Maps of the mind

Virgil, in gloomy mood, once declared that "time wastes all things including the mind". He was echoed in Shakespeare's portrayal of man's final age "sans everything". But the inevitability of mental decline in old age has not been universally accepted. Cicero believed that the mind, like the body, was more likely to keep functioning well if it was given plenty of exercise. He practised what he preached by learning Greek in the last years of his life, and by taking time each evening to recall carefully everything he had heard during the day. He called this his intellectual exercise, his "running track for the brain".

Cicero's optimistic and practical approach to ageing finds support in research into Alzheimer's disease reported by David Smith in "Ageing and the brain: is mental decline inevitable?", his contribution to From Brains to Consciousness? He presents evidence that Alzheimer's and other severe forms of dementure are indeed diseases and not a natural or inevitable consequence of growing old. There are several risk factors related to Alzheimer's, and age is certainly a major one, but there is no reason to accept what Smith calls "the fatalistic view that mental decline is an inevitable accompaniment of ageing". That view can simply become a self-fulfilling prophecy, and there is evidence that simple cognitive tasks like jigsaw and crossword puzzles, and even light physical activities such as knitting or pottering in the garden, can slow or reverse mental decline in the elderly.

The scientific explanation for such self-help towards successful ageing draws on the observation that it is not the absolute number of brain cells that is significant but the density of the connections between them. The story of the brain's formation and basic anatomy, with its variety of glia cells and neurons (nerve cells) with their connecting dendrites and axons, are described clearly and relevantly by John Parna-valas in an early chapter "The human brain: 100 billion connected cells". The correlations between physical brain processes - especially the chemical and electrical communication between neurons - and mental processes such as thoughts and emotions - and especially memory - are explored from different angles by Trevor Robbins ("The pharmacology of thought and emotion"), Larry Squire ("Memory and brain systems") and Tim Bliss ("The physiological basis of memory").

These papers together prepare the ground for Smith's article, with which I began this review. His insistence that dementia is a symptom of disease, rather than a necessary natural process, leads naturally into a pair of chapters that focus on mental illness. Ever since the pioneer psychiatric work of Kraeplin and Bleuler about a century ago, mental disorders have been divided into two major categories: schizophrenia (or dementia praecox ) and manic depressive illness. In a very accessible survey of the history and current state of the diagnosis and treatment of insanity, Richard Bentall argues that this division is misconceived and seeks to show "Why there will never be a theory of schizophrenia".

By way of contrast, Tim Crow, in a paper boldly titled "Nuclear schizophrenic symptoms as the key to the evolution of modern homo sapiens ", argues that not only does schizophrenia exist, but its ubiquity and even spread across all human populations point to its causes lying at the very roots of our modern species. He links those causes with the emergence of language, suggesting that the same genetic changes that allowed the development of language also opened the door to schizophrenia.

Might a computer suffer from schizophrenia? Not according to Roger Penrose, who answers the question posed in his own chapter title, "Can a computer understand?", with a firm, "no". The familiar Penrose arguments against artificial intelligence are deployed with their customary clarity.

A quite different stance is taken in "A neurocomputational view of consciousness" by Igor Aleksander, who is well known as a champion not only of artificial intelligence but of artificial consciousness. He disarmingly begins his paper by attributing to a film interviewer the original use of the term consciousness in relation to his "artificial neural nets", and he goes on to proclaim his own conversion to the concept of machine consciousness and to explain what it would involve.

Aleksander finds that much contemporary research into consciousness, whether by scientists such as Penrose or mathematically trained philosophers such as David Chalmers, has more than a smack of "good old dualism" and he believes that "this is precisely where the conscious machine may be able to clarify things". He would agree with Penrose that a pre-programmed computer, such as the chess-playing Deep Blue, could never be conscious. But for him it is not the computation but the programming that is the problem.

The essence, says Aleksander, is to grasp "that the organism must own and recognise its knowledge-building processes. Putting a programmer in the loop removes the sense of control and ownership which the organism needs to possess. Without these its consciousness would be a rum thing". So even when discussing neural networks with a learning capacity, he again agrees with Penrose that many of these "simply learn to obey their trainers' wishes", which is tantamount to programming and so unable to support artificial consciousness. But Aleksander claims that by the process he calls "iconic learning" his artificial neural nets can advance beyond this stage into an independent state that could be conscious.

Richard Gregory (himself no stranger to artificial intelligence - more than 30 years ago he co-founded at Edinburgh the first department in Europe dedicated to it) contributes a short article exploring the consequences of the growing conviction among researchers: that visual perception owes as much to previously stored "top-down" knowledge as it does to "bottom-up" processing of signals from the retina. "Flagging the present with qualia" suggests that the role of qualia - those felt qualities of redness, brilliance, etc, that comprise our conscious experience - is to help us distinguish between images formed entirely of "top-down" processes, such as memories, and images partly created by the "bottom-up" processing of current stimuli. The dire consequences of failing to make such a distinction can be easily imagined; so if Gregory is right, he has found an answer to one of the most persistent questions about conscious experience: what is it actually for?

Susan Greenfield brings us back to neurophysiology with the $64,000 question: "How might the brain generate consciousness?" She accepts the standard line that the answer lies in the connections between neurons and goes on to tackle the question of "hardwiring" versus "plasticity". Neuroscientists split between those (such as Michael Gazzaniga) who believe the bulk of our neuronal connections are genetically determined and so "hardwired", and those (such as Michael Merzenich, whose distasteful experiments on monkeys provide some of the key evidence) who think the brain is very adaptable to change and displays "plasticity" by reorganising large sets of connections to meet new circumstances.

Greenfield has studied the question by focusing on the concept of varying degrees of consciousness and has developed a theory of temporary "neuronal assemblies". The more neurons assembled (interconnected) at a particular time, the higher the degree of consciousness. This approach could accommodate experiences such as dreaming and dementure, and also provide support for the intuition that babies are somehow "less conscious" than adults. It might also provide a way into the scientific study of consciousness in non-human animals - still a hotly disputed topic.

In "Consciousness from a neurobiological perspective", Wolf Singer uses visual perception to illustrate both the explanatory power and the limitations of neurobiology. He concludes that not only is consciousness a property of the whole brain rather than of individual brain cells, it also has a social dimension.

If proof were needed that Cicero's optimism is not entirely misplaced, it comes in the form of Mary Midgley, who took up a writing career as she retired from university teaching and 20 years on is still producing sharp, informed and original comment. In the concluding chapter, "One world, but a big one", she uses the analogy of the variety of world maps in an atlas - political, climatic, geological - to argue for the value of cooperation among a multiplicity of approaches to the science of consciousness.

To those familiar with the field of consciousness studies there is little new in this volume, but Steven Rose is to be congratulated on putting together an excellent overview of the present state of the sciences of the mind. It provides an accessible introduction for the general reader by some of the leading researchers into consciousness.

The Revd Anthony Freeman is managing editor, Journal of Consciousness Studies .