Authors:

DUNCAN, J.

Reference:

Year of publication:

2005

CBU number:

6009

Abstract:

This chapter addresses one of Pat Rabbitt's enduring concerns - the basis for general intelligence or Spearman's g. I begin with a striking finding from functional brain imaging: In specific regions of frontal and parietal cortex, there is a pattern of similar activation for many different cognitive demands. Plausibly, this pattern of broad cognitive activation (BCA) could reflect functions basic to g. A related finding comes from monkey studies: In the lateral prefrontal cortex, neurons produce a dense, selective representation of information relevant to a current task, whatever that task may be. This representation is reminiscent of a basic idea from artificial intelligence - task modelling in working memory - and I suggest that g may largely be a reflection of this task modelling function. Moving on to behavioural studies, I show how competition in a task model can result in loss of vulnerable task components. This "goal neglect" is closely related to g. Finally I review recent lesion data, and the prediction that g deficits should reflect damage to BCA regions. Among difficulties in testing this hypothesis are confounding of BCA damage with total lesion volume, and highly variable effects of apparently similar lesions.