Cognitive neuropsychology and functional brain imaging: implications for functional and anatomical models of cognition.

We discuss the relations between functional imaging and cognitive neuropsychological research. We begin by elaborating on some of the problems of traditional neuropsychological research, which attempted to provide accounts of cognitive performance at a neural as well as at a functional level of description. The difficulties in making neural-level arguments from neuropsychological data include: problems of associated deficits, problems due to interactive effects between brain regions, problems with analyses based on behavioural syndromes, problems due to the influence of compensatory strategies, and problems in separating damaged from disconnected representations. We discuss how cognitive neuropsychology by-passed many of these problems by emphasising functional rather than neural-level theories, though problems with inferences at the neural-level remain. We then consider the contribution that functional imaging can make to cognitive neuropsychology. Using evidence drawn from studies of language, object recognition and visual attention, we argue that functional imaging complements cognitive neuropsychology by: (i) not being reliant on accidents of nature and by enabling effects of lesions on 'distant' neural areas to be measured, (ii) revealing the brain systems necessary and sufficient for a given task, (iii) providing tests of neural-level models of cognition, and by (iv) providing novel evidence on the mechanisms of functional recovery in patients. In addition to this, imaging studies can contribute directly to functional-level theories, by providing converging evidence on the neural locus of cognition--knowing 'where' can allow new inferences about 'how' a given task is performed.