The Private Life of the Brain
Susan Greenfield
Allen Lane, £18.99, 258pp

Shine a light at a cat. A quarter of a second later, 10m cat brain cells are telling each other that they saw something. (The same, presumably, goes for you and me, though we are less likely to have fluorescent dye applied to our exposed brains to check the numbers.) At some point, as a spreading wave of activation courses through neura networks, the cat will become conscious of the light, in a feline sort of way. If we want to understand consciousness, we have to explain how.

This is the challenge Susan Greenfield takes on, and this is the level at which she attacks the problem. As a neuroscientist with a penchant for pharmacology, she is convinced that real understanding of the mind will come from studying the soft, moist organ inside the skull, not by making smart computer models or by clever philosophising.

Her book is good on the detail, interpreting experiments and proposing new ones, and she is clear about the pitfalls of trying to move from statements about brain cells to ideas about the self which inhabits the neural maze. But the big picture, perhaps inevitably, is less satisfying. Her grand idea is that the most basic form of consciousness is intense emotion, not cool reason. People seem driven to seek a loss of identity or letting go as a respite from the more complex brainwork of maintaining meaning and memory. They find it through orgasm, dancing, drug-induced euphoria, or a cold dip on a hot day. Through such things we recover a childlike state of pure feeling.

The job of these cheap thrills, according to Greenfield, is to build a transient neural network which registers sensation intensely but is prevented from engaging larger networks which would bring in the slowly grown personality laid down inside an adult brain. And we regard thrills as cheap because somehow we sense that they have to be kept in check in deference to the mind's larger concerns.

Greenfield elaborates her sex and drugs and rock 'n' roll view of consciousness throughout the book, looking at children's emotions, the minds of depressives and schizophrenics, dreamers and junkies. She ends with the suggestion that the quality of succeeding moments of consciousness depends on the size of the transient neural networks activated. But this again betrays the difficulty of finding a useful language for following how a simple quantitative measure - the size of a network of cells signalling to each other - turns into changes in qualities of mind. Nor is the book's other closing motif, that of consciousness as being like ripples from a stone lobbed into a puddle, much help.

A cumbersome metaphor like this is hardly a theory; but all Greenfield really wants to claim, it turns out, is that her ideas suggest new experiments. The notion that the size of evoked neural networks is crucial to whatever is happening inside brains is testable. She can go back to the lab with a slightly modified research programme, while readers can note that at least one book on the mind resists the temptations either to oversell the science or to claim that no real progress has been made. And cats who would prefer to keep their brains to themselves had better watch out.

• Jon Turney teaches science communication at University College London