Computational Modeling of Individual Differences in Working Memory and Strategy Adaptivity

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

Lynne Reder Abstract Performance on cognitive tasks varies among individuals. This project is a of study two of the sources of these individual differences; namely, working memory capacity (people's limited resources for retrieving and maintaining information during cognitive processing) and strategy adaptivity (people's ability to change their approach to a task in order to achieve greater success). The first goal of this project is to develop, test, and refine a theory of how individual differences in working memory capacity impact performance across multiple tasks. This theory will be developed as a computational model that can make accurate predictions of individual subjects' performances across multiple tasks at a fine-grained, quantitative level. Specifically, the computational model will enable the estimation of an individual's working memory capacity from one task and then use that estimate to make predictions of performance on the second task. The parameters can be interpreted to represent stable differences between subjects and can be used to predict the same individual's performance on other tasks. Predicting individuals' performances in this way has not been achieved before now and will be a major contribution of this research. The second goal of this project is to explore how differences in strategy adaptivity can be understood in terms of differences in working memory capacity. In sum, the project will result in several unique achievements: 1) the development of a new way to understand individual differences in working memory capacity and strategy adaptivity; 2) provision of a mechanistic account of these differences; and 3) a determination of whether computational modeling can be used to predict performance in terms of zero parameter model fitting.