In 1948 the philosopher Bertrand Russell remarked that ‘the unitary space of common sense is a construction…it is part of the business of psychology to make us aware of the steps in this construction’ (Ref. 1, p. 234). Fifty years later, psychology has made considerable progress in understand- ing our perception of the space around us 2 . Dramatic and complementary advances have also been made on the topic in neuroscience 3,4 . It has become abundantly clear that the ‘space of common sense’ is indeed constructed, by compu- tations that involve considerable integration between our different senses. The outcome might more properly be de- scribed as common spaces for multiple senses, rather than the single space of a common sense. People normally experience external space as a seamless whole, but this belies the complexity of the underlying processes. The various senses can each sample different regions of space (e.g. we can hear sounds from the invisible region behind our head), and they do so in different co- ordinates (e.g. vision is initially retinotopic, touch is somato- topic, and hearing is first tonotopic then head-centred). Our perception of the location of external events depends on in- tegration of information from all these different senses. Such integration is required even for many tasks that might initially seem unimodal. For instance, perceiving the lo- cation of a light in darkness does not depend solely on visual inputs, but also on diffuse proprioceptive inputs, signalling the current position of the eye in its orbit, and of the neck on the trunk. This is shown when stimulation of the appro- priate muscles produces illusions of visual displacement 5 . Although initially surprising, such results make sense when one considers what is required for a stable representation of space. When taken in isolation, each modality signals stimu- lus location with respect to its own receptor surface only (e.g. on the eye for vision, or on the skin for touch). Since the receptors for each modality can move freely relative to external objects (as in eye- or hand-movements), and can also move relative to each other (as when making an eye- movement but no hand-movement, or vice versa), a single modality alone cannot provide a stable representation of external space. Instead, this requires integration of infor- mation from multiple sensory modalities, so that current posture is taken into consideration, as well as the location of stimuli on receptor surfaces. Until recently, such crossmodal issues were overlooked by the extensive literature on ‘spatial attention’; that is, on our ability to focus selectively on sensory information from particular locations in external space. However, in the last five years, there has been an explosion of crossmodal atten- tion studies 6–11 . These have revealed numerous crossmodal links in spatial attention, and their relation to the internal construction of external space. Crossmodal attention Traditional research on spatial attention considers only a single sensory modality at a time (e.g. just vision 12 or just hearing 13,14 ). However, in the real world our attention often has to be coordinated crossmodally, so that we can select in- formation from a common external source across several modalities, despite vast differences in the initial coding of that source by each modality. For instance, when listening to someone speaking in a noisy environment, it is useful to pick out not only the relevant speech-sounds from among 254 Attention and the crossmodal construction of space Jon Driver and Charles Spence Traditional studies of spatial attention consider only a single sensory modality at a time (e.g. just vision, or just audition). In daily life, however, our spatial attention often has to be coordinated across several modalities. This is a non-trivial problem, given that each modality initially codes space in entirely different ways. In the last five years, there has been a spate of studies on crossmodal attention. These have demonstrated numerous crossmodal links in spatial attention, such that attending to a particular location in one modality tends to produce corresponding shifts of attention in other modalities. The spatial coordinates of these crossmodal links illustrate that the internal representation of external space depends on extensive crossmodal integration. Recent neuroscience studies are discussed that suggest possible brain mechanisms for the crossmodal links in spatial attention. J. Driver is at the Institute of Cognitive Neuroscience, Department of Psychology, University College London, Gower Street, London, UK WC1E 6BT. C. Spence is at the Department of Experimental Psychology, University of Oxford, South Parks Road, Oxford, UK OX1 3UD. tel: +44 171 387 7050 fax: +44 171 436 4276 e-mail: j.driver@ucl. ac.uk Review Driver and Spence – Crossmodal spatial attention Copyright © 1998, Elsevier Science Ltd. All rights reserved. 1364-6613/98/$19.00 PII: S1364-6613(98)01188-7 Trends in Cognitive Sciences – Vol. 2, No. 7, July 1998