Sunday, 10 February 2013

Koch consciousness

Just finished reading an interesting book by Christof Koch called 'Consciousness'. An engaging and accessible canter through the fields of consciousness research, fields which are only now becoming a respectable place for respectable (and career conscious) scientists to do their ploughing.

Koch does some if not most of his work for the Allen Institute for Brain Science in Seattle (http://www.alleninstitute.org/), a not-for-profit medical research institute funded in some large part by Paul Allen, one of the founders of Microsoft. An institute which makes all of its product publically available on the internet, including, for example, far more than most of us want to know about the inner workings of a mousy brain.

Not aware that any UK businessmen fund public works in this way, let alone on this scale. What do Richard Branson or James Dyson do in this line? But it was not always thus: once upon time we had Mr Tate the Sugar who gave us the Tate Gallery (the Millbank one that is) and a chain of London Libraries, one of which can be observed from the terrace outside the Estrela Bar in Vauxhall. Perhaps we have declined to the point where we just don't have people rich enough to make a difference at this sort of thing? Perhaps we have declined to a nation of benefit junkies who expect all that sort of thing to be done by government?

But no immediate need to tap into the Allen Insitute as there is lots of good stuff in this book, from which I share a few snippets.

First, the human brain contains a lot of neurons, maybe 75 billion of them, coming in maybe 1,000 rather than 57 varieties. So there is no way that a computer built the way that computers are built now is going to be able to match these numbers. But it does not follow that a computer cannot model interesting aspects of the brains behaviour - although many people think that such modelling will never include all aspects. Human brains are special.

By way of comparison, the human gut contains just a few hundred million neurons. But there seems to be a consensus that there is not a rival, parallel consciousness banging away, unbeknown to us, in the abdomen.

One way in which computers are not like neurons is the environment, in particular the electrical and chemical environments, in which neurons live. Transistors do their stuff, we want them to do their stuff, quite unaffected by their environment. They just respond in a reliable way to signals from other transistors. Certainly one might model the behaviour of neurons in a computer, plenty of people do, but the foundations are fundamentally different: neurons are a lot more complicated than transistors.

Second, the human brain includes proper nouns. If I know of, for example, Gordon Brown, my brain will contain a small cluster of neurons from among all the billions which respond specifically to the mention of Gordon Brown. Show me a picture of the man and you can watch the neurons fire.

Third, all kinds of bizarre complaints are caused by damage to very specific and very small parts of the brain, perhaps the result of a stroke. For example, the ability to distinguish one human face from another. An ability which can, in effect, be switched off, leaving the rest of one pretty much intact. (One can, it seems, get along with such a disability. There are workarounds).

Fourth, while we do not yet know how the brain makes the mind, we do know that, at least in a limited way, the mind can prevail over the matter of the brain. If you wire someone up in the right sort of way, they are able to control some of their brain waves.

Fifth and last, one seems to be unconscious when the various areas of the brain stop talking to each other. All kinds of stuff - like the liver and the legs - can be looked after without such talking. But no such talking and no consciousness - talking which can be turned off with remarkable speed by modern anaesthetics. Consciousness seems to require a degree of integration - even if it needs neither spinal cord nor cerebellum.

Thoroughly recommended to anyone with an interest in the subject.

No comments:

Post a Comment