How to mind your own business

As you are reading this page, your attention is rapidly shifting from word to word, but how is this phenomenon of attention effected by your brain? How does your brain manage to filter out all the other sensory information that is impinging upon it at this very moment (perhaps the noise of traffic outside, or birds singing in the garden) and focus on the task in hand? These are the questions which David LaBerge attempts to answer in Attentional Processing, one volume in a Harvard series on perspectives in cognitive neuroscience.

Attention, especially visual attention, has often been characterised metaphorically. Some researchers have argued that attention is like a spotlight. Others that it is more like a zoom lens. LaBerge begins by examining several such metaphors, making explicit some of the baggage that the metaphors carry. In doing so he confronts some difficult questions, such as whether attention can be divided (or whether we can only "simulate" divided attention by rapidly switching full attention between tasks), and whether in switching visual attention between two points it is necessary mentally to traverse the space in between.

LaBerge adopts a multidisciplinary perspective, drawing on empirical cognitive psychology, neurophysiology, and computational simulation work. He is concerned with what constitutes attention, and how is attention realised in neural circuits. LaBerge is therefore not concerned with just the results of attention, or the algorithm which produces those results, or even the neurophysiology which implements that algorithm. He is simultaneously concerned with all three levels, and as such his answer takes the form of a necessarily speculative model of attentional processes with subfunctions of those processes anchored to various neural elements. One of the more intriguing aspects of the model is that it suggests that attention cannot be localised. Attention is not, according to LaBerge, the task of a single neural structure, but the result of a number of neural structures and regions working in concert.

The psychological evidence for LaBerge's attention algorithm stems from reaction time experiments which aim to measure the angle of visual attention and the rate at which attention can be shifted. The neurophysiological evidence for the implementation of that algorithm is based on neuroanatomy (known connections between structures and regions within the brain), recordings of the activity of single cells in animal brains and non-invasive brain-imaging techniques applied to human brains. LaBerge proffers convincing surveys of each of these fields as he builds up to the presentation of his model. Fewer details are given of the computational work used in developing the arguments, and, more disappointingly, virtually no use is made of evidence from cognitive neuropsychology concerning attentional deficits in patients suffering brain damage. This is surprising given the attempted anchoring of functions to neural elements and hence the insights that damage to those neural elements would appear to offer.

A second disappointing aspect of the text is the rather anti-climactic presentation of the final model. After 200 pages of build-up the model is presented and discussed in a mere ten pages. Nevertheless, the attempt to marry the various fields is of considerable interest in itself, and LaBerge's effort in this respect represents a much better approximation to the elusive cognitive science approach than is generally found. Having said this, the readership may unfortunately be limited. The initial chapter is very lucid and written for a general audience, but later chapters will prove difficult without specialist knowledge of the methods of empirical cognitive psychology and the intricacies of neuroanatomy.

Richard Cooper is a lecturer in the department of psychology, Birkbeck College, London.