Order and item information in working memory: Neuroimaging, patient, and behavioral evidence.

摘要

Working memory is thought to include a mechanism that allows for the coding of order information. One question of interest is how order information is coded, and how that code is neurally implemented. This dissertation is divided into two parts. The first is an empirical study using functional magnetic resonance imaging (fMRI) to examine the neural regions involved in order processing. It includes behavioral and fMRI findings from an experiment which involved comparing two tasks, an item-memory task and an order-memory task. In each case, five letters were presented for storage, followed after a brief interval by a set of probe letters. In the case of the item-memory task, the two letters were identical, and the subject responded to the question, "Was this letter one of the items you saw?". In the case of the order-memory task, the letters were different, and subjects responded to the question, "Are these two letters in the order in which you saw them?". Behaviorally, items that were farther apart in the sequence elicited faster reaction times and higher accuracy in the Order task. Areas that were significantly more activated in the Order condition included parietal and prefrontal cortex. Parietal activations overlapped those involved in number processing, leading to the suggestion that the underlying representation of order and numbers may share a common process, coding for magnitude.; The second part is a review of the behavioral, neuroimaging, and patient evidence for the separability of item and order memory and potential cognitive and neural mechanisms of the latter. The mechanisms discussed are inter-item associations, direct codes, dual codes, and magnitude codes, and it is suggested that different coding mechanisms may underlie order operations in different tasks. Furthermore, patient and neuroimaging evidence suggests that the prefrontal cortex and, perhaps, the parietal cortex are involved in order processing.