Poster A39, Saturday, March 25, 5:00 – 7:00 pm, Pacific Concourse
Neuroimaging, Neurostimulation, and Neuropsychological Evidence for Different States of Representation in Working Memory
Nathan Rose1, Bradley R Postle2; 1University of Notre Dame, 2University of Wisconsin-Madison
Previous fMRI and EEG research has supported the idea that dynamic attention processes enable the retention of items in working memory (WM) at varying levels of priority. Specifically, it has been suggested that information in WM, but outside of focal attention (FA), may be retained either by short-term, synaptic plasticity mechanisms or long-term potentiation/episodic-retrieval processes. Here, we consider causal investigations of this idea. In Study 1, multivariate pattern analysis (MVPA) of fMRI data from an individual with developmental amnesia and healthy controls performing WM tasks with either one item, or for two items with prioritization cues, revealed that items held in FA could be decoded from a classifier using data from the whole brain, but not the hippocampus. In contrast, passively retained items could not be decoded. For both the amnesic and controls, subsequent long-term memory (LTM) tests showed that word-stem completion priming was relatively insensitive to the amount of time an item was held in FA; moreover, subsequent cued recall and recognition was not better for passively retained items. In Experiment 2, we administered TMS to participants and recorded EEG while they performed the two-item WM task with prioritization cues. TMS transiently reactivated neural patterns associated with passively retained items, and induced greater connectivity within frontoparietal cortex than sensory cortex. Collectively, these results are both consistent with the hypothesis that short-term plasticity mechanisms underlie the prioritized state of information in WM but outside FA, and inconsistent with the idea that these items in WM are processed with LTM mechanisms.
Topic Area: EXECUTIVE PROCESSES: Working memory