The management of cognitive resources

Abstract

It is argued that an understanding of complex cognitive performance can best be achieved by considering both processing and representational cognitive resources. In any given task, control processes are important for configuring such resources appropriately and passing information between them. A computer controlled alphabet counting task which allows storage and processing requirements to be independently manipulated is used to gain a better understanding of the organisation and utilisation of resources by providing access to the microstructure of performance. Three main directions are explored. The first establishes baseline conditions for varying parameters of the task. Most notably, it demonstrates that resources are typically set up for the expected task difficulty prior to the task commencing, rather than as a consequence of immediate task demands. The second theme explores individual differences in carrying out one of the more complex conditions of the task, and shows that subgroups of subjects can be isolated who exhibit distinct patterns of performance. Moreover, in a task of this complexity, gross predictors of individual differences, such as IQ, do not relate to overall performance in any simple way, although they can be understood within each subgroup. The third group of experiments explore sensitivity to stressors external to the immediate task. Two 'environmental stressors' (alcohol and noise) and one 'cognitive stressor' (an additional concurrent memory load) are examined. Reliable differential effects are observed on the storage and processing phases of the task within individual subjects, but variations in the precise pattern of effects between subjects result in group data being potentially misleading. Finally, the requirements for an appropriate framework which can capture the most important aspects of resource management are considered, and a framework incorporating components of contemporary models of working memory is presented.