability. As its name suggests, it is a way of looking at the living human brain at work in conscious patients while they are doing relatively sophisticated tasks. One particular technique is called positron emission tomography (PET) scanning. Like other imaging techniques, PET exploits the fact that your brain is very, very greedy for oxygen and glucose. It works much harder than other parts of your body at rest, relative to the weight it contributes. So, what you do in this case is to tag oxygen or glucose with a radioactive label, and introduce it into the body. The radioactive label will emit positrons that will then collide with electrons in the brain; the annihilation will cause energy waves to come out through the skull and strike sensors that can then be transformed into a computer image. You can then effectively see the brain lighting up.

What we see in the PET scan of a conscious subject performing different language skills tells us something already hinted at from the previous observations of the last century: that more than one brain area is involved.

My point is that there is no "single centre", but for whatever aspect of the language that is occurring, different brain areas are configured differently. Different constellations of brain areas are active for different aspects of a particular function. So the brain divvies up what we see as a unitary function in many different brain regions - and those regions will vary according to the particular task or aspect of the task that is occurring.

This means that the Numbskull idea is not right, but rather that many brain regions will contribute to a single function. It is a bit more like this (rather a simple-minded analogy): as if all your brain regions were different ingredients for a culinary dish, and in themselves could not account for the final taste, the net function - the whole being more than the sum of its parts of the final stew.

Perhaps a more lyrical analogy would be a symphony; different instruments all make their own contribution to the final sound. In attempting to analyse a symphony, something would be lost if you adopted that reductionist approach.